Papers
| Paper presentations : Monday | |||
|---|---|---|---|
| 1‑A | Avatar Representation, Embodiment, and Social Perception | 11:15‑12:15 | Chateaubriand |
| 1‑B | Perception, Interaction, and Learning in XR | 11:15‑12:15 | Maupertuis |
| 1‑C | XR technology infrastructure | 11:15‑12:15 | Lammenais 1/2 |
| 1‑D | Tracking and sensing | 11:15‑12:15 | Lammenais 3/4/5 |
| 2‑A | Immersive Learning | 14:00‑15:00 | Chateaubriand |
| 2‑B | Redirection | 14:00‑15:00 | Maupertuis |
| 2‑C | Multimodal Interaction and Audio Perception | 14:00‑15:00 | Lammenais 1/2 |
| 2‑D | Ethical, diversity and gender issues | 14:00‑15:00 | Lammenais 3/4/5 |
| 3‑B | Co-Presence, Agency, and Social Interaction | 15:15‑16:15 | Maupertuis |
| 3‑D | Haptic interfaces and rendering | 15:15‑16:15 | Lammenais 3/4/5 |
| 4‑B | Immersive analytics and visualization | 17:15‑18:15 | Maupertuis |
| 4‑C | Human factors and ergonomics | 17:15‑18:15 | Lammenais 1/2 |
| 4‑D | Evaluation methods | 17:15‑18:15 | Lammenais 3/4/5 |
| Paper presentations : Wednesday | |||
|---|---|---|---|
| 10‑A | Immersive Applications for Training, Shopping, and Social Interaction | 08:30‑09:30 | Chateaubriand |
| 10‑B | Enhancing Interaction and Feedback in Virtual and Cross-Reality Systems | 08:30‑09:30 | Maupertuis |
| 10‑C | Interactive Experience and System Usability | 08:30‑09:30 | Lammenais 1/2 |
| 10‑D | 3D authoring | 08:30‑09:30 | Lammenais 3/4/5 |
| 11‑A | Perception, Embodiment, and Spatial Awareness | 11:15‑12:15 | Chateaubriand |
| 11‑B | Accessibility of immersive interfaces | 11:15‑12:15 | Maupertuis |
| 11‑C | Embodiment, Agency, and Interaction | 11:15‑12:15 | Lammenais 1/2 |
| 11‑D | Social Interaction and Collaboration | 11:15‑12:15 | Lammenais 3/4/5 |
| 12‑A | Cybersickness | 14:00‑15:00 | Chateaubriand |
| 12‑B | Locomotion and Spatial Awareness | 14:00‑15:00 | Maupertuis |
| 12‑C | User experience and usability | 14:00‑15:00 | Lammenais 1/2 |
| 12‑D | Teleoperation and telepresence | 14:00‑15:00 | Lammenais 3/4/5 |
| 13‑B | Immersive applications and 3DUI | 15:15‑16:15 | Maupertuis |
| 13‑D | Computer graphics techniques | 15:15‑16:15 | Lammenais 3/4/5 |
Session: Avatar Representation, Embodiment, and Social Perception (1 - A)
GaussianHand: Real-Time 3D Gaussian Rendering for Hand Avatar Animation
Lizhi Zhao
Rendering animatable and realistic hand avatars is pivotal for enhancing user experiences in human-centered AR/VR applications. While recent initiatives have utilized neural radiance fields to forge hand avatars with lifelike appearances, these methods are often hindered by high computational demands and the necessity for extensive training views. In this paper, we introduce GaussianHand, the first Gaussian-based real-time 3D rendering approach that enables efficient free-view and free-pose hand avatar animation from sparse view images. Our approach encompasses two key innovations. We first propose Hand Gaussian Blend Shapes that effectively models hand surface geometry while ensuring consistent appearance across various poses. Secondly, we introduce the Neural Residual Skeleton, equipped with Residual Skinning Weights, designed to rectify inaccuracies involved in Linear Blend Skinning deformations due to geometry offsets. Experiments demonstrate that our method not only achieves far more realistic rendering quality with as few as 5 or 20 training views, compared to the 139 views required by existing methods, but also excels in efficiency, achieving up to 125 frames per second for real-time rendering and remarkably surpassing recent methods.
Effective VR Intervention to Reduce Implicit Bias towards People with Physical Disabilities: The Interplay between Experience Design and Individual Characteristics
Hyuckjin Jang, KAIST; Jeongmi Lee, KAIST
Studies on reducing bias against physical disabilities with VR have shown inconsistent results based on experience design and overlooked potential moderating effects of individual characteristics. This study examined how experience design components and individual characteristics (gender, preexisting bias) moderate VR's effect on bias towards physical disabilities. We designed a VR wheelchair experience manipulating situational context and avatar visualization. Participants' implicit and explicit bias levels were assessed before and after the VR experience. The findings underscore the importance of considering the interaction between experience design and individual characteristics in understanding VR's efficacy as an empathy-inducing tool.
The Impact of Avatar Clothing on the Proteus Effect and Stereotype Threat
Pauline W Cha, Davidson College; Fangyi Li, Davidson College; Tabitha C. Peck, Davidson College
We investigate if avatar clothing-swaps induce a Proteus effect. Specifically, we investigate the impact of a woman avatar wearing business or science attire during a stereotype threatening situation. Participants in a successful woman business avatar experienced gender stereotype threat while participants in a successful woman science avatar were instead buffered from gender stereotype threat. Further, subjective embodiment was a significant predictor of response measures such that participants with higher subjective embodiment scores had better cognitive performance. This work demonstrates the potential of the Proteus effect to promote and support traditionally marginalized groups.
Self-Similarity Beats Motor Contor in Augmented Reality Body Weight Perception
Marie Luisa Fiedler, University of Würzburg; Mario Botsch, TU Dortmund University; Carolin Wienrich, University of Würzburg; Marc Erich Latoschik, University of Würzburg
Our work examines how self-similarity and motor control affect sense of embodiment, self-identification, and body weight perception in Augmented Reality (AR). In a 2x2 mixed design experiment, 60 participants interacted with synchronously or independently moving virtual humans, either with self-similar or generic appearance, across two AR sessions. Results show that self-similarity enhances sense of embodiment, self-identification, and weight estimation accuracy, while motor control effects were weaker than in similar VR studies. Participants' body weight, self-esteem, and body shape concerns also impacted estimates. These findings deepen understanding of AR body weight perception, highlighting real-world coherence as a key factor.
Investigating the Impact of Video Pass-Through Embodiment on Presence and Performance in Virtual Reality
Kristoffer Waldow, TH Köln; Constantin Kleinbeck, Technical University of Munich; Arnulph Fuhrmann, TH Köln; Daniel Roth, Technical University of Munich
Creating a compelling sense of presence and embodiment can enhance the user experience in VR. Traditional approaches use personalized or video self-avatars but require external hardware and focus mainly on hand representations. This paper introduces video Pass-Through Embodiment (PTE), a method leveraging per-eye depth maps from Head-Mounted Displays to integrate users' real bodies into VR environments without additional hardware. In a study with 40 participants performing a sorting task, PTE enhanced presence and embodiment despite minor visual artifacts, with no negative impact on performance, cognitive load, or VR sickness. PTE therefore offers a practical alternative to avatar-based methods in VR.
Session: Perception, Interaction, and Learning in XR (1 - B)
On the Benefits of Sensorimotor Regularities as Design Constraints for Superpower Interactions in Mixed Reality
Jingyi Li, University of Cambridge; Per Ola Kristensson, University of Cambridge
Creating intuitive action-outcome mappings for augmented superpowers in Mixed Reality (MR) is a key challenge for designers. We explore sensorimotor regularities (SRs) as design constraints for superpower interactions by comparing three temporal manipulation methods: mid-air button control, SRs-incongruent gestures, and SRs-congruent gestures. Results show that SRs-congruent gestures significantly enhance task performance, reduce workload, and increase agency and presence compared to the other two methods, while also enabling faster learning. No significant differences were found between SRs-incongruent and button-based methods. These findings highlight the potential of SRs in designing intuitive superpower interactions for MR systems.
Perceptually-Guided Acoustic 'Foveation'
Realistic spatial audio rendering improves immersion in virtual environments. However, the computational complexity of acoustic propagation increases linearly with the number of sources. Consequently, real-time accurate acoustic rendering becomes challenging in highly dynamic scenarios such as virtual and augmented reality (VR/AR).
The Impact of Navigation on Proxemics in an Immersive Virtual Environment with Conversational Agents
Rose Connolly, Trinity College Dublin; Lauren Buck, University of Utah; Victor Zordan, Roblox Corporation; Rachel McDonnell, Trinity College Dublin
The management of interpersonal space - the distance people maintain between each other - is crucial for user comfort in VR. This is particularly relevant in social VR where interactions are frequent. Teleportation is a widely used locomotion method in these environments yet its effect on proximity remains unexplored. To address this, we measured the interpersonal distance of 70 participants interacting with virtual agents, comparing teleportation to natural walking. Results showed that participants maintained closer proximity during teleportation, emphasising the need for further research on how locomotion influences spatial perception and social dynamics in VR.
Implicit Learning of Professional Skills through Immersive Virtual Reality: a Media Comparison Study
Pierre Bondesan, Arts et Métiers Institute of Technology; Audrenne CANAL, Arts et Métiers Institute of Technology; Sylvain Fleury, Arts et Métiers Institute of Technology; Andréa Boisadan, CESI; Simon Richir, Arts et Métiers Institute of Technology
This study compares Immersive Virtual Reality (IVR) and traditional slideshow lessons for teaching implicit knowledge in professional decision-making tasks in viticulture. In a 2x1 between-subject design, forty engineering students learned to assess vine vigor through implicit learning, followed by real-world evaluation. The findings suggest that while immersion does not directly improve decision-making accuracy, it enhances intrinsic motivation and influences how knowledge is transferred to real-world tasks.
Misperception of the distance of virtual augmentations
Domenic Au, York University; Robert S. Allison, York University; Iroshini Gunasekera, York University; Laurie M Wilcox, York University
Binocular disparity provides metric depth, while monocular cues like occlusion indicate depth order. In augmented reality (AR), conflicts arise when virtual objects fail to be occluded by real-world surfaces, disrupting depth perception. Using distance-matching paradigms, we investigated these conflicts within (0.35-0.5 m) and beyond reach space (0.9-1.5 m). Observers matched a virtual letter's distance using a probe or manual reach. Results showed consistent underestimation of the letter's distance when rendered beyond a surface, with errors increasing with distance. These biases persisted despite proprioceptive input, underscoring occlusion conflicts' impact on depth perception and the need for careful AR design.
Session: XR technology infrastructure (1 - C)
VirtuEleDent: A Compact XR Tooth-Cutting Training System Using a Physical EMR-based Dental Handpiece and Teeth Model
Yuhui Wang, Tohoku University; Kazuki Takashima, Shibaura Institute of Technology; Masamitsu Ito, Wacom Co.; Ltd.; Takeshi Kobori, Wacom Co.; Ltd. EMR Technology; Tomo Asakura, Wacom Co.; Ltd. EMR Technology; Kazuyuki Fujita, Tohoku University; Hong Guang, Tohoku University; Yoshifumi Kitamura, Tohoku University
VirtuEleDent is a compact XR tooth-cutting training system designed for dental education. It combines a 3D-printed tooth model and a tracked handpiece rendered in a virtual environment via a mobile HMD. Using electromagnetic resonance (EMR) and IMU sensors, the system accurately tracks handpiece movements, providing realistic haptic feedback. Compared to conventional bulky and costly haptic devices, VirtuEleDent enhances accessibility and training opportunities. An expert user study with licensed dentists demonstrated its effectiveness and potential for future development in dental training applications.
Keep it Clean: The Current State of Hygiene and Disinfection Research and Practices for Immersive Virtual Reality Experiences
Emil Rosenlund Høeg, Aalborg University; Stefania Serafin, Aalborg University; Belinda Lange, Flinders University
Despite the growing commercial availability of VR headsets and state of the art hygiene solutions, the efficacy and safety remains largely understudied. This is a challenge that is affecting both private and public stakeholders in various fields who have adopted VR. This paper presents findings from a survey (n=42) of the most popular cleaning methods and procedures in various fields and respondents' consensus on the availability of research and guidelines to support their continuous safe operation. Moving forward, we seek to raise awareness of the limited state of VR hygiene research, and to initiate a call-for-action through future cooperative efforts.
Complex Virtual Environments on All-In-One VR Headsets Through Continuous From-Segment Visibility Computation
Voicu Popescu, Purdue University; Elisha Sacks, Purdue University; Zirui Zhang, Purdue University; Jorge Askur Vazquez Fernandez, Purdue University
All-in-one VR headsets have limited rendering power which limits the complexity of the virtual environments (VEs) that can be used in VR applications. This paper describes a novel visibility algorithm for making complex VEs tractable on all-in-one VR headsets. Given a view segment, the algorithm finds the set of triangles visible as a camera translates on the view segment. The visibility algorithm supports static and dynamic VEs, and it solves visibility with either triangle, particle, or object granularity. The visible sets yield output frames that are virtually indistinguishable from ground truth frames rendered from the original VEs.
No More Head-Turning: Exploring Passthrough Techniques for Addressing Rear Interruptions from the Front in VR
Zixuan Guo, Xi'an Jiaotong-Liverpool University; Yuekai Shi, Xi'an Jiaotong-Liverpool University; Tiantian Ye, Xi'an Jiaotong-Liverpool University; Tingjie Wan, Xi'an Jiaotong-Liverpool University; Hai-Ning Liang, The Hong Kong University of Science and Technology (Guangzhou)
Virtual reality (VR) users often encounter interruptions. While the Passthrough feature can help, when interruptions come from the rear, users still need to turn their heads to see the real world. Our Study 1's semi-structured interviews found that users were less likely to use Passthrough for rear interruptions due to large head-turning movements which were inconvenient and increased cybersickness. These findings led to three Passthrough techniques for displaying the rear view in front of the user. A second study showed that they reduced physical and temporal demands, alleviated cybersickness, and improved users' experience. Our work offers solutions that effectively manage rear interruptions while maintaining user comfort and experience.
XRXL: A System for Immersive Visualization in Large Lectures
Kabir Batra, Purdue University; Zirui Zhang, Purdue University; Shuwen Yang, Purdue University; Arnima Agrawal, Purdue University; Yiyin Gu, Purdue University; Bedrich Benes, Purdue University; Alejandra Magana, Purdue University; Voicu Popescu, Purdue University
This paper describes XRXL, an extended-reality system for increasing student engagement in large lectures. Students wear XR headsets to see 3D visualizations controlled by the instructor. The instructor can virtually retract the roof and walls of the classroom to allow for large-scale visualizations that extend beyond the physical boundaries of the classroom, or to turn the classroom into a 360 degree theater. The instructor can also partition the classroom into small groups of students and to assist individual groups as needed. XRXL was tested in a user study with 82 students in the context of a mock-lecture on neural networks, with good results.
Session: Tracking and sensing (1 - D)
An Early Warning System Based on Visual Feedback for Light-Based Hand Tracking Failures in VR Head-Mounted Displays
Mohammad Raihanul Bashar, Concordia University; Anil Ufuk Batmaz, Concordia University
State-of-the-art VR Head-Mounted Displays (HMDs) enable hand interaction with virtual objects, but hand tracking accuracy is limited by camera hardware and software. This work explores a visual feedback mechanism to create an early warning system that alerts users of potential hand skeleton recognition failures. Two user studies evaluated its effectiveness: a cup stacking task and a ball sorting task. Visual warnings based on tracking confidence improved system usability and reduced frustration. These findings, while focused on VR HMDs, have broader implications for any system utilizing hand tracking, such as robotics, to enhance reliability and user experience.
FovealNet: Advancing AI-Driven Gaze Tracking Solutions for Efficient Foveated Rendering in Virtual Reality
Wenxuan Liu, New York University; Budmonde Duinkharjav, New York University; Qi Sun, New York University; Sai Qian Zhang, New York University
This paper introduces FovealNet, an advanced AI-driven gaze tracking framework designed to optimize system performance by strategically enhancing gaze tracking accuracy. To further reduce the implementation cost of the gaze tracking algorithm, FovealNet employs an event-based cropping method that eliminates over 64.8% of irrelevant pixels from the input image. Additionally, it incorporates a simple yet effective token-pruning strategy that dynamically removes tokens without compromising tracking accuracy. Finally, to support different runtime rendering configurations, we propose a system performance-aware multi-resolution training strategy, allowing the gaze tracking DNN to adapt and optimize overall system performance more effectively.
SplatLoc: 3D Gaussian Splatting-based Visual Localization for Augmented Reality
Hongjia Zhai, Zhejiang University; Xiyu Zhang, Zhejiang University; Boming Zhao, Zhejiang University; Hai Li, RayNeo; Yijia He, RayNeo; Zhaopeng Cui, Zhejiang University; Hujun Bao, Zhejiang Univeristy; Guofeng Zhang, Zhejiang University
In this paper, we propose an efficient visual localization method capable of high-quality rendering with fewer parameters. Specifically, our approach leverages 3D Gaussian primitives as the scene representation. To ensure precise 2D-3D correspondences for pose estimation, we develop an unbiased 3D scene-specific descriptor decoder for Gaussian primitives, distilled from a constructed feature volume. Additionally, we introduce a salient 3D landmark selection algorithm that selects a suitable primitive subset based on the saliency score for localization. We further regularize key Gaussian primitives to prevent anisotropic effects, which also improves localization performance.
EX-Gaze: High-frequency and Low-latency Gaze Tracking with Hybrid Event-frame Cameras for On-Device Extended Reality
Ning Chen, Shandong University; Yiran Shen, Shandong University; Tongyu Zhang, Shandong University; Yanni Yang, Shandong University; Hongkai Wen, University of Warwick
We present EX-Gaze, an event-based real-time eye tracking system designed for on-device extended reality. EX-Gaze achieves a high tracking frequency of 2KHz, providing decent accuracy and low tracking latency. We have developed a lightweight tracking framework that enables real-time pupil region localization and tracking on mobile devices. To effectively leverage the sparse nature of event-streams, we introduce the sparse event-patch representation and the corresponding sparse event patches transformer as key components to reduce computational time. At last, EX-Gaze achieves real-time tracking at 2KHz without accumulating latency.
Synthesizing Six Years of AR/VR Research: A Systematic Review of Machine and Deep Learning Applications
Sarker Monojit Asish, Florida Polytechnic University; Bhoj Bahadur Karki, University of Louisiana at Lafayette; Bharat KC, University of Louisiana at Lafayette; Niloofar Kolahchi, University of Louisiana at Lafayette; Shaon Sutradhar, AIMEN; Spain.
Augmented Reality (AR), Virtual Reality (VR), and Mixed Reality (MR), combined with machine learning (ML) and deep learning (DL), represent both challenging and promising research areas. Despite their rapid growth, there has been a lack of comprehensive reviews of their contributions. This paper analyzes AR/VR/MR with ML/DL-focused 154 publications from IEEE VR and ISMAR (2018-2023), covering 12 categories like 3D reconstruction, cybersickness, gesture recognition, tracking, privacy/security, etc. It highlights ML/DL's impact, discusses emerging trends and challenges, and provides a curated list of publicly available datasets to support further research. This review serves as a valuable resource for advancing AR/VR/MR research.
Session: Immersive Learning (2 - A)
Enhancing Plant Variety Discovery Process with Visual Trait Assessment in VR
Muhammad Moiz Sakha, DFKI; Saarland Informatics Campus Florian Daiber, DFKI; Saarland Informatics Campus Matthias Enders, NPZ Innovation GmbH (NPZi); Christoph Tieben, DFKI; Benjamin KisLiuk, DFKI; Antonio Krüger, DFKI; Saarland Informatics Campus
Plant breeders evaluate crop varieties through field trials, but comparing candidates across locations and years is resource-intensive. We developed an integrated system combining robotic data collection with virtual reality for remote assessments. The robot autonomously captures images, spectral data, and 3D scans of canola trials. Our VR interface enables breeders to compare candidates across locations and growth stages' capabilities unavailable in traditional field assessments. In a user study with five breeders, the system demonstrated consistent scoring patterns. Breeder feedback highlighted that cross-location and temporal comparisons enhanced trait assessment decisions, showing potential for transforming agricultural workflows.
Peripheral Teleportation: A Rest Frame Design to Mitigate Cybersickness During Virtual Locomotion
Tongyu Nie, University of Minnesota; Courtney Hutton Pospick, University of Minnesota; Ville Cantory, University of Minnesota; Danhua Zhang, University of Minnesota; Jasmine Joyce DeGuzman, University of Central Florida; Victoria Interrante, University of Minnesota; Isayas Berhe Adhanom, Texas State University; Evan Suma Rosenberg, University of Minnesota
Mitigating cybersickness can improve the usability of VR and increase its adoption. We propose peripheral teleportation, a novel technique that creates a rest frame (RF) in the user's peripheral vision. Specifically, the peripheral region is rendered by a pair of RF cameras whose transforms are updated by the user's physical motion. We apply alternating teleportations or snap turns to the RF cameras to keep them close to the current viewpoint. We compared peripheral teleportation with a black FOV restrictor and an unrestricted control condition. The results showed that peripheral teleportation significantly reduced discomfort and enabled participants stay immersed longer.
Reduction of Motion Complexity as an Objective Indicator of Cybersickness in Virtual Reality
Jasmine Joyce DeGuzman, University of Central Florida; Kaori Hirano, Carleton College; Alice Guth, Davidson College; Tabitha C. Peck, Davidson College; Evan Suma Rosenberg, University of Minnesota; Tongyu Nie, University of Minnesota
Subjective measures, such as the Simulator Sickness Questionnaire, are widely used to assess cybersickness, but they often interrupt the user experience and are prone to bias. To overcome these limitations, researchers also investigated objective indicators. Based on the loss of complexity hypothesis, which suggests that disease or aging can produce a reduction of complexity in physiological system dynamics, we conducted an initial investigation of the relationship between movement complexity and cybersickness. We analyzed motion tracking collected from two previous VR locomotion studies using the d95 score. The results revealed a systematic relationship between movement complexity and cybersickness across both experiments.
Exploring Large Language Model-Driven Agents for Environment-Aware Spatial Interactions and Conversations in Virtual Reality Role-Play Scenarios
Ziming Li, Rochester Institute of Technology; Huadong Zhang, Rochester Institute of Technology; Chao Peng, Rochester Institute of Technology; Roshan L Peiris, Rochester Institute of Technology
Recent research explores using Large Language Model (LLM) agents to enhance Virtual Reality (VR) interactions, focusing on immersive chatbot experiences. However, current studies primarily address dialogue generation from user speech, overlooking the potential for richer, environment-aware interactions. This work proposes an approach that enables LLM agents to perceive VR environments and generate environment-aware interactions for more immersive human-AI experiences. We present a schema for describing VR environments and interactions via text prompts and evaluate it using five role-play scenarios with 14 participants. The findings highlight opportunities and challenges in developing environment-aware agents to enhance VR interactions.
Decoding Learner Behavior in Virtual Reality Education: Insights from Epistemic Network Analysis and Differential Sequence Mining
Antony Prakash, Indian Institute of Technology Bombay; Ramkumar Rajendran, IIT Bombay
This study investigates learner behavior patterns in Virtual Reality Learning Environments (VRLE) through an Interaction Behavioral Data (IBD) logging mechanism. We analyzed interaction traces from 30 undergraduate students learning electromagnetic induction concepts, using Epistemic Network Analysis and Differential Sequence Mining to compare high and low performers' behaviors. Results reveal that high performers engage in structured, iterative experimentation patterns, while low performers show less focused exploration. These insights inform the development of adaptive VR learning content and personalized scaffolding to enhance immersive VR educational experiences.
Session: Redirection (2 - B)
Dynamic Redirection for Safe Interaction with ETHD-Simulated Virtual Objects
Yuqi Zhou, Purdue University; Voicu Popescu, Purdue University
This paper proposes a redirection strategy for safe and effective haptic feedback with both stationary and moving virtual objects in virtual reality that are simulated by a table-top encountered-type haptic device (ETHD). Users interact with the virtual environment via a handheld stick, ensuring safety by avoiding contact with the ETHD during movement and synchronizing visual and physical feedback for effectiveness. A controlled within-subject user study (N = 26) evaluated static and moving virtual objects. Results demonstrate that dynamic redirection of the virtual stick outperforms no redirection and static redirection, achieving both safety and effectiveness across objective and subjective metrics.
ShiftingGolf: Gross Motor Skill Correction using Redirection in VR
Chen-Chieh Liao, Institute of Science Tokyo; Zhihao Yu, Sichuan Gaolu Information Technology Co.; Ltd.; Hideki Koike, Institute of Science Tokyo
Unintentional habits hinder sports performance, especially in golf, where correcting ingrained swing paths is challenging. This study proposes a VR-based training system that uses forward ball shifts to prompt golfers to modify their swing motion, eliminating undesirable habits. We develop and evaluate three ball shift modes, including a novel gradual-interspersed mode designed for post-training retention. A comprehensive user study demonstrates that the proposed redirection approach effectively corrects swing paths and promotes sustained learning effects. This work highlights the potential of VR-based visuomotor redirection for broader applications in sports training.
Detection Thresholds for Replay and Real-Time Discrepancies in VR Hand Redirection
Kiyu Tanaka, The University of Tokyo; Takuto Nakamura, The University of Tokyo; Keigo Matsumoto, The University of Tokyo; Hideaki Kuzuoka, The University of Tokyo; Takuji Narumi, The University of Tokyo
Hand redirection can modify perception and movement by providing real-time corrections to motor feedback. In the context of motor learning, observing replays of movements can enhance motor function. The application of hand redirection to these replays by making movements appear larger or smaller than they actually are has the potential to improve motor function. We conducted two psychophysical experiments to evaluate how much discrepancy between replayed and actual movements can go unnoticed by users, both with hand redirection (N=20) and without (N=18). Our findings reveal that the detection threshold for discrepancies in replayed movements is significantly different from that for real-time discrepancies.
Redirected Drawing: Expanding the Perceived Canvas Size in VR
Kumpei Ogawa, Research Institute of Electrical Communication; Tohoku University; Kazuyuki Fujita, Tohoku University; Kazuki Takashima, Shibaura Institute of Technology; Yoshifumi Kitamura, Tohoku University
Drawing, writing, and painting on 2D surfaces in VR offers arbitrary interaction with virtual canvases of unlimited size. We propose Redirected Drawing, a novel methodology that applies visual manipulation to the user's drawing movement in VR to expand the perceived surface size required for the drawing experience. To this end, this study specifically explores translation gain manipulation, where the displacement of a drawing stroke in reality is simply multiplied by a gain to represent it in VR. Our results show that pen-based drawing on a physical surface allows for a gain of up to 1.16 without being noticed by the user, which is more suited to expanding a virtual canvas size than finger-based drawing on a physical surface or in mid-air.
Redirection Detection Thresholds for Avatar Manipulation with Different Body Parts
Ryutaro Watanabe, The University of Tokyo; Azumi Maekawa, The University of Tokyo; Michiteru Kitazaki, Toyohashi University of Technology; Yasuaki Monnai, The University of Tokyo; Masahiko Inami, University of Tokyo
This study examines how the body part controlling a VR avatar and the avatar's appearance impact redirection detection thresholds. Experiments compared hand and foot manipulation of a hand-shaped avatar and an abstract spherical avatar. Results showed a 21% higher detection threshold with the hand avatar than the abstract avatar, regardless of the body part used. Redirecting the avatar toward the body midline increased thresholds by 49% compared to redirection away from it. No differences were found between hand and foot manipulations. These findings highlight the importance of avatar appearance and redirection direction for designing effective VR interactions.
Session: Multimodal Interaction and Audio Perception (2 - C)
Multimodal Neural Acoustic Fields for Immersive Virtual Reality
Guansen Tong, UNC Chapel Hill; Johnathan Chi-Ho Leung, UNC Chapel Hill; Xi Peng, UNC Chapel Hill; Haosheng Shi, UNC Chapel Hill; Liujie Zheng, UNC Chapel Hill; Shengze Wang, UNC Chapel Hill; Arryn Carlos O'Brien, UNC Chapel Hill; Ashley Paula-Ann Neall, UNC Chapel Hill; Grace Fei, UNC Chapel Hill; Martim Gaspar, UNC Chapel Hill; Praneeth Chakravarthula, UNC Chapel Hill
We propose multimodal neural acoustic fields for synthesizing spatial sound and creating immersive auditory experiences from novel viewpoints and unseen environments. Extending neural radiance fields to acoustics, our hybrid transformer-convolutional network captures scene reverberation and generates spatial sound using sparse audio-visual inputs. By representing spatial acoustics, our method enhances presence in augmented and virtual reality. Validated on synthetic and real-world data, it outperforms existing methods in spatial audio quality and nonlinear effects like reverberation. Studies confirm significant improvement in audio perception for immersive mixed reality applications.
Enhancing Empathy for Visual Impairments: A Multi-modal Approach in VR Serious Games
Yuexi Dong, Beijing Jiaotong University; Haonan Guo, Beijing Jiaotong University; Jingya Li, Beijing Jiaotong University
This paper explores the design and evaluation of SEEK VR, a multi-modal virtual reality game aimed at fostering empathy for visually impaired (VI) individuals. Through a combination of visual feedback, 3D spatial audio, and haptic feedback, SEEK VR simulates the challenges faced by VIs, allowing non-disabled players to experience their daily difficulties. Results from a user study of 24 participants indicate significant improvements in empathy and awareness. This work demonstrates the potential of multi-modal VR-based serious games to raise awareness and encourage social change by immersing players in the perspectives of VIs.
Perceptual Alignment of Spatial Auditory and Tactile Stimuli for Effective Directional Cueing
Dajin Lee, Pohang University of Science and Technology (POSTECH); Seungmoon Choi, Pohang University of Science and Technology (POSTECH)
Spatial audio and directional tactile cues are crucial for target localization and collision avoidance. Spatially aligned auditory and tactile stimuli can perform better than misaligned ones, but achieving precise alignment is challenging. This study aims to (1) identify just noticeable differences (JNDs) of direction between auditory and tactile stimuli in the horizontal plane and (2) evaluate hazard avoidance performance with misaligned cues within the JNDs. The estimated JNDs increase from 26° to 84° as the stimuli shifted from the center to the side of the body. Perceptual alignment of auditory and tactile cues within the JND ranges yielded comparable hazard avoidance performance and usability to exact physical alignment.
VR Whispering: A Multisensory Approach for Private Conversations in Social Virtual Reality
Xueyang Wang, Tsinghua University; Kewen Peng, Shanghai Jiao Tong University; Chonghao Hao, Shanghai Jiao Tong University; Wendi Yu, Tsinghua University; Xin Yi, Tsinghua University; Hewu Li, Tsinghua University
We introduce Whisper, a novel multisensory interaction technique that enhances private conversations in social Virtual Reality. Based on formative research (N=20), Whisper simulates natural whispering through synchronized visual (avatar proximity and gestures), auditory (voice conversion), and tactile (simulated airflow) feedback, while maintaining public conversation awareness through contextual records. Our comparative evaluation (N=24) in social scenarios demonstrates that Whisper provides an intuitive and immersive private communication method, significantly outperforming existing approaches in privacy, mode distinguishability, intimacy, and social presence.
Evaluating Dynamic Delivery of Audio+Visual Message Notifications in XR
Hyunsung Cho, Carnegie Mellon University; Drew Edgar, University of Glasgow; David Lindlbauer, Carnegie Mellon University; Joseph O'Hagan, University of Glasgow
Notifications play a crucial role in XR, informing users of received messages they might otherwise miss. However, it is unclear how attention-grabbing they should be, how much information they should present, and how the presentation should adapt to the message's content. We evaluated nine audiovisual notification designs on their usability and user preferences in XR environments, and how these adapt to varying contexts. Our results show differing effects of multimodal notification designs and that individuals want notification modality and design to vary based on delivered message content. These results offer new insights into developing context-aware multimodal interaction strategies for notifications in XR.
Session: Ethical, diversity and gender issues (2 - D)
Exploring the Uncoordinated Privacy Protections of Eye Tracking and VR Motion Data for Unauthorized User Identification
Samantha Aziz, Texas State University; Oleg Komogortsev, Texas State University
Virtual reality (VR) sensors capture large amounts of user data, including body motion and eye tracking, that contain personally identifying information. While privacy-enhancing techniques can obfuscate this data, incomplete privacy protections risk privacy leakage, which may allow adversaries to leverage unprotected data to identify users without consent. This work examines the extent to which unprotected body motion data can undermine privacy protections for eye tracking data, and vice versa, to enable user identification in VR. These findings highlight a privacy consideration at the intersection of eye tracking and VR, and emphasize the need for privacy protections that address these technologies comprehensively.
Embedding Human Values into the Designing of Mixed-Reality Technologies
Mengxing Li, Monash University; Taghreed Alshehri, Monash University; Tim Dwyer, Monash University; Sarah Goodwin, Monash University; Joanne Evans, Monash University
Current mixed reality designs predominantly prioritise functionality and usability, often overlooking individual's diverse value needs. To create more meaningful MR experiences, this paper aims to explore how human values can be integrated into MR design. We propose a values-based design process and evaluate it through three design workshops considering a remote collaborative learning scenario. By comparing our approach to an existing MR application, we demonstrate how embedding human values into MR design can lead to more ethical and human-centred outcomes. Our findings contribute to advancing the design and application of MR technologies to be more aligned with people's diverse values.
The Impact of Environment Design Bias on Working Memory
Jack Schwanewede, Davidson College; Alice Guth, Davidson College; Tabitha C. Peck, Davidson College
This work investigates subtle design differences in virtual environments on women+ performance. Non-male participants completed the Stroop Interference Task in two virtual environments: a neutral and a stereotypically STEM environment altered by four wall posters depicting positive gender-neutral to positive male figures, such as Albert Einstein. Results support that participants in the stereotypical environment were more distracted and responded more slowly and less accurately than in the neutral environment. We emphasize the importance of bias awareness in study design, and recommend that researchers consider how their experiments could impact participants of all demographics, in order to enhance inclusivity and non-biased results.
Observations on Virtual Reality Avatar Alignment with Research Participants' Skin Tone and Gender
Kristin Bartlett, University of Kentucky; Claudia Krogmeier, University of Wisconsin
In this paper, we look specifically at the avatars used in VR research. We analyzed 40 recent studies in which research participants embodied full avatars or were given virtual hands. While nearly every study matched the gender of the participant to the gender of the avatar, similar efforts were not made regarding skin color. We draw attention to two serious problems in VR research: the explicit and implicit exclusion of participants on the basis of skin color through avatar design. We provide recommendations for researchers with the goal of both mitigating and bringing awareness to current issues in VR research.
Enhancing Obstacle Visibility with Augmented Reality Improves Mobility in People with Low Vision
Lior Maman, University of Haifa; Ilan Vol, Ben Gurion University; Sarit Szpiro, University of Haifa
Low vision (LV) refers to functional yet reduced visual abilities, which pose safety challenges, particularly for mobility. We tested how increasing obstacle visibility by highlighting the obstacle outline in AR supports people with LV walk an obstacle course. Our study reveals that LV participants walked faster and took shorter paths in the AR-enhanced condition as compared to the control condition without enhancements. Our findings demonstrate the potential of object-specific AR enhancements to improve accessibility in real-life scenarios for people with LV and underscores the promise of AR and computer vision in the development of future assistive technologies.
Session: Co-Presence, Agency, and Social Interaction (3 - B)
It's My Fingers' Fault: Investigating the Effect of Shared Avatar Control on Agency and Responsibility Attribution
Xiaotong Li, The University of Tokyo; Yuji Hatada, The University of Tokyo; Takuji Narumi, the University of Tokyo
Previous studies introduced a system known as 'virtual co-embodiment,' where control over a single avatar is shared between two people. We investigate how this experience influences people's agency and responsibility attribution through: (1) explicit agency questionnaires, (2) implicit intentional binding (IB) effects, (3) responsibility attribution measured through financial gain/loss distribution, and (4) interviews. Agency questionnaires indicated that losing control over the fingers' movements negatively affected the agency over both the fingers and the entire upper limb. IB demonstrated that participants felt greater causality for failed attempts, they were reluctant to take responsibility and accept financial penalties.
Behavioral Measures of Copresence in Co-located Mixed Reality
Pierrick Uro
When several people are co-located and immersed in a mixed reality environment, they may feel like they share the virtual environment or not.
Avatars, Should We Look at Them Directly or through a Mirror?: Effects of Avatar Display Method on Sense of Embodiment and Gaze
Kizashi Nakano
In this study, by using head-mounted display with an extended downward field of view (FOV) and an eye-tracking function, we investigated the effects of presentation on the avatar body, which cannot be easily viewed with a limited downward FOV, on the sense of embodiment, presence, difficulty, speed of reaching tasks, and head behavior patterns.
A Comparison Study Understanding the Impact of Mixed Reality Collaboration on Sense of Co-Presence
Jianing Yin, Tsinghua University; Weicheng Zheng, Duke Kunshan University; Yuntao Wang, Tsinghua University; Xin Tong, Hong Kong University of Science and Technology (Guangzhou); Yukang Yan, University of Rochester
Sense of co-presence, the perceived closeness and interaction in collaboration, is key to effective teamwork. This study examines the novel impact of Mixed Reality (MR) on co-presence in remote settings compared to video conferencing. In a comparison with 14 dyads performing block assembly tasks across MR, video conferencing, and co-located scenarios, MR achieved co-presence levels comparable to physical co-location. Co-presence was also predicted through behavioral indicators such as deictic actions, confirmation behaviors, and language use. By enabling natural, action-based interactions and reducing reliance on verbal communication, MR demonstrates its transformative potential for remote collaboration.
Simulating Social Pressure: Evaluating Risky Behaviors in Construction Using Augmented Virtuality
Shiva Pooladvand, Arizona State University; Sogand Hasanzadeh, Purdue University; George Takahashi, Purdue University; Kenneth Park, Purdue University; Jacob Marroquin, Purdue University
Drawing on social influence and behavioral intention theories, coworkers' risk-taking serves as an extra motive -an exogenous factor- for risk-taking behaviors among workers in the workplace. Social influence theories have shown that social factors, such as social pressure and coworker risk-taking, may predict risk-taking behaviors and significantly affect decision-making. This study developed a state-of-the-art Augmented Virtuality (AV) environment to investigate roofers' risk-taking behaviors when exposed to social stressors (working alongside a safe/unsafe peer). The results demonstrated that the developed AV model could successfully enhance participants' sense of presence and social presence.
Session: Haptic interfaces and rendering (3 - D)
Influence of haptic feedback on perception of threat and peripersonal space in social VR
Vojta Smekal, Maastricht University; Jeanne Hecquard, Inria; Sophie Kühne, Maastricht University; Nicole Occidental, Maastricht University; Anatole Lécuyer, Inria; Marc J-M Macé, CNRS / Rennes 1; Beatrice de Gelder, Maastricht University
Humans notably interpret social interactions through nonverbal cues like body language, personal space, and touch. In VR, personal space violations can feel more severe, with psychological effects akin to real life. While 'personal bubble' features in VR aim to prevent harassment, they may not suffice. This study evaluates how social touch influences personal space perception. Participants wearing a haptic compression belt encountered virtual agents displaying threatening or neutral behavior. Personal space violations were perceived earlier with angry agents and haptic feedback. Results align with previous studies, highlighting haptic devices' potential to refine VR social interaction design.
PantographHaptics: A Technique for Large-Surface Passive Haptic Interactions using Pantograph Mechanisms
Marcus Friedel, Vireo Interfaces; Zachary McKendrick, University of Waterloo; Ehud Sharlin, University of Calgary; Ryo Suzuki, University of Colorado Boulder
In Virtual Reality, existing hand-scale passive interaction techniques are unsuitable for continuous large-scale renders: room-scale proxies lack portability, and wearable robotic arms are energy-intensive and induce friction. This paper introduces PantographHaptics, a technique which uses the scaling properties of a pantograph to passively render two-degree-of-freedom body-scale surfaces. We demonstrate PantographHaptics through a grounded prototype, HapticLever, and a wearable prototype, Feedbackpack. We evaluate these prototypes with technical and user evaluations. A 9-participant study compares HapticLever against traditional haptics, while a 7-participant second study verifies Feedbackpack's usability and interaction fidelity.
Shiftly: A Novel Origami Shape-Shifting Haptic Device for Virtual Reality
Tobias Batik, Complexity Science Hub; Hugo Brument, TU Wien; Khrystyna Vasylevska, TU Wien; Hannes Kaufmann, Vienna University of Technology
Shiftly is a low-weight shape-shifting haptic device that enhances Virtual Reality (VR) by providing plausible haptic feedback for differently shaped virtual objects. Using three actuators and foldable curved origamis, Shiftly approximates various touch surfaces, from flat to curved. We detail its design, kinematic model, and integration into VR setups. Two user studies evaluated its performance: one tested shape approximation without visuals, and the other assessed haptic realism for different virtual objects. Results showed Shiftly effectively rendered flat, convex, and edge-like geometries, though concave shapes and fine details were less realistic, demonstrating its potential for immersive VR haptics.
The Impact of Airflow and Multisensory Feedback on Immersion and Cybersickness in a VR Surfing Simulation
Premankur Banerjee, University of Southern California; Mia P Montiel, University of Southern California; Lauren Tomita, University of Southern California; Olivia Means, University of Southern California; Jason Kutch, University of Southern California; Heather Culbertson, University of Southern California
This paper explores enhancing immersion and reducing cybersickness in a VR surfing simulation through multisensory feedback integration. Using a custom airflow system and a surfboard on a 6-DoF motion platform, we conducted two studies. The first examines the effects of variable airflow synchronized with VR speed versus constant airflow, highlighting significant benefits for active user engagement. The second evaluates the combined impact of visual, motion, and airflow cues on acceleration perception, user immersiveness and cybersickness. Results show that synchronized multisensory feedback delivers the most immersive and comfortable experience, offering insights for designing dynamic, motion-intensive VR simulations.
Bimanual Ultrasound Mid-Air Haptics for Virtual Reality Manipulation
Lendy Mulot
The ability to manipulate and physically feel virtual objects without any real object being present and without equipping the user has been a long-standing goal in virtual reality (VR). Emerging ultrasound mid-air haptics (UMH) technology could potentially address this challenge, as it enables remote tactile stimulation of unequipped users. However, to date, UMH has received limited attention in the field of haptic exploration and manipulation in virtual environments. Existing work has primarily focused on interactions requiring a single hand and thus the delivery of unimanual haptic feedback. Despite being fundamental to a large part of haptic interactions with our environments, bimanual tasks have rarely been studied in the field of UMH interaction in VR. In this paper, we propose the use of non-coplanar mid-air haptic devices for providing simultaneous tactile feedback to both hands during bimanual VR manipulation. We discuss coupling schemes and haptic rendering algorithms for providing bimanual haptic feedback in bimanual interactions with virtual environments. We then present two human participant studies, assessing the benefits of bimanual ultrasound haptic feedback in a two-handed grasping and holding task and in a shape exploration task. Results suggest that the use of multiple non-coplanar UMH devices could be an interesting approach for enriching unencumbered haptic manipulation in virtual environments.'
Session: Immersive analytics and visualization (4 - B)
Minimalism or Creative Chaos? On the Arrangement and Analysis of Numerous Scatterplots in Immersive 3D Knowledge Spaces
Melanie Derksen, TU Dortmund University; Torsten Wolfgang Kuhlen, RWTH Aachen University; Mario Botsch, TU Dortmund University; Tim Weissker, RWTH Aachen University
Working with scatterplots is a common task in data analysis, becoming more complex as more plots are needed to understand the data. This paper explores how different immersive virtual environments (iVEs) affect users' ability to quickly retrieve scatterplots from a larger collection. We tested three scenarios: an Empty scene, a single landmark (Desk), and multiple landmarks (Office). In all conditions, users could freely arrange plots in 3D space. Results from a study with 45 participants reveal that the time and effort spent arranging the plots had a greater impact on memory performance than the iVE design. The findings highlight the significance of an active arrangement phase in supporting spatial memorization of scatterplots within iVEs.
Perception of Visual Variables on Virtual Wall-Sized Tiled Displays in Immersive Environments
Dongyun Han, Utah State University; Anastasia Bezerianos, LISN; Université Paris-Saclay; CNRS; INRIA; Petra Isenberg, Université Paris-Saclay; CNRS; Isaac Cho, Utah State University
We investigate the perception of visual variables on wall-sized tiled displays within an immersive environment. We conducted two user studies focusing on elementary visualization reading tasks in VR. The first study compared three different virtual display arrangements. Following that, we compared the results with those from a previous study conducted in a real-world setting. The comparative analysis showed that virtual curved walls resulted in smaller errors than the real-world flat wall display, but with longer task completion time.The second study evaluated the impact of four 3D user interaction techniques on performing the elementary task on the virtual wall display. We report our finings and discuss the limitations and future work.
Immersive Analytics as a Support Medium for Data-Driven Monitoring in Hydropower
Marina L. Medeiros, VRVis Forschungs-GmbH; Hannes Kaufmann, Vienna University of Technology; Johanna Schmidt, VRVis Zentrum für Virtual Reality und Visualisierung Forschungs-GmbH
Our Immersive Analytics application is part of a research project that combines and compares simulated water turbine flows and sensor-measured data, looking for data-driven predictions of the lifetime of the mechanical parts of hydroelectric power plants. Our VR prototype combines spatial and abstract data in which the user can navigate through a full-scale model of a turbine, view simulated water flows of three different energy supply conditions, and visualize and interact with sensor-collected data. Our findings confirm the current literature that IA applications add value to the presentation and analysis of situated data, and advance in the design directions for IA applications for domain experts.
Immersive Analytics for Understanding Ecosystem Services Data
Benjamin Powley, Linnaeus University; Craig Anslow, Victoria University of Wellington; Mairéad de Róiste, Victoria University of Wellington; Stuart Marshall, Victoria University of Wellington
Planning land use changes requires input from experts across several domains with diverse expertise. An immersive VR system, Immersive ESS Visualizer, was developed and implemented for analysing and visualizing ecosystem services. Our system was evaluated and compared to existing methods for ecosystem services analysis (paper maps, 2D screen). In VR, participants arranged handheld maps into complex 3D layouts, intersecting and tilting elevation. Although VR presented some difficulties, the benefits of spatially arranging maps, inspecting fine details, zoom navigation, and hillshade inspection made Immersive ESS Visualizer effective. Our research demonstrates that geospatial analysis can be supported by comparative visualization in VR.
PwP: Permutating with Probability for Efficient Group Selection in VR
Jian Wu, Beihang University; Weichen Zhang, Beihang University; Handong Chen, Beihang University; Wei Lin, Beihang University; Xuehuai Shi, Nanjing University of Posts and Telecommunications; Lili Wang, Beihang University
Group selection in virtual reality is an important means of multi-object selection, which allows users to quickly group multiple objects and can significantly improve the operation efficiency of multiple types of objects. In this paper, we propose a group selection method based on multiple rounds of probability permutation, in which the efficiency of group selection is substantially improved by making the object layout of the next round easier to be batch-selected through interactive selection, object grouping probability computation, and position rearrangement in each round of the selection process.
Session: Human factors and ergonomics (4 - C)
Virtual teleoperation performance under load factor in real flight
Mathurin Franck, Fédération ENAC ISAE-SUPAERO ONERA; Maëlis Lefebvre, Fédération ENAC ISAE-SUPAERO ONERA; Raphaëlle N. Roy, Fédération ENAC ISAE-SUPAERO ONERA; Vsevolod Peysakhovich, Fédération ENAC ISAE-SUPAERO ONERA
Manned-unmanned teaming (MUM-T) allows an operator aboard a manned aircraft to control UAVs. As MUM-T interfaces evolve, understanding how dynamic environments affect operators' perception and control is crucial. This study investigates the effects of load factor and visual perspective on teleoperation performance, workload, and motion sickness. 14 participants completed a simulated task aboard a Robin DR400 aircraft during turns of 30°, 45°, and 60° (load factors of 1.2, 1.4, and 2g). Results show increased mental and physical demands under higher load factors, but unaffected teleoperation performance. No differences were observed between first- and third-person perspectives. The performance was maintained despite the increased workload.
Performance and ergonomics of automated versus manual validation for AR-supervised industrial operations
Gaspard Laouénan, Univ Evry; Université Paris-Saclay; Jean-Yves Pascal Didier, Université Paris-Saclay; Paul-Eric Dossou, University Gustave Eiffel
Validation tasks in AR-supervised environments can be automated to improve process quality in industrial contexts. However, human-centered approaches urge us to study the impact of such methods on operator's acceptance. The paper presents a study on 24 participants to compare 4 modalities of validation of operations. SCRAM, a WebXR application is developed to compare validation tasks on 2 types of devices: tablets and HMDs. All scenarios are tested with an assembly use case through performance, task load, usability and acceptance criteria. Result show significant differences regarding the choice of the device but do not highlight any difference between scenarios with automated validation of operations and scenarios with manual validation.
AdaptiveCoPilot: Design and Testing of a NeuroAdaptive LLM Cockpit Guidance System in both Novice and Expert Pilots
Shaoyue Wen, New York University; Michael Middleton, New York University; Songming Ping, imperial college london; Nayan N Chawla, Virginia Tech; Guande Wu, New York University; Bradley S Feest, Northrop Grumman Mission Systems; Chihab Nadri, Virginia Tech; Yunmei Liu, University of Louisville; David Kaber, University of Florida; Maryam Zahabi, Texas A&M University; Ryan P. McMahan, Virginia Tech; Sonia Castelo Quispe, New York University; Ryan McKendrick, Northrop Grumman; Jing Qian, New York University; Claudio Silva, New York University
Pilots in modern cockpits face high cognitive demands from complex interfaces and multitasking, often leading to overload and decreased performance. AdaptiveCoPilot, a neuroadaptive guidance system, dynamically adjusts visual, auditory, and textual cues based on cognitive workload measured via functional Near-Infrared Spectroscopy (fNIRS). A formative study (N=3) with expert pilots informed adaptive rules for modality switching and information load adjustments. Evaluated in a VR-simulated cockpit with licensed pilots (N=8), the system showed higher rates of optimal cognitive load for working memory and perception. This highlights the potential of neuroadaptive systems to enhance pilot performance and aviation safety.
Measuring the Impact of Objects'™ Physicalization, Avatar Appearance, and their Consistency on Pick-and-Place Performance in Augmented Reality
Antonin Cheymol, Univ Rennes; INSA Rennes; Inria; Jacob Wallace, Ecole Centrale de Nantes; Juri Yoneyama, Inria; Rebecca Fribourg, Nantes Université; Ecole Centrale Nantes LS2N-PACCE; UMR 6004; F-44000; Jean-Marie Normand, Ecole Centrale de Nantes; Ferran Argelaguet Sanz, Inria; Univ Rennes; CNRS; IRISA
In order to support the future development of efficient and usable AR interactions, this paper investigates the impact of objects' physicalization (object's set composed of (i) virtual, (ii) real, or (iii) a composite mix of real and virtual objects) and hand appearance (hand's appearance displayed as (i) the real hand, (ii) an avatar, or (iii) dynamically adapting to the surrounding objects' physicalization) on the speed performance of a pick-and-place task. Overall, our results provide a rich characterization of different factors influencing different phases of a pick-and-place interaction, which could be employed to improve the design of future AR applications.
Exploring Worker-Drone Interaction in Mixed Reality: Balancing Distraction and Situational Awareness
Woei-Chyi Chang, Purdue University; Lap-Fai Yu, George Mason University; Sogand Hasanzadeh, Purdue University
Mixed-reality (MR) technology has been widely used to simulate high-risk workplaces to minimize safety concerns. However, its use in understanding worker attentional allocation during interactions with drones in future construction environments remains underexplored. This study developed a futuristic bricklaying MR environment, where human-drone interaction was mandatory, to capture participants' naturalistic behaviors (attention, productivity, and distraction) across different interaction levels (no interaction, coexistence, and collaboration). The findings highlight the value of immersive technology in investigating workers' naturalistic behaviors in future construction scenarios where workers and robots must function as teammates.
Session: Evaluation methods (4 - D)
Robotic Characterization of Markerless Hand-Tracking on Meta Quest Pro and Quest 3 Virtual Reality Headsets
Eric Godden, Queen's University; William Kenneth Steedman Mr., Queen's University; Matthew Pan, Queen's University
Markerless hand-tracking has become increasingly common on commercially available virtual and mixed reality headsets to improve the naturalness of interaction and immersivity of virtual environments. However, there has been limited examination of the performance of markerless hand-tracking on commercial head-mounted displays. Here, we propose an evaluation methodology that leverages a robotic manipulator to measure the positional accuracy, jitter, and latency of such systems and provides a standardized characterization framework of markerless hand-tracking. We apply this methodology to evaluate the hand-tracking performance of two recent mixed reality devices from Meta: the Quest Pro and Quest 3.
ViDDAR: Vision Language Model-based Task-Detrimental Content Detection for Augmented Reality
Yanming Xiu, Duke University; Tim Scargill, Duke University; Maria Gorlatova, Duke University
In Augmented Reality (AR), improperly designed or placed virtual content can obstruct key real-world information or provide users misrepresented information, degrading the overall experience. To address this, we introduce ViDDAR (Vision language model-based Task-Detrimental content Detector for Augmented Reality), the first system using Vision Language Models to detect task-detrimental content, including obstruction and information manipulation attacks. ViDDAR is tested on real-world datasets and deployed via a mobile app, achieved 92.15% obstruction detection accuracy with a 533ms latency and 82.46% information manipulation detection accuracy with a 9.62s time cost, demonstrating its effectiveness in detecting task-detrimental AR content.
Investigating Virtual Reality for Alleviating Human-Computer Interaction Fatigue: A Multimodal Assessment and Comparison with Flat Video
Xinyi Wang, Shandong University; Jing Qu, Shandong University; Lingguo Bu, Shandong University; Shantong Zhu, Communication University of China
This study collected electroencephalogram, functional near-infrared spectroscopy, eye movement, and subjective scale data to explore the validity of a multimodal approach in assessing human-computer interaction fatigue. Fatigue significantly increased brain activation, particularly in the Theta band of the prefrontal cortex, reduced effective connectivity between the Alpha and Theta bands, decreased pupil diameter, increased blink frequency, and raised scale scores. During recovery, both the VR and flat video groups viewed the same natural grass scene. The VR group showed higher prefrontal cortex activation in the Alpha and Theta bands, while the video group had stronger Alpha band connectivity.
From PINs to Gestures: Analyzing Knowledge-Based Authentication Schemes for Augmented and Virtual Reality
Naheem Noah, University of Denver; Sanchari Das, George Mason University
As Augmented and Virtual Reality (AR/VR) advances, the need for secure and user-friendly authentication methods grows critical. We present a systematic evaluation of 17 distinct authentication schemes across five categories: gaze-based, gesture-based, PIN-based, spatial, and recognition-based. Our analysis revealed varying performance in effectiveness, security, and usability. Beat-PIN achieved strong security with 140-bit entropy, while RubikAuth excelled in usability with authentication times of 1.69 seconds. Users preferred balanced schemes like In-Air-Handwriting and Things. By applying Bonneau et al.'s framework, we identified promising schemes and future research directions for AR/VR authentication.
The Reliability and Validity of the Avatar Embodiment Questionnaire
Tabitha C. Peck, Davidson College; Pauline W Cha, Davidson College
We assess the reliability of the Avatar Embodiment Questionnaire (aeQ) proposed by Peck and Gonzalez-Franco through 2-5 repeated exposures, each approximately two-weeks apart. Participants were embodied in same or cross-gendered avatars and results were assessed using three common reliability analyses. For participants in same-gendered avatars, the aeQ demonstrated acceptable or better reliability. For participants in cross-gendered avatars, the aeQ demonstrated questionable reliability for the initial trial and acceptable to excellent reliability thereafter. Results support that the aeQ is a reliable questionnaire, however minor experimental design changes should be considered when using avatars that are diversely different from the user.
Session: Navigation and Locomotion Strategies (5 - A)
Multimodal Turn in Place: A Comparative Analysis of Visual and Auditory Reset UIs in Redirected Walking
Ho Jung Lee, Yonsei University; Hyunjeong Kim, Yonsei University; In-Kwon Lee, Yonsei University
Resetting in redirected walking (RDW) ensures collision-free walking in VR in limited spaces, but frequent resets hinder immersion. Research has proposed visual reset UIs to effectively guide users, but their effectiveness remains underexplored. Expert interviews reveal that users sometimes fail to detect reset UIs in a timely manner. To address this, a novel gauge-based visual reset UI is proposed to improve user navigation and experience quality. Study 1 validates its effectiveness by comparing it to existing UIs such as Direction, End Point, and Arrow Alignment. This work also introduces a multimodal reset paradigm that integrates visual and auditory sensory cues for a comprehensive reset solution.
Techniques for Multiple Room Connection in Virtual Reality: Walking Within Small Physical Spaces
Ana Rita Rebelo, NOVA LINCS; NOVA School of Science and Technology Pedro A. Ferreira, NOVA LINCS; NOVA School of Science and Technology Rui Nóbrega, NOVA LINCS; NOVA School of Science and Technology
This paper presents three room-connection techniques, portals, corridors, and central hubs, that can be used in virtual environments (VEs) to create 'impossible spaces'. These spaces use overlapping areas to maximize available physical space, promising for natural walking even in constrained spaces. We conducted a user study with 33 participants to assess the effectiveness of these techniques within a small physical area (2.5 x 2.5 m). The results show that all three techniques are viable for connecting rooms in VR, each offering distinct characteristics. Each method positively impacts presence, cybersickness, spatial awareness, orientation, and overall user experience. Specifically, portals offer a flexible and straightforward solution, corridors provide a seamless and natural transition between spaces, and central hubs simplify navigation. The primary contribution of this work is demonstrating how these room-connection techniques can be applied to dynamically adapt VEs to fit small, uncluttered physical spaces, such as those commonly available to VR users at home. Applications such as virtual museum tours, training simulations, and emergency preparedness exercises can greatly benefit from these methods, providing users with a more natural and engaging experience, even within the limited space typical in home settings.
F-RDW: Redirected Walking With Forecasting Future Position
Sang-Bin Jeon
In order to serve better VR experiences to users, existing predictive methods of Redirected Walking (RDW) exploit future information to reduce the number of reset occurrences. However, such methods often impose a precondition during deployment, either in the virtual environment's layout or the user's walking direction, which constrains its universal applications. To tackle this challenge, we propose a novel mechanism F-RDW that is twofold: (1) forecasts the future information of a user in the virtual space without any assumptions, and (2) fuse this information while maneuvering existing RDW methods. The backbone of the first step is an LSTM-based model that ingests the user's spatial and eye-tracking data to predict the user's future position in the virtual space, and the following step feeds those predicted values into existing RDW methods (such as MPCRed, S2C, TAPF, and ARC) while respecting their internal mechanism in applicable this http URL results of our simulation test and user study demonstrate the significance of future information when using RDW in small physical spaces or complex environments. We prove that the proposed mechanism significantly reduces the number of resets and increases the traveled distance between resets, hence augmenting the redirection performance of all RDW methods explored in this work.
Follow Me: Confirmation-based Group Navigation in Collocated Virtual Reality
Tony Jan Zoeppig, Bauhaus-Universität Weimar; Anton Benjamin Lammert, Bauhaus-Universität Weimar; Ephraim Schott, Bauhaus-Universität Weimar; Sebastian Mühlhaus, Bauhaus-Universität Weimar; Bernd Froehlich, Bauhaus-Universität Weimar
In collocated social virtual reality, spatial synchronization reduces collision risks but limits user autonomy during navigation of large virtual worlds. We introduce a two-phase, confirmation-based group navigation concept allowing users to stay behind and rejoin their group later, using confirmation mechanisms triggered by either the navigator, the passenger, or both. To prevent physical collisions between spatially desynchronized users, we utilize ghost avatars. Our study (N = 24) shows that users valued the autonomy provided by our approach during a guided tour. Furthermore, all techniques achieved comparable levels of co-presence, spatial awareness, and collision avoidance as a baseline group navigation technique.
Can I Get There? Negotiated User-to-User Teleportations in Social VR
Miao Wang, Beihang University; Wentong Shu, Beihang University; Yi-Jun Li, Tsinghua University; Wanwan Li, University of South Florida
The growing adoption of social virtual reality platforms underscores the importance of safeguarding personal VR space to maintain user privacy and security. Teleportation facilitates user engagement but can also inadvertently intrude upon personal VR space. This paper introduces three innovative negotiated teleportation techniques designed to secure user-to-user teleportation and protect personal space privacy, all under a unified small-group development framework. We conducted a user study with 20 participants who performed social tasks within a virtual campus environment. The findings demonstrate that our techniques significantly enhance privacy protection and alleviate anxiety associated with unwanted proximity in social VR.
Session: Modeling and simulation (5 - B)
Editable Mesh Animations Modeling Based on Controlable Particles for Real-Time XR
Xiangyang Zhou, University of Science and Technology Beijing; Yanrui Xu, University of Science and Technology Beijing; Chao Yao, University of Science and Technology Beijing; Xiaokun Wang, University of Science and Technology Beijing; Xiaojuan Ban, University of Science and Technology Beijing
The real-time generation of editable mesh animations in XR applications is challenging due to slow generation speeds and difficulty controlling complex shapes. We propose a universal particle control scheme based on target sampling features to address these issues. This includes a spatially adaptive control algorithm that adjusts forces according to model features, eliminating parameter dependency. Additionally, we implement boundary correction techniques and a distance-adaptive fragmentation mechanism. Experimental results show improved performance in controlling complex structures, enhancing accuracy and efficiency in generating multiple targets.
A Cost-effective Tangible Rehearsal Interface for Microsurgical Clipping of Intracranial Aneurysm
Wei Cao, China University of Petroleum; Zhengyang Li, China University of Petroleum; Yuchen He, China University of Petroleum; Zehua Liu, The Chinese University of Hong Kong; Meng Zhang, Shenzhen Secand People's Hospital; Lv Jianping, Guangzhou First People's Hospital; Weixin Si, Shenzhen Institute of Advanced Technology; Chinese Academy of Sciences
This work presents a mixed reality based microsurgical clipping rehearsal system for intracranial aneurysm treatment. To ensure realistic tangible feedback from the clip holder, we design a marker-based tracking method to adopt the real microsurgical instrument as the input device. Besides, a specific-to-general stitching method is proposed to register the patient-specific aneurysm lesion to a standard brain for personalized model reconstruction. In addition, to well simulate the interactions between the aneurysm and the clip, we design a collision distance constrained contact mechanics to enhance the efficiency and stability of simulation. Experiments demonstrated our system had the potential to be applied in real clinical practice.
Blind Augmentation: Calibration-free Camera Distortion Model Estimation for Real-time Mixed-reality Consistency
Siddhant Prakash, University College London; David R Walton, Birmingham City University; Rafael Kuffner dos Anjos, University of Leeds; Anthony Steed, University College London; Tobias Ritschel, University College London
Real camera footage is subject to noise, motion blur (MB) and depth of field (DoF). In some applications these might be considered distortions to be removed, but in others it is important to model them because it would be ineffective to simply remove them. We propose a method which estimates parameters for noise, MB and DoF instantly, allowing use of off-the-shelf real-time simulation methods in compositing augmented content. Our main idea is to unlock both features by showing how to use modern computer vision methods that can remove noise, MB and DoF from the video stream, essentially providing self-calibration. This allows to auto-tune any black-box real-time noise+MB+DoF method to deliver fast and high-fidelity augmentation consistency.
Look at the Sky: Sky-aware Efficient 3D Gaussian Splatting in the Wild
Yuze Wang, Beihang University; Junyi Wang, Shandong University; Ruicheng Gao, Beihang University; Yansong Qu, XMU; Wantong Duan, Beihang University; Shuo Yang, Beihang University; Yue Qi, Beihang University
Photographs taken in unconstrained tourist environments often present challenges for accurate scene reconstruction due to variable appearances and transient occlusions. With the advancement of 3DGS, some methods have enabled real-time rendering by reconstructing scenes from unconstrained photo collections. However, the rapid convergence of 3DGS is misaligned with the slower convergence of neural network-based appearance encoder and transient mask predictor, hindering the reconstruction efficiency. To address this, we propose a novel sky-aware framework for in-the-wild scene reconstruction. Extensive experiments on multiple datasets demonstrate that the proposed framework outperforms existing methods in novel view and appearance synthesis.
EventPointMesh: Human Mesh Recovery Solely From Event Point Clouds
Ryosuke Hori
Session: Collaborative interactions (5 - C)
Environment Spatial Restitution for Remote Physical AR Collaboration
Bruno Caby, Arts et Metiers Institute of Technology; Guillaume Bataille, Orange; Florence Danglade, Arts et Metiers ParisTech; Jean-Rémy Chardonnet, Arts et Metiers Institute of Technology
This paper presents a remote physical collaborative AR system designed to enhance remote collaboration through augmented physical tasks. This system features verbal and non-verbal communication tools, allowing users to embody avatars and interact via hand, head and eye tracking. It captures and renders real-world environments remotely in near real time, achieving latency of less than 250ms. We present a cascading user study comparing our system with a commercial 2D video AR application, showing improved empathy between collaborators, increased cognitive load and slightly reduced usability. Our findings highlight the potential of RPC-AR to improve the effectiveness and usability of remote collaboration.
Do We Still Need Human Instructors? Investigating Automated Methods for Motor Skill Learning in Virtual Co-embodiment
Haruto Takita, The University of Tokyo; Kenta Hashiura, The University of Tokyo; Yuji Hatada, The University of Tokyo; Daiki Kodama, Kuzuoka Amemiya Narumi Lab; Takuji Narumi, the University of Tokyo; Tomohiro Tanikawa, the University of Tokyo; Michitaka Hirose, The University of Tokyo
Virtual reality enhances motor skill acquisition by enabling unique physical activities. Co-embodiment learning, where instructors and learners share a single avatar blending their movements, has proven effective but requires real-time participation from skilled instructors, limiting its accessibility. This study addresses the issue by developing an AI instructor trained on human motor data using supervised learning. Performance evaluation compared the AI instructor, recorded human data, a human instructor, and solo learning. Results showed that the AI instructor significantly improved learning efficiency over solo or recorded data practice, achieving comparable outcomes to learning with a human instructor.
Just stop doing everything for now!'': Understanding security attacks in remote collaborative mixed-reality
Maha Sajid, Virginia Tech; Syed Ibrahim Mustafa Shah Bukhari, Virginia Tech; Bo Ji, Virginia Tech; Brendan David-John, Virginia Tech
The unique immersive features of Mixed Reality(MR) devices introduce new vulnerabilities leading to interaction obstruction and misdirection. We implemented four types of attacks within a remote collaborative MR platform and evaluated user behavior and mitigations through a user study. We compared responses to MR-specific attacks, which exploit the unique characteristics of remote collaborative immersive environments, and traditional security attacks implemented in MR. Our findings indicate that users generally exhibit lower recognition rates for immersive attacks. Our results demonstrate a clear gap in user awareness and responses and emphasize the importance of enhanced security measures to maintain trust in such systems.
Investigating the Influence of Playback Interactivity during Guided Tours for Asynchronous Collaboration in Virtual Reality
Alexander Giovannelli, Virginia Tech; Leonardo Pavanatto, Virginia Tech; Shakiba Davari, Virginia Tech; Haichao Miao, Lawrence Livermore National Laboratory; Vuthea Chheang, Lawrence Livermore National Laboratory; Brian Giera, Lawrence Livermore National Laboratory; Peer-Timo Bremer, Lawrence Livermore National Laboratory; Doug Bowman, Virginia Tech
While collaborative virtual environments have been investigated for working in different spaces at the same time, few have investigated the best practices for collaborating in those spaces at different times. We designed a system that allows experts to create engaging tours inside a virtual inspection space, preserving knowledge and insights while requiring observer interaction to follow the tour's playback. A user study evaluating the influence of these interactions on an observing user's information recall and user experience indicated that additional interactivity can improve auditory and spatial recall of key information conveyed during the tour and enhance the user experience.
TeamPortal: Exploring Virtual Reality Collaboration Through Shared and Manipulating Parallel Views
Xian Wang, The Hong Kong Polytechnic University; Luyao Shen, Computational Media and Arts Thrust; The Hong Kong University of Science and Technology (Guangzhou); Lei Chen, Hebei GEO University; Mingming Fan, The Hong Kong University of Science and Technology (Guangzhou); Lik-Hang Lee, Hong Kong Polytechnic University
Parallel views in Virtual Reality (VR) can enhance collaborative efficiency by enabling users to observe and interact with partners' perspectives. We introduce TeamPortal, a shared view system, and present two user studies to investigate the effects of interactive, shared perspectives. Study 1 compares ShaView and TeamPortal with a baseline in collaborative tasks, showing that TeamPortal significantly reduced movement and increased efficiency and social presence in complex tasks. Study 2 explores three enhanced variants (TeamPortal+, SnapTeamPortal+, DropTeamPortal+), the later two improved task efficiency and acceptance, although SnapTeamPortal+ reduced co-presence. We proposed three design implications for future VR collaboration systems.
Session: Input devices (5 - D)
CoordAuth: Hands-Free Two-Factor Authentication in Virtual Reality Leveraging Head-Eye Coordination
Sheng Zhao, Tsinghua University; Junrui Zhu, Tsinghua University; Shuning Zhang, Tsinghua University; Xueyang Wang, Tsinghua University; Hongyi Li, Tsinghua University; Fang Yi, Tsinghua University; Xin Yi, Tsinghua University; Hewu Li, Tsinghua University
We present CoordAuth, a gaze-based two-factor authentication technique in VR utilizing implicit head-eye motion features to offer a more secure and natural alternative to traditional pattern-based authentication. Users authenticate by performing gestures across a 3x3 grid with their eyes. We optimized CoordAuth's UI by evaluating input performance across different grid sizes and constructed its authentication algorithm using head-eye coordination features and Random Forest classifiers. CoordAuth showed strong security with a 1.6% FAR and 1.5% FRR across 24 users and a 0.6% Attack Success Rate against shoulder-surfing. Usability study in various postures revealed faster authentication and lower rejection rates compared to baselines.
A Dual-Stick Controller for Enhancing Raycasting Interactions with Virtual Objects
Nianlong Li, Lenovo Research; Tong Wu, Human-Computer Interaction Technology and Intelligent Information Processing Laboratory; Zhenxuan He, Institute of software; Chinese Academy of Sciences; Luyao Shen, Computational Media and Arts Thrust; The Hong Kong University of Science and Technology (Guangzhou); Tianren Luo, Institute of Software; Teng Han, Institute of Software; Chinese Academy of Sciences; BoYu Gao, Jinan University; Yu Zhang, Lenovo Research; Liuxin Zhang, Lenovo Research; Feng Tian, Institute of software; Chinese Academy of Sciences; Qianying Wang, Lenovo Research
This work presents Dual-Stick, a novel controller with two sticks connected at the end, introducing the Dual-Ray interaction paradigm to enhance raycasting input in Virtual Reality. By leveraging human dexterity with tools like clamps and tweezers, Dual-Stick adjusts the relative angle between sticks, supporting flexible manipulation with additional degrees of freedom. Two studies were conducted to assess Dual-Ray's effectiveness in target selection and manipulation tasks. Results showed that Dual-Ray improved selection efficiency over single-ray input but did not outperform enhanced single-ray techniques. For manipulation, Dual-Ray reduced completion time and mode switching compared to single-ray input.
Omnidirectional VR Treadmills Walking Techniques: Comparing Walking-in-Place and Sliding vs Natural Walking
Helia Homami, Dalhousie University; Adria Quigley, Dalhousie University; Mayra Donaji Barrera Machuca, Dalhousie University
In Virtual Reality (VR), navigating large environments remains challenging due to space constraints and motion sickness from software-based methods. This study compares natural walking, walking-in-place, and sliding in terms of user performance and experience using VR omnidirectional treadmills (VR-ODTs). Eighteen participants completed a maze task using each method. Results showed that natural walking led to the fastest completion and highest satisfaction, sliding balanced effort and immersion, and walking-in-place required the most exertion. These findings provide insights for selecting VR locomotion methods for training and rehabilitation applications.
The Impact of Visual and Haptic Feedback on Keyboard Typing in Immersive Virtual Environments
Amine Chellali, Université d'Evry Paris Saclay; Lucas Herfort, Université d'Evry Paris-Saclay; Guillaume Loup, Université d'Evry Paris-Saclay; Marie-helene Ferrer, The French Armed Forces Biomedical Research Institute
We investigate the effects of two visual feedback (hover and keypress) and three haptic feedback mechanisms (physical keyboard, physical surface, and a mid-air virtual keyboard) and the possible interactions between them on typing using the two index fingers in VR. Results show that keypress visual feedback enhanced typing speed and reduced workload, while hover feedback lowered error rates but negatively impacted typing speed. Additionally, using a physical keyboard with passive haptic feedback increased the error rate. Regarding eye gaze behavior, participants spent more time looking at the keyboard with the keypress visual feedback and when no haptic feedback was provided.
AirtypeLogger: How Short Keystrokes in Virtual Space Can Expose Your Semantic Input to Nearby Cameras
Tongyu Zhang, Shandong University; Yiran Shen, Shandong University; Ning Chen, Shandong University; guoming zhang, Shandong University; Yuanfeng Zhou, Shandong University
We present AirtypeLogger, a video-based attack on the air-typing activities of XR users. AirtypeLogger considers a scenario in which the users are typing a short text fragment with semantic meaning under the spy of video cameras. It detects and localizes the air-typing events in video streams and encodes the keystrokes' relative positions and temporal order. Then, high-precision inference can be achieved by applying a Transformer-based network to the spatial and temporal encodings of the keystroke sequences. Finally, AirtypeLogger can achieve a Character Error Rate of less than 0.1 as long as 7 air-typing events are observed.
Session: Spatial Experiences and AI-Driven Content Generation (6 - A)
Semantics-Controlled Gaussian Splatting for Outdoor Scene Reconstruction and Rendering in Virtual Reality
Hannah Schieber, Technical University of Munich (TUM); Jacob Young, University of Otago; Tobias Langlotz, University of Otago; Stefanie Zollmann, University of Otago; Daniel Roth, Technical University of Munich
We propose Semantics-Controlled GS (SCGS) enabling the separation of large scene parts outdoor scenes for VR. Additionally, we introduce a novel challenging outdoor dataset, overcoming the ''circling' setups. We outperform the state-of-the-art in visual quality on our dataset and in segmentation quality on the 3D-OVS dataset. We conducted an exploratory user study, comparing a 360-video, plain GS, and SCGS in VR with a fixed viewpoint. In our subsequent main study, users were allowed to move freely, evaluating plain GS and SCGS. Our main study results show that participants clearly prefer SCGS. We overall present an innovative approach that surpasses the state-of-the-art both technically and in user experience.
MineVRA: Exploring the Role of Generative AI-Driven Content Development in XR Environments through a Context-Aware Approach
Lorenzo Stacchio, University of Macerata; Emanuele Balloni, Universita Politecnica delle Marche; Marina Paolanti, University of Macerata; Emanuele Frontoni, University of Macerata; Primo Zingaretti, UNIVPM; Roberto Pierdicca, Universita Politecnica delle MArche
The convergence of AI, Computer Vision, Graphics, and Extended Reality (XR) drives innovation in immersive environments. A key challenge is creating personalized 3D assets, which is traditionally a manual, time-consuming process. Generative AI (GenAI) offers a promising solution for automated, context-aware content creation. This paper introduces MineVRA, a Human-In-The-Loop XR framework integrating GenAI for adaptive 3D asset generation. A user study compared GenAI-generated objects to Sketchfab assets in immersive contexts. Results highlight GenAI's potential to complement traditional libraries, offering design insights for human-centered XR environments and advancing efficient, personalized 3D content creation.
LLMER: Crafting Interactive Extended Reality Worlds with JSON Data Generated by Large Language Models
Jiangong Chen, Pennsylvania State University; Xiaoyi Wu, Pennsylvania State University; Tian Lan, George Washington University; Bin Li, Pennsylvania State University
In this paper, we present LLMER, a novel framework that leverages JSON data generated by large language models (LLMs) to create interactive Extended Reality (XR) environments. LLMER enables users to interact with a virtual agent via natural language, seamlessly generating virtual objects and animations. The framework employs a multi-stage strategy to provide only the most relevant contextual information tailored to the user's requests and integrates multiple modules to address diverse XR tasks. Our preliminary user study demonstrates the effectiveness of LLMER, achieving an over 80% reduction in token consumption and approximately 60% decrease in task completion time compared to state-of-the-art methods.
Navigating Decision-Making Frontiers: Virtual Reality and Spatial Skills in Strategic Planning
Jerson G Geraldo Neto, Brazilian Army; Anderson Maciel, Universidade de Lisboa; Luciana Nedel, Federal University of Rio Grande do Sul (UFRGS)
Navigating topographic charts requires nuanced skills and real-world correspondence, shaped by individual learning styles. Traditional 2D and 3D representations help bridge map-reality gaps but often rely on manual methods prone to errors. This study explores VR's potential in military planning, focusing on terrain interpretation, data visualization, and scale transitions. With 36 army cadets, results show VR enhances spatial perception, improves position choices and grades, and reduces disparities, especially aiding users with lower spatial skills. The research also examines the impact of scale on VR planning, offering insights into its benefits and challenges. Future studies should address control issues and task completion times.
MagicScroll: Enhancing Immersive Storytelling with Controllable Scroll Image Generation
Bingyuan Wang, The Hong Kong University of Science and Technology (Guangzhou); Hengyu Meng, The Hong Kong University of Science and Technology (Guangzhou); Rui Cao, Xiamen University Malaysia; Zeyu Cai, HKUST(GZ); Lanjiong Li, The Hongkong University of Science and Technology(Guangzhou)); Yue Ma, HKUST; Qifeng Chen, HKUST; Zeyu Wang, The Hong Kong University of Science and Technology (Guangzhou)
Scroll images are a unique medium commonly used in virtual reality (VR) providing an immersive visual storytelling experience. This paper proposes a multi-layered, diffusion-based scroll image generation framework with a novel semantic-aware denoising process. We incorporate layout prediction and style control modules to generate coherent scroll images of any aspect ratio. We use different multi-window strategies to render diverse visual forms such as chains, rings, and forks for VR storytelling. Quantitative and qualitative evaluations demonstrate that our techniques can significantly enhance text-image consistency and visual coherence in scroll image generation, as well as the level of immersion and engagement of VR storytelling.
Session: Virtual Agents, Embodiment, and Social Presence (6 - B)
EmBARDiment: an Embodied AI Agent for Productivity in XR
Riccardo Bovo, Imperial College London; Steven Abreu, University of Groningen; Karan Ahuja, Northwestern University; Eric J Gonzalez, Google; Li-Te Cheng, Google; Mar Gonzalez-Franco, Google
XR devices running chat-bots powered by Large Language Models (LLMs) have the to become always-on agents that enable much better productivity scenarios. Current screen based chat-bots do not take advantage of the the full-suite of natural inputs available in XR, including inward facing sensor data, instead they over-rely on explicit voice or text prompts, sometimes paired with multi-modal data dropped as part of the query. We propose a solution that leverages an attention framework that derives context implicitly from user actions, eye-gaze, and contextual memory within the XR environment. Our work minimizes the need for engineered explicit prompts, fostering grounded and intuitive interactions that glean user insights for the chat-bot.
'Age Isn't Just a Number': Effects of Virtual Human Age and Gender on Persuasion, Social Presence and Influence in Interpersonal Social Encounters in VR
Elizabeth A Schlesener, Clemson University; Vyomakesh Shivakumar, Clemson University; Dave Breeze, BMW Group; Brooke Rennison, BMW Group; Burak Soehmelioglu, BMW Group; Sabarish V. Babu, Clemson University
This study examines the impact of user and virtual human (VH) age and gender on persuasion and social influence in virtual reality (VR), focusing on factors such as brand attitude, interpersonal attraction, source credibility, and presence. Using a 3 (VH age) ž 2 (VH gender) ž 3 (module topic) multi-factorial design (N=45), results reveal that VH age and gender significantly affect their persuasiveness and social impact, which vary depending on the user's age and gender. Our findings provide novel insights for designing effective virtual sales strategies and tailoring persuasive technologies to diverse user profiles, thus filling a vital gap in understanding the interaction between user and agent characteristics in virtual environments.
Coverage of Facial Expressions and Its Effects on Avatar Embodiment, Self-Identification, and Uncanniness
Peter Kullmann, University of Würzburg; Theresa Schell, JuLius-Maximilians-Universität Würzburg; Timo Menzel, TU Dortmund University; Mario Botsch, TU Dortmund University; Marc Erich Latoschik, University of Würzburg
Facial expressions are crucial for many eXtended Reality (XR) use cases. We investigate effects of the coverage of facial animations on important factors of self-perception. We exposed 83 participants to their mirrored personalized avatar. They were shown their mirrored avatar face with upper and lower face animation, upper face animation only, lower face animation only, or no face animation. Whole animations were rated higher in embodiment and slightly lower in uncanniness. Partial animations were similar to no animation in embodiment, but less eerie, less human, and less attractive. Self-identification effects were mixed. Qualitative results show participants perceived their virtual representation as fascinating yet uncanny.
Trust in Virtual Agents: Exploring the Role of Stylization and Voice
Advancements in AI have improved the realism of virtual agents. However, little research explores how data-driven methods combined with voice cues affect user perceptions. This study reconstructs stylized agents with synthesized voices to examine their impacts on trust and other perceptions. An experiment with 27 participants showed that stylized agents build trust similarly to realistic ones, while voice has a negligible effect. Elder agents were more trusted. The style of the agents also plays a key role in participants' perceived realism, and audio-visual matching significantly enhances perceived empathy.
X's Day: Personality-Driven Virtual Human Behavior Generation
Haoyang Li, Beijing Institute of Technology; Zan Wang, Beijing Institute of Technology; Wei Liang, Beijing Institute of Technology; Yizhuo Wang, Beijing Institute of Technology
This paper introduces a novel task focused on generating long-term behaviors for virtual agents, guided by specific personality traits and contextual elements within 3D environments. We present a comprehensive framework capable of autonomously producing daily activities autoregressively. By modeling the intricate connections between personality characteristics and observable activities, we establish a hierarchical structure of Needs, Task, and Activity levels. Integrating a Behavior Planner and a World State module allows for the dynamic sampling of behaviors using large language models (LLMs), ensuring that generated activities remain relevant and responsive to environmental changes. Project website: https://behavior.agent-x.cn/.
Session: Exploring Perception, Haptics, and Learning in XR (6 - C)
Explainable XR: Understanding User Behaviors of XR Environments using LLM-assisted Analytics Framework
Yoonsang Kim, Stony Brook University; Zainab Aamir, Stony Brook University; Mithilesh Kumar Singh, Stony Brook University; Saeed Boorboor, Stony Brook University; Klaus Mueller, Stony Brook University; Arie Kaufman, Stony Brook University
We introduce Explainable XR, a novel framework for analyzing user behavior across varied eXtended Reality (XR) environments using Large Language Models for data interpretation. Addressing the challenges of cross-virtuality transitions and multimodal data, Explainable XR offers a virtuality-agnostic solution for data collection, analysis, and visualization. It features a User Action Descriptor for capturing multimodal actions and intents, a platform-independent XR session recorder, and a visual analytics interface coupled with LLM-assisted insights. Our evaluation showcases its effectiveness in both individual and collaborative XR applications, providing valuable, actionable insights into user behaviors in immersive environments.
Hit Around: Substitutional Moving Robot for Immersive and Exertion Interaction with Encountered-Type Haptic
Yu-Hsiang Weng, National Tapei University of Technology; Ping-Hsuan Han, National Taipei University of Technology; Kuan Ning Chang, National Taipei University of Technology; Chi-Yu Lin, National Taipei University of Technology; Chia-Hui Lin, National Taipei University of Technology; Ho Yin Ng, National Taipei University of Technology; Chien-Hsing Chou, National Taipei University of Education; Wen-Hsin Chiu, National Tsing Hua University
Previous works have highlighted immersive technologies' potential to enhance physical activities. However, existing encountered-type haptic feedback systems have yet to extensively explore handling intense human forces and simultaneous user movement, limiting their application in combat sports. We introduce Hit Around, a substitutional moving robot enabling users to punch virtual opponents while receiving realistic haptic feedback. Through iterative prototyping and insights from exhibitions, we designed the system, conducted technical evaluations, and validated its feasibility in a laboratory experiment. A field study explored the limitations and challenges of developing immersive exertion interactions with encountered-type haptics.
Mentor-Guided Learning in Immersive Virtual Environments: The Impact of Visual and Haptic Feedback on Skill Acquisition
Flavien Lebrun, Université d'Evry Paris-Saclay; Cassandre Simon, Université d'Evry Paris-Saclay; Assia Boukezzi, Université d'Evry Paris-Saclay; Samir Otmane, Université d'Evry Paris-Saclay; Amine Chellali, Université d'Evry Paris-Saclay
We present a study in which two experimental groups received visual and/or haptic-augmented feedback from an expert, while a control group received no feedback. Results indicate that both experimental groups showed significant performance improvement, with no significant difference between them. Analysis of their learning curves revealed similar performance improvements, outperforming the control group. The visual-haptic feedback was linked to a lower workload and a higher perceived interdependence with the expert. This suggests that augmented feedback enhances the learning of tool manipulation skills. Although adding haptic feedback did not lead to better learning outcomes, it did improve the user experience.
Unified Approach to Mesh Saliency: Evaluating Textured and Non-Textured Meshes Through VR and Multifunctional Prediction
Kaiwei Zhang, Shanghai Jiao Tong University; Dandan Zhu, East China Normal University; Xiongkuo Min, Shanghai Jiao Tong University; Guangtao Zhai, Shanghai Jiao Tong University
Mesh saliency enhances AI adaptability by identifying regions that naturally attract visual attention. While existing methods focus on geometric shapes, capturing the interplay between geometry and texture patterns remains challenging. Therefore, we introduce a comprehensive dataset with saliency distributions for identical 3D models under non-textured and textured conditions. Additionally, we propose a unified saliency prediction model that integrates geometric and texture features within a coherent topological framework. Extensive validation highlights the model's scalability, generalizability, and potential to enhance modeling and rendering while uncovering insights into 3D visual feature interaction.
Impact of Visual Virtual Scene and Localization Task on Auditory Distance Perception in Virtual Reality
Sarah Roákopf, University of Regensburg; Andreas Mühlberger, University of Regensburg; Felix Stärz, Jade Hochschule; Matthias Blau, Jade Hochschule; Steven van de Par, Carl-von-Ossietzky-University; Leon O. H. Kroczek, University of Regensburg
Virtual reality (VR) enables controlled study of auditory perception. However, visual information affects sound processing due to multimodal integration, and head-mounted displays compress egocentric distances. We investigated how different visual scenes affect auditory distance perception in VR using two paradigms. First, participants localized sounds in audiovisually congruent or incongruent rooms. Second, we manipulated room visibility and localization tasks. Audiovisual incongruency impaired distance perception, while room visibility had no main effect but interacted with the task. Also, differential effects on quality of experience were found, highlighting the need to consider the complex interplay of acoustic and visual scenes in VR.
Session: Next-Generation Display Technologies and Visual Processing (6 - D)
Slim Diffractive Waveguide Glasses for Beaming Displays with Enhanced Head Orientation Tolerance
Yuta Itoh, The University of Tokyo; Tomoya Nakamura, Osaka University; Yuichi Hiroi, Cluster Metaverse Lab; Kaan Ak?it, University College London
Augmented Reality (AR) glasses require slim, lightweight, and energy-efficient designs for widespread adoption. Beaming Displays offer a solution by shifting active components to the environment, leaving only passive optics in the wearable device. Current designs, however, face challenges in balancing compactness and wide-angle projection tolerance. This study introduces light-receiving glasses with a diffractive waveguide incorporating in-coupling and out-coupling gratings. Our design achieves a wider incident angle range (20-30 degrees) compared to the previous limit of 5 degrees, while maintaining a compact form factor. A proof-of-concept prototype validates the approach, addressing key limitations of existing AR glasses.
LAPIG: Language Guided Projector Image Generation with Surface Adaptation and Stylization
Yuchen Deng, Southwest University; Haibin Ling, Stony Brook University; Bingyao Huang, Southwest University
We propose LAPIG, a language guided projector image generation method with surface adaptation and stylization. LAPIG consists of a projector-camera system and a target textured projection surface. LAPIG takes the user text prompt as input and aims to transform the surface style using the projector. LAPIG's key challenge is to generate a projector image that follows the user's instruction while also displays minimum surface artifacts. To address this, we propose projection surface adaptation (PSA) that can generate compensable surface stylization. First, we simulate projector compensation and project-and-capture processes. Then, we design content and saturation losses to guide the projector image generation for pleasing projection effects.
Privacy-Enabled Parallax Display
Sizhuo Ma, Snap Inc.; Karl Bayer, Snap Inc.; Gurunandan Krishnan, Snap Inc.; Mohit Gupta, University of Wisconsin-Madison; Shree Nayar, Columbia University
Privacy filters aim to restrict visual content to intended viewers but often suffer from either wide viewing angles or limited user positioning. We propose a privacy-enabled parallax display with a narrow, adaptive viewing field directed to arbitrary user locations. Unlike conventional parallax barriers, which suffer from repeated views due to periodic patterns, our approach randomizes barrier designs in both 1D and 2D, enhancing privacy along both horizontal and vertical directions. Using a prototype built from off-the-shelf LCDs, we demonstrate a steerable viewing field of ñ6°, offering better privacy protection on high-frequency contents compared to traditional screens.
Frequency-aware Uncertainty Gaussian Splatting for Dynamic Scene Reconstruction
Mingwen Shao, China University of Petroleum (East China); Yuanjian Qiao, China University of Petroleum (East China); Kai Zhang, China University of Petroleum (East China); Lingzhuang Meng, China University of Petroleum (East China)
In this paper, we propose a novel Frequency-aware Uncertainty Gaussian Splatting, termed FUGS, for adaptively reconstructing dynamic scenes in the Fourier space. Specifically, we design an Uncertainty-aware Deformation Model (UDM) that explicitly models motion attributes using learnable uncertainty relations with neighboring Gaussian points. Subsequently, a Dynamic Spectrum Regularization (DSR) is developed to perform coarse-to-fine Gaussian densification through low-to-high frequency filtering. Extensive experiments on synthetic and real-world datasets show that our FUGS exhibits significant superiority over state-of-the-art methods.
Session: 3D user interfaces (7 - A)
Beyond the Phone: Exploring Phone-XR Integration through Multi-View Transitions for Real-World Applications
Fengyuan Zhu, University of Toronto; Xun Qian, Google; Daniel Kalmar, Google; Mahdi Tayarani, Google; Eric J Gonzalez, Google; Mar Gonzalez-Franco, Google; David Kim, Google Research; Ruofei Du, Google
Although XR headsets have become increasingly prevalent, their integration with mobile phones remains limited. Existing approaches typically replicate phone interfaces in XR or employ the phone as a 6DoF controller. We propose a framework that enables seamless transitions between mirrored, magnified, and augmented views of mobile interfaces in XR. Our system dynamically adapts these interfaces to application content, creating a generalizable solution for phone's XR integration. Guided by literature reviews and expert workshops, we established a design space and developed a prototype evaluated in real-world applications. A user study evaluated the system, underscoring its potential to foster a cohesive cross-device ecosystem.
Order Up! Multimodal Interaction Techniques for Notifications in Augmented Reality
Lucas Plabst, Colorado State University; Florian Niebling, TH Köln ? University of Applied Sciences; Sebastian Oberdörfer, University of Würzburg; Francisco Raul Ortega, Colorado State University
As augmented reality (AR) becomes integral to professional and social settings, managing frequent notifications effectively is crucial. With users receiving nearly 200 notifications daily on smartphones, translating this interaction to AR requires seamless integration with real-world behaviors. This study examines multimodal interaction with notifications during a virtual cooking task, where participants managed customer orders. We tested Point and Pinch, Gaze and Pinch, Point and Voice, Gaze and Voice, and Touch interaction techniques, analyzing their impact on workload, usability, and performance. The findings highlight challenges in multimodal interaction and provide recommendations for optimizing notification management in AR systems.
SeamlessVR: Bridging the Immersive to Non-Immersive Visualization Divide
Shuqi Liao, Purdue University; Sparsh Chaudhri, Purdue University; Maanas K Karwa, Purdue University; Voicu Popescu, Purdue University
SeamlessVR is a method for smoothly transitioning from immersive VR headset visualization to non-immersive on-screen views. It continuously morphs 3D visualization into 2D, matching what users see after removing the headset. This seamless transition reduces cognitive effort, allowing users to continue visualization tasks effortlessly on screen. In an IRB-approved study with 30 participants, SeamlessVR outperformed the conventional headset removal approach in time and accuracy for target tracking in complex abstract and realistic scenes. Additionally, users reported better perception of the transition and enhanced usability. SeamlessVR did not induce cybersickness.
Evaluating 3D Visual Comparison Techniques for Change Detection in Virtual Reality
Although recent advances in algorithmic methods in Change detection (CD), CD still requires human intervention. Virtual Reality (VR) can improve visual comparison through 3D interaction and stereoscopic depth, enhancing spatial reasoning.
FocalSelect: Improving Occluded Objects Acquisition with Heuristic Selection and Disambiguation in Virtual Reality
Duotun Wang, The Hong Kong University of Science and Technology (Guangzhou); Linjie Qiu, The Hong Kong University of Science and Technology (Guangzhou); Boyu Li, Hong Kong University of Science and Technology (Guangzhou); Qianxi Liu, The Hong Kong University of Science and Technology (Guangzhou); Xiaoying Wei, IIP(Computational Media and Arts); Jianhao Chen, The Hong Kong University of Science and Technology (Guangzhou); Zeyu Wang, The Hong Kong University of Science and Technology (Guangzhou); Mingming Fan, The Hong Kong University of Science and Technology (Guangzhou)
We present FocalSelect, a heuristic selection technique that builds 3D disambiguation through head-hand coordination and scoring-based functions. Our interaction design adheres to the principle that the intended selection range is a small sector of the headset's viewing frustum, allowing optimal targets to be identified within this scope. We also introduce a density-aware adjustable occlusion plane for effective depth culling of rendered objects. Two experiments are conducted to assess the adaptability of FocalSelect across different input modalities and its performance against five techniques. The results indicate that FocalSelect enhances selections in occluded and remote scenarios while preserving the spatial context among objects.
Session: Cognitive Load, Attention, and Perception (7 - B)
Predicting and Explaining Cognitive Load, Attention, and Working Memory in Virtual Multitasking
Jyotirmay Nag Setu, University of Texas at San Antonio; Joshua M Le, University of Texas at San Antonio; Ripan Kumar Kundu, University of Missouri-Columbia; Barry Giesbrecht, University of California -- Santa Barbara; Tobias Höllerer, University of California; Santa Barbara; Khaza Anuarul Hoque, University of Missouri; Kevin Desai, The University of Texas at San Antonio; John Quarles, University of Texas at San Antonio
As VR technology advances, multitasking in immersive environments escalates that comes with cognitive challenges, emphasizing the importance of attention, working memory, and managing physical and mental strain. To address this, we analyzed the VRWalking dataset, which includes eye and head tracking, heart rate, and galvanic skin response (GSR) data with labels for physical load, mental load, working memory, and attention. Using deep learning models, we achieved accuracies of 91%, 96%, 93%, and 91%, respectively. SHAP analysis revealed key predictive features, offering insights for optimizing user experience and minimizing cognitive strain. Our findings guide future research and enhance virtual reality application design.
Immersion, Attention, and Collaboration in Spatial Computing: A Study on Work Performance with Apple Vision Pro
Carolin Wienrich, University of Würzburg; David Obremski, University of Würzburg
Spatial computing will enable focused work through a higher degree of immersion and collaborative work through enhanced integration of shared interactions. With the Apple Vision Pro, the level of immersion can be adjusted seamlessly. As a first study, the present work investigates different levels of immersion while solving various tasks with the Apple Vision Pro. Selective attention improves significantly with increasing immersion levels. In contrast, social presence decreases with increasing immersion. In addition, we could not determine any adverse effects on depth perception after use. The present study provides an initial contribution to the systematic investigation of various conditions of spatial computing on work performance.
Cognitive Load Inference Using Physiological Markers in Virtual Reality
Jishang Wei, HP Labs; Erika H Siegel, HP Labs; Prahalathan Sundaramoorthy, HP Labs; Ant“nio Carlos D Almeida Gomes, HP Inc.; Shibo Zhang, HP Inc.; Mithra Vankipuram, HP Inc; Kevin Lee Smathers, PG&E; Sarthak Ghosh, HP Inc.; Hiroshi Horii, Samsung Research America; Jeremy N. Bailenson, Stanford University; Rafael 'Tico' Ballagas, HP Inc.
We conducted a largescale study (N=738) to collect behavioral and physiological measures under different cognitive load conditions in a VR environment, and developed an AI model to predict cognitive load in real time. Our model predicts cognitive load as a continuous value from 0 to 1, where 0 and 1 correspond to the lowest and highest reported cognitive loads across all participants. Besides the point estimation, our model quantifies prediction uncertainty using a prediction interval. We release data from 100 participants. This dataset includes recordings from multiple sensors (pupillometry, eye-tracking, and pulse plethysmography), self-reported cognitive effort, behavioral task performance, and demographic information.
Comparing Absolute Size Perception in Optical See-Through Augmented Reality and Real World Viewing Using Verbal and Physical Judgments
Chao-Kuo Chiu, National Yang Ming Chiao Tung University; Jung-Hong Chuang, National Chiao Tung University; Christopher Pagano, Clemson University; Sabarish V. Babu, Clemson University
This empirical evaluation aimed to investigate how size perception operates between OST AR and the real world, using verbal reports and physical judgments. In a within-subjects experimental design, participants viewed target objects of different sizes in both AR and real-world conditions and estimated their sizes using verbal and physical judgment methods across multiple trials. Our findings revealed that participants consistently underestimated object sizes when using verbal reports in both AR and real-world conditions, with more pronounced errors in AR. In contrast, physical judgments yielded more accurate size estimates under both viewing conditions. The accuracy of verbal reports decreased as target sizes increased, especially in AR.
Perception-driven Soft-Edge Occlusion for Optical See-Through Head-Mounted Displays
Xiaodan Hu
Systems with occlusion capabilities, such as those used in vision augmentation, image processing, and optical see-through head-mounted display (OST-HMD), have gained popularity. Achieving precise (hard-edge) occlusion in these systems is challenging, often requiring complex optical designs and bulky volumes. On the other hand, utilizing a single transparent liquid crystal display (LCD) is a simple approach to create occlusion masks. However, the generated mask will appear defocused (soft-edge) resulting in insufficient blocking or occlusion leakage. In our work, we delve into the perception of soft-edge occlusion by the human visual system and present a preference-based optimal expansion method that minimizes perceived occlusion leakage. In a user study involving 20 participants, we made a noteworthy observation that the human eye perceives a sharper edge blur of the occlusion mask when individuals see through it and gaze at a far distance, in contrast to the camera system's observation. Moreover, our study revealed significant individual differences in the perception of soft-edge masks in human vision when focusing. These differences may lead to varying degrees of demand for mask size among individuals. Our evaluation demonstrates that our method successfully accounts for individual differences and achieves optimal masking effects at arbitrary distances and pupil sizes.
Session: Therapy and rehabilitation (7 - C)
MRUCT: Mixed Reality Assistance for Acupuncture Guided by Ultrasonic Computed Tomography
Xinkai Wang, Southeast University; Yue Yang, Stanford University; Kehong Zhou, Southeast University; Xue Xie, Shanghai Sixth People's Hospital; Lifeng Zhu, Southeast University; Aiguo Song, Southeast University; Bruce Daniel, Stanford University
We present MRUCT, a system combining ultrasonic computed tomography (UCT) and mixed reality (MR) to visualize acupuncture points in real-time, addressing the challenges faced by novice practitioners. The system provides offline image registration and real-time guidance for accurate needle insertion based on anatomical structures. We detail non-rigid registration methods for UCT data and key MR system design considerations. Evaluations with two 3D user interfaces show improved performance over traditional workflows for new practitioners and medical students, highlighting MR's potential to enhance acupuncture practices and facilitate clinical training.
Setting the Stage: Using Virtual Reality to Assess the Effects of Music Performance Anxiety in Pianists
Nicalia Elise ThompSon, Goldsmiths; Xueni Pan, Goldsmiths; Maria Herrojo Ruiz, Goldsmiths
Music Performance Anxiety (MPA) is highly prevalent among musicians and often debilitating. We systematically assessed the effect of MPA on performance, physiology, and anxiety ratings using virtual reality (VR) to induce heightened MPA levels in pianists. Twenty pianists completed a performance task under two conditions: a public -audition- and a private -studio- rehearsal. Participants experienced VR pre-performance settings before transitioning to live piano performance in the real world. Pianists in the Audition condition experienced higher somatic anxiety and increased in performance accuracy over time, with a reduced error rate. Additionally, their performances were faster and featured increased note intensity.
ArmVR: Innovative Design Combining Virtual Reality Technology and Mechanical Equipment in Stroke Rehabilitation Therapy
Jing Qu, Shandong University; Lingguo Bu, Shandong University; Zhongxin Chen, North University of China; Yalu Jin, Shandong Yiyang Health Group Zhangqiu Yile Hospital; Lei Zhao, Shandong Jianzhu University; Shantong Zhu, Communication University of China; Fenghe Guo, Shandong Jianzhu University
The rising incidence of stroke presents a global health challenge. Virtual reality (VR) offers promise for rehabilitation due to its immersive qualities. This study developed the ArmVR system, combining VR with rehabilitation hardware to aid stroke recovery. Assessments showed VR supports cognitive and motor function recovery. Forest scenarios help relax emotions, while city scenarios activate motor centers more effectively. Normal users experienced cognitive changes, while stroke patients demonstrated improvements in motor function. Personalized VR systems have significant potential to enhance recovery, meet diverse patient needs, and improve quality of life.
Focus-Driven Augmented Feedback: Enhancing Focus and Maintaining Engagement in Upper Limb Virtual Reality Rehabilitation
Kai-Lun Liao, Xi'an Jiaotong-Liverpool University; Mengjie Huang, Xi'an Jiaotong-Liverpool University; Jiajia Shi, Kunshan Rehabilitation Hospital; Min Chen, Xi'an Jiaotong-Liverpool University; Rui Yang, Xi'an Jiaotong-Liverpool University
The Focus-Driven Augmented Feedback (FDAF) system leverages eye-tracking to enhance focus and maintain engagement in upper limb VR rehabilitation. The FDAF system creates a personalised rehabilitation experience through dynamic real-time augmented visual feedback based on participants' gaze and focus rate. Three experimental studies involving 71 healthy participants and 17 patients tested varying feedback levels. Results identified 30% augmentation as optimal for healthy participants and 20% for patients, balancing focus and engagement. This research demonstrates the potential of eye-tracking to improve VR rehabilitation outcomes and contributes valuable insights into developing dynamic feedback systems for therapeutic applications.
Examining the Validity of An Endoscopist-patient Co-participative Virtual Reality Method (EPC-VR) in Pain Relief during Colonoscopy
Yulong Bian, Shandong University; Juan Liu, Shandong University; Yongjiu Lin, Shandong University; Weiying Liu, School of Mechanical; Electrical&Information Engineering; Yang Zhang, School of Mechanical; Electrical & Information Engineering; Tangjun Qu, Shandong University; Sheng Li, Peking University; Zhaojie Pan, The First Affiliated Hospital of Jinan University; Wenming Liu, The First Affiliated Hospital of Jinan University; Wei Huang, Jinan University; Ying Shi, Jinan University
To relieve perceived pain in patients undergoing colonoscopy, we developed an endoscopist-patient co-participative VR tool (EPC-VR) based on A Neurocognitive Model of Attention to Pain. It allows the patient actively play a VR game and supports endoscopist to trigger a distraction mechanism to divert the patient's attention away from the medical procedure. We performed a comparative clinical study with 40 patients. Patients' perception of pain and affective response were evaluated and the results support the effectiveness of EPC-VR: active VR playing with endoscopists' participation can help relieve the perceived pain and scare of patients undergoing colonoscopy. Finally, 87.5% of patients opt to use the VR application in next colonoscopy.
Session: Enhancing Immersive Experiences (7 - D)
Adaptive 360-Degree Video Streaming with Super-Resolution and Interpolation
Siyuan Hong, The Pennsylvania State University; Ruiqi Wang, Pennsylvania State University; Guohong Cao, Penn State University
360° video streaming requires considerable bandwidth, and previous research has proposed to apply super-resolution (SR) to solve this issue. However, most mobile devices lack the computational capacity to apply SR for all tiles in real time. To solve this problem, we utilize linear interpolation, which requires much less computational overhead. We develop a system that adaptively selects the most suitable downloading strategy, whether interpolation, SR, or ABR at the appropriate bitrate, for each tile to maximize user QoE while considering bandwidth limitations and computational constraints of mobile devices. Extensive evaluations show that our solution significantly outperforms existing techniques in terms of QoE under various scenarios.
Interactions between vibro-acoustic discomfort and visual stimuli: Comparison of real, 3D and 360 environment
Charlotte Scarpa, Centre Scientifique et Technique du Batiment; Toinon Vigier, Université de Nantes; Gwénaelle Haese, Centre Scientifique et Technique du Batiment; Patrick Le Callet, Université de Nantes
This study explores the use of Virtual Reality (VR) in enhancing sensory analysis for optimizing office environments. We compare the ecological validity of 360° and 3D VR formats against real-life settings, focusing on visual perception and immersion. Additionally, we investigate the impact of varying intensities of vibroacoustic stimuli on the perception and cognition of 30 participants. Employing a combination of subjective assessments, physiological, and cognitive data, this preliminary study introduces an immersive methodology in VR. This approach offers new insights into multisensory user experiences, advancing VR applications in building and interior environment evaluations.
Comparative Analysis of AR, VR, and Desktop Tools for Prototyping Augmented Reality Services
Jérémy Lacoche, Orange Labs; Anthony Foulonneau, Orange; Stéphane Louis Dit Picard, Orange
This paper explores the challenges novice users face in creating Augmented Reality (AR) services for specific environments due to complex 3D editors and coding requirements. It compares three tools, desktop, Virtual Reality (VR), and AR editors focusing on AR asset positioning. Using a between-subjects design, users tested one editor across two scenarios (smart-home and smart-building) to evaluate performance, usability, workload, and user experience. Results showed AR and VR tools outperformed the desktop editor in task completion, usability, and enjoyment, indicating their potential as preferred options. Scenario and environment size had minimal impact, warranting further investigation.
ESIQA: Perceptual Quality Assessment of Vision-Pro-based Egocentric Spatial Images
Xilei Zhu, Shanghai Jiao Tong University; Liu Yang, Shanghai Jiao Tong University; Huiyu Duan, Shanghai Jiao Tong University; Xiongkuo Min, Shanghai Jiao Tong University; Guangtao Zhai, Shanghai jiaotong university; Patrick LE CALLET, Université de Nantes
With the development of eXtended Reality (XR), egocentric spatial images and videos are emerging as a compelling form of stereoscopic XR content. However, the image quality assessment (IQA) research for egocentric spatial images is still lacking. We establish the first Egocentric Spatial Images Quality Assessment Database (ESIQAD), including 500 images and their MOSs in 2D, 3D-window, and 3D-immersive display modes. Furthermore, we propose a novel mamba2-based multi-stage feature fusion model ESIQAnet, which predicts the perceptual quality of egocentric spatial images under the three display modes. Extensive experimental results show that the ESIQAnet outperforms 22 state-of-the-art IQA models on the ESIQAD under all three display modes.
Beyond Mute and Block: Adoption and Effectiveness of Safety Tools in Social Virtual Reality, from Ubiquitous Harassment to Social Sculpting
Maheshya Weerasinghe, University of Glasgow; Shaun Alexander Macdonald, University of Glasgow; Cristina Fiani, University of Glasgow; Joseph O'Hagan, University of Glasgow; Mathieu Chollet, University of Glasgow; Mark McGill, University of Glasgow; Mohamed Khamis, University of Glasgow
Harassment in Social Virtual Reality (SVR) is increasing, yet it is unclear whether existing safety measures adequately protect users. We surveyed 100 users across 12 SVR platforms about harassment like hate speech, stalking, and physical abuse, and how they use muting, blocking, personal boundaries, and safety gestures. We found reactive features are often insufficient. Users struggle with inconsistencies across platforms, relying on familiar tools and approaches at the cost of safety. Proactive safety strategies help shape personalised experiences but risk fragmenting communities. Our findings highlight the need for standardised, proactive anti-harassment measures, guiding future development of more effective, platform-wide safety tools.
Session: Visual Perception and Interaction (8 - B)
ChromaGazer: Unobtrusive Visual Modulation using Imperceptible Color Vibration for Visual Guidance
Rinto Tosa, University of Tsukuba; Shingo Hattori, Cluster Metaverse Lab; Yuichi Hiroi, Cluster Metaverse Lab; Yuta Itoh, The University of Tokyo; Takefumi Hiraki, Cluster Metaverse Lab
Visual guidance (VG) is essential for directing user attention. However, traditional methods rely on explicit visual annotations, which often compromise visual clarity and increase cognitive load. we propose an unobtrusive VG method based on color vibration, a phenomenon in which rapidly alternating colors at frequencies above 25 Hz are perceived as an intermediate color. Firstly, we explored a perceptual state that lies between complete color fusion and clear flicker, demonstrating that color differences can be detected without conscious awareness. Secondly, we validated the effectiveness of this technique using natural images and eye-tracking. Consequently, this study provides an unobtrusive VG approach that achieves low cognitive demand.
On Virtual Pointing Rays and Motion Artefacts
Gareth Rendle, Bauhaus-Universität Weimar; Adrian Kreskowski, Bauhaus-Universität Weimar; Bernd Froehlich, Bauhaus-Universität Weimar
Virtual pointing rays allow VR users to select out-of-reach objects. However, fast hand movements can cause the ray to move quickly across the visual field, causing multiple copies of the ray to appear in discrete locations. We analyse gaze and ray movement data during a selection task to show that pointing actions create the conditions necessary for motion artefacts to appear. In addition, we propose a motion blur effect that we apply to the pointing ray with the aim of mitigating motion artefacts. The blur effect is found to improve the perceived smoothness of ray selection, without affecting selection performance.
Children's Calibration of Reaching Estimates in Virtual Reality
Hunter C Finney, University of Utah; Holly C Gagnon, Vanderbilt University; Sarah Creem-Regehr, University of Utah; Bobby Bodenheimer, Vanderbilt University; Jeanine Stefanucci, University of Utah
This study compares children's and adults' perceptions of reaching affordances (up and out) in immersive virtual environments (IVEs). Both groups overestimate their reaching abilities, but especially for reaching out. However, by allowing them to physically reach up or out to the targets (action-outcome feedback), both groups significantly recalibrated their judgments to align their estimates more closely with their actual capabilities. Children showed greater initial overestimation but recalibrated at a similar rate as adults. These findings contribute to the understanding of affordance perception across age groups in IVEs, marking an essential step in adapting virtual reality technology to younger users.
Comparison of Visual Saliency for Dynamic Point Cloud: Task-free vs. Task-dependent
Xuemei Zhou, CWI; Irene Viola, Centrum Wiskunde en Informatica (CWI); Silvia Rossi, Centrum Wiskunde & Informatica (CWI); Pablo Cesar, Centrum Wiskunde & Informatica (CWI)
We present TF-DPC, a task-free eye-tracking dataset for dynamic point clouds in VR, capturing gaze and head movements from 24 participants observing 19 scanned dynamic point clouds with 6 degrees of freedom. Comparing visual saliency maps from task-free and task-dependent (quality assessment) settings, we analyze task influence on visual attention using Pearson correlation and a spatially adapted Earth Mover's Distance. Results reveal significant task-driven differences in attention and provide insights into gaze behavior and movement trajectories. This work informs the development of visual saliency models and enhances VR perception, particularly for dynamic human figure point clouds.
From Novelty to Knowledge: A Longitudinal Investigation of the Novelty Effect on Learning Outcomes in Virtual Reality
Joomi Lee, University of Arkansas; Chen (Crystal) Chen, University of Miami; Aryabrata Basu, University of Arkansas at Little Rock
This study examines the novelty effect of immersive virtual reality (VR) and its influence on long-term learning. A three-wave longitudinal study revealed that while novelty initially boosted user engagement and exploration, it significantly waned over time. Importantly, learning outcomes steadily improved, indicating that the novelty effect may obscure VR's sustained educational benefits. These findings highlight VR's potential as a powerful educational tool and provide guidelines for mitigating the novelty effect in long-term learning strategies, ensuring that immersive environments can support meaningful, lasting learning outcomes beyond initial excitement.
Session: Tactile interfaces and haptic feedback (8 - D)
Smart Pipette: Elevating Laboratory Performance with Tactile Authenticity and Real-Time Feedback
Juan Pieschacon
Mastering the correct use of laboratory equipment is a fundamental skill for undergraduate science students involved in laboratory-based training. However, hands-on laboratory time is often limited, and remote students may struggle as their absence from the physical lab limits their skill development. An air-displacement micropipette was selected for our initial investigation, as accuracy and correct technique are essential in generating reliable assay data. Handling small liquid volumes demands hand dexterity and practice to achieve proficiency. This research assesses the importance of tactile authenticity during training by faithfully replicating the micropipette's key physical and operational characteristics. We developed a custom haptic training approach called 'Smart Pipette' which promotes accurate operation and enhances laboratory dexterity training. A comparative user study with 34 participants evaluated the effectiveness of the Smart Pipette custom haptic device against training with off-the-shelf hardware, specifically the Quest VR hand controller, which was chosen because it is held mid-air similar to a laboratory micropipette. Both training conditions are integrated with the same self-paced virtual simulation displayed on a computer screen, offering clear video instructions and realtime guidance. Results demonstrated that participants trained with the Smart Pipette custom haptic exhibited increased accuracy and precision while making fewer errors than those trained with off-the-shelf hardware. The Smart Pipette and the Quest VR controller had no significant differences in cognitive load and system usability scores. Tactile authentic interaction devices address challenges faced by online learners, while their applicability extends to traditional classrooms, where real-time feedback significantly enhances overall training performance outcomes.
Automatic Generation of Haptic Motion Effects Expressing Human Dance
Jaehyeok Ahn, Pohang University of Science and Technology (POSTECH); Seungmoon Choi, Pohang University of Science and Technology (POSTECH)
Haptic motion effects stimulate the human vestibular system, enhancing multisensory experiences in 4D films.This paper focuses on the automatic creation of motion effects that express the physical movements in human dance using motion capture data. The algorithm processes six-degree-of-freedom motion effects by (1) separating external and internal movements of a dancer's body, (2) segmenting internal movements into a few segments, and (3) merging all external and internal movements with appropriate scaling and weight. Our novel method can elicit better user experiences than two previous motion effect generation algorithms, which respectively represent the movements in general articulated bodies and all dynamic components in the scene.
HIPS - A Surgical Virtual Reality Training System for Total Hip Arthroplasty (THA) with Realistic Force Feedback
Mario Lorenz, Chemnitz University of Technology; Maximilian Kaluschke, University of Bremen; Annegret Melzer, Chemnitz University of Technology; Nina Pillen, Youse GmbH; Magdalena Sanrow, Youse GmbH; Andrea Hoffmann, Chemnitz University of Technology; Dennis Schmidt, FAKT Software GmbH; Andre Dettmann, Chemnitz University of Technology; Angelika Bullinger, Chemnitz University of Technology; Jerome Perret Perret, Haption GmbH; Gabriel Zachmann, University of Bremen
VR training is essential for enhancing patient safety in surgical education. Simulators require visual realism and accurate haptic feedback. Although some surgical simulators are available, authentic training tools are limited in high-force areas like total hip arthroplasty (THA). This research introduces an innovative VR simulation for five steps of THA, with realistic haptic feedback for the first time. The simulation employs a novel haptic hammering device, the enhanced Virtuose 6D from Haption, alongside novel algorithms for collision detection, haptic rendering, and material removal. A study involving 17 surgeons across various skill levels confirmed the realism, practicality, and user-friendliness of our pioneering methods.
In Touch We Decide: Physical Touch by Embodied Virtual Agent Increases the Acceptability of Advice
Atsuya Matsumoto, The University of Tokyo; Takashige Suzuki, the Graduate school of Information Science and Technology; Chi-Lan Yang, The University of Tokyo; Takuji Narumi, the University of Tokyo; Hideaki Kuzuoka, The University of Tokyo
We investigated the use of social touch to enhance trust repair in virtual agents after speech recognition errors. Participants interacted with an agent in VR while a robotic arm provided touch to the back of their hand. Results showed that while social touch did not directly repair trust, it significantly increased the acceptance of the agent's advice, regardless of trust levels. This suggests that physical haptic feedback plays a crucial role in shaping human-agent interactions in VR, particularly influencing decision-making.
Transtiff: A Stylus-shaped Interface for Rendering Perceived Stiffness of Virtual Objects via Stylus Stiffness Control
Ryoya Komatsu, Aoyama Gakuin University; Ayumu Ogura, Aoyama Gakuin University; Shigeo Yoshida, OMRON SINIC X; Kazutoshi Tanaka, OMRON SINIC X Corporation; Yuichi Itoh, Aoyama Gakuin University
The replicating object stiffness is critical for realistic haptic feedback in virtual environments. Existing research has largely overlooked the role of stylus stiffness in shaping the perception of virtual object stiffness. To address this, we conducted a psychophysical experiment showing that dynamic changes in stylus stiffness, combined with visual stimuli, alter user perceptions. Based on these findings, we developed Transtiff, a stylus-shaped interface with on-demand stiffness control using a McKibben artificial muscle. User study confirmed its effectiveness in simulating material properties such as sponge, plastic, and tennis balls, thereby advancing haptic design for VR applications.
Session: Immersive Experiences (9 - A)
Visual Presentation Method for Paranormal Phenomena Through Binocular Rivalry Induced by Dichoptic Color Differences
Kai Guo, Graduate School of Frontier Sciences; Juro Hosoi, The University of Tokyo; Yuki Shimomura, The University of Tokyo; Yuki Ban, The University of Tokyo; Shinichi Warisawa, The University of Tokyo
Paranormal visual effects, such as spirits and miracles, are frequently depicted in games and media design. However, current methods do not express paranormal experiences as aspects of the sixth sense. We propose a novel visual presentation method using binocular rivalry by displaying different images to each eye. In this study, we conducted two experiments. Experiment 1 assessed paranormal sensation, color perception controllability, and visual discomfort caused by mismatched colors in each eye in relation to color difference. Experiment 2 assessed our proposed visual presentation method in three application scenarios. The results indicate that our proposed method improves the visual experience of more realistic paranormal phenomena.
Movement- and Traffic-based User Identification in Commercial Virtual Reality Applications: Threats and Opportunities
Sara Baldoni, University of Padova; Salim Benhamadi, University of Padova; Federico Chiariotti, University of Padova; Michele Zorzi, University of Padova; Federica Battisti, University of Padova
Commercial and gaming VR applications have become pervasive and the need for the secure and convenient identification of users, often overlooked by the research in immersive media, is getting more and more pressing. In this work, we investigate the possibility of identifying users from their movement patterns and data traffic traces while playing four commercial games, using a publicly available dataset. If, on the one hand, this paves the way for easy identification and automatic customization of the virtual reality content, it also represents a serious threat to users' privacy due to network analysis-based fingerprinting. Based on this, we analyze the threats and opportunities for virtual reality users' security and privacy.
Evaluating the Impact of Immersiveness in Virtual Reality Simulations on Anxiety Reduction for MRI Procedures: A Preliminary Study
Hamideh Hosseini Toudeshky, Concordia University; Sarah Seidnitzer, Uniklinikum Erlangen; Sebastian Bickelhaupt, Universitätsklinikum Erlangen; Rola Harmouche, National Research Council Canada; Marta Kersten-Oertel, Concordia University
Magnetic Resonance Imaging (MRI) often triggers anxiety, impacting imaging quality and compliance. This study explores Virtual Reality Exposure Therapy (VRET) using three scenarios: 2D video, 360° video, and fully immersive VR. Thirteen participants underwent all scenarios, with heart rate (HR), skin temperature (SKT), electrodermal activity (EDA), and anxiety self-reports monitored. While no significant HR differences were found, immersive VR elicited higher HR peaks, increased EDA, and reduced SKT. Anxiety decreased post-session, particularly in those with moderate to high pre-session anxiety. These results suggest personalized VRET approaches and real-time biofeedback could improve outcomes.
From Picks to Pixels: An Exploration of Virtual Reality in Geoscience Education
Jacob Young, University of Otago; Matthew Wood, Victoria University of Wellington; Nadia Pantidi, Victoria University of Wellington; Dene Carroll, Te Herenga Waka - Victoria University of Wellington; James Crampton, Te Herenga Waka - Victoria University of Wellington; Cliff Atkins, Te Herenga Waka - Victoria University of Wellington
In this work we present our system for teaching practical geology field skills through a combination of 360° video, photogrammetry, and virtual content. The system was evaluated with first- and second-year undergraduate geoscience students to determine if it was effective in teaching skills that could be transferred to the real world. Students who had performed the task previously saw a significant improvement in their abilities, however this improvement was absent in novice users, suggesting the tool's suitability for revision rather than first-time learning. We discuss these findings and the challenges we encountered in the system's development and deployment.
Designing VR Music Game for Stress Reduction
Kirak Kim, KAIST; Hyojin Kim, KAIST; Youjin Choi, KAIST; Juhan Nam, KAIST; Jeongmi Lee, KAIST
Music therapy is widely used to reduce stress and enhance the well-being of various clinical groups. VR music therapy has been studied to overcome physical constraints of music therapy, but most research focuses on receptive music therapy, failing to utilize VR's interactive potential fully. This paper proposes VR active music therapy based on conventional rhythm-following music therapy methods and VR gamified active music therapy by adding game elements based on Self-Determination Theory (SDT). Our between-subject comparative study (n=33) revealed the stress reduction effects of VR receptive music therapy, VR active music therapy, and VR gamified active music therapy.
Session: Perception and cognition (9 - B)
Seeing is not Thinking: Testing Capabilities of VR to Promote Perspective-Taking
Eugene Kukshinov, University of Waterloo; Federica Gini, University of Trento; Anchit Mishra, University of Waterloo; Nick Bowman, Syracuse University; Brendan Rooney, University College Dublin; Lennart E. Nacke, University of Waterloo
Virtual Reality (VR) offers immersive experiences, often considered a tool for fostering perspective-taking (PT) by enabling users to see through another's eyes. However, our study reveals that VR's commonly assumed ability to facilitate PT is overstated. In a 2x2 design, we examined the effects of point of view (first-person, 1PP vs. third-person, 3PP) and a PT task on users' PT as thinking about others. Results showed that PT significantly improved only when participants were explicitly tasked with considering another's perspective. These findings highlight the importance of intentional design, emphasizing targeted tasks over mere viewpoint shifts to achieve authentic perspective-taking in VR.
When Fear Overshadows Perceived Plausibility: The Influence of Incongruencies on Acrophobia in VR
Larissa Brübach, University of Würzburg; Deniz Celikhan, University of Würzburg; Lennard Rüffert, JuLius-Maximilians-Universität; Franziska Westermeier, Human-Computer Interaction Group; University of Würzburg; Marc Erich Latoschik, University of Würzburg; Carolin Wienrich, University of Würzburg
Virtual Reality Exposure Therapy (VRET) is an effective and affordable treatment for anxiety disorders, including acrophobia. To ensure positive outcomes, it is crucial to understand the effects of VR technology itself. This study examines the impact of the Plausibility illusion (Psi) on fear of heights. In two experiments, 30 participants experienced two heights with congruent and incongruent object behaviors in a 2x2 within-subject design. Results show that only strong incongruencies are recognized, particularly in high-fear conditions. Subtle incongruencies are overshadowed by fear responses. This highlights the need to control incidental factors in VRET to avoid unintended effects on fear responses.
Interpupillary to Inter-Camera Distance of Video See-Through AR and its Impact on Depth Perception
Franziska Westermeier, Human-Computer Interaction Group; University of Würzburg; Chandni Murmu, Clemson University; Kristopher Kohm, Clemson University; Christopher Pagano, Clemson University; Carolin Wienrich, University of Würzburg; Sabarish V. Babu, Clemson University; Marc Erich Latoschik, University of Würzburg
Interpupillary distance (IPD) is crucial to consider for correct depth perception in stereoscopic displays. While HMDs offer lens adjustments for individual IPDs, modern Video See-Through Augmented Reality (VST AR) devices place cameras at fixed positions, potentially causing IPD-ICD mismatches and perceptual incongruencies. In a 2x3 mixed-factor evaluation, we used a near-field reaching task in VR and VST AR, including a feedback phase for perceptual calibration. Linear Mixed Model analysis indicated differences between VR and VST AR, a notable effect of IPD-ICD mismatch, and their interaction. These findings underscore the importance of matching IPD and ICD in depth-sensitive XR applications and practical design considerations.
Brain Signatures of Time Perception in Virtual Reality
Sahar Niknam, University of Luxembourg; Saravanakumar Duraisamy, University of Luxembourg; Jean Botev, University of Luxembourg; Luis A. Leiva, University of Luxembourg
Achieving immersion in VR requires precise measurement of user state. While physical characteristics like locomotion and pose can be accurately tracked in real-time, capturing mental states is harder. Time perception reflects mental states like stress, focus, and boredom, but objective measurement is scarce. This work explores EEG to monitor time perception in VR using neural oscillations and time-frequency analysis. We implemented various time perception modulators, recorded EEG data, and labeled samples as overestimation, correct estimation, or underestimation. Clear EEG spectral signatures for these states were found, enabling VR environments to adapt dynamically to users, increasing immersion and improving user experience.
Adaptive Score Alignment Learning for Continual Perceptual Quality Assessment of 360-Degree Videos in Virtual Reality
Kanglei Zhou, Beihang University; Zikai Hao, Beihang University; Liyuan Wang, Tsinghua University; Xiaohui Liang, Beihang University
Virtual Reality Video Quality Assessment (VR-VQA) aims to evaluate the perceptual quality of 360-degree videos, which is crucial for ensuring a distortion-free user experience. Traditional methods trained on static datasets with limited distortion diversity struggle to balance correlation and precision. To this end, we propose a novel approach for VR-VQA, Adaptive Score Alignment Learning (ASAL). ASAL integrates correlation loss with error loss to enhance alignment with human subjective ratings and precision in predicting perceptual quality. Furthermore, we extend ASAL with adaptive memory replay as a novel Continul Learning (CL) framework. Our experiments demonstrate that ASAL outperforms recent strong baseline models.
Session: Display Technologies and Rendering Techniques (9 - C)
Color Correction for Occlusion-Capable Optical See-Through Head-Mounted Displays by Using Phase-Modulation
Yan Zhang, Shanghai Jiao Tong University; Shulin Hong, shanghai jiao tong university; Weike Qian, Shanghai JiaoTong University; Keyao You, Shanghai Jiao Tong University; Hangyu Zhou, Shanghai Jiao Tong University; Kiyoshi Kiyokawa, Nara Institute of Science and Technology; Xubo Yang, SHANGHAI JIAO TONG UNIVERSITY
OC-OSTHMDs overcome the deficiency of semi-transparent virtual images in AR scenarios, considerably improving the graphics performance. Existing OC-OSTHMDs achieve compact form factors by compressing the optical system based on the modulation of light polarization. However, the wavelength sensitivity of polarizing optical elements causes color aberration in the see-through view. We propose the spectrum-tuning method that mitigates color aberration of the see-through view caused by the wavelength sensitivity of OC-OSTHMDs. The methods operate on a spectrum-based color perception model that formulates the variation of the visible spectrum through OC-OSTHMDs. The optimization is performed globally, requiring minimal computation at runtime.
SRBF-Gaussian: Streaming-Optimized 3D Gaussian Splatting
Dayou Zhang, The Chinese University of Hong Kong; Shenzhen Zhicheng LIANG, The Chinese University of Hong Kong (Shenzhen); Zijian Cao, The Chinese University of Hong Kong; Shenzhen Dan Wang, The Hong Kong Polytechnic University; Fangxin Wang, The Chinese University of Hong Kong; Shenzhen
3D Gaussian Splatting (3DGS) offers high-quality scene representation but faces challenges in streaming applications due to its large data volume. We propose SRBF-Gaussian, which introduces viewport-dependent color encoding using Spherical Radial Basis Functions (SRBFs) and HSL color space. Our system implements adaptive Gaussian pruning and multi-level representations, enabling efficient streaming based on viewport and network conditions. In cloud VR scenarios, our approach achieves up to 14.17% higher PSNR, 59.16% lower latency, and 30.12% better Quality of Experience (QoE).
DPCS: Path Tracing-based Differentiable Projector-Camera Systems
Jijiang Li, Southwest University; Qingyue Deng, Southwest University; Haibin Ling, Stony Brook University; Bingyao Huang, Southwest University
Projector-camera systems (ProCams) simulation aims to model the physical project-and-capture process and associated scene parameters of a ProCams, and is crucial for spatial augmented reality (SAR) applications. In this paper, we introduce a novel path tracing-based differentiable projector-camera systems (DPCS). Our DPCS models the physical project-and-capture process using differentiable physically-based rendering (PBR), thus the scene parameters can be explicitly decoupled and learned. In experiments, DPCS demonstrates clear advantages over previous approaches in ProCams simulation, offering better interpretability, more efficient handling of complex interreflection and shadow, and requiring fewer training samples.
MPGS: Multi-plane Gaussian Splatting for Compact Scenes Rendering
Deqi Li, Beijing Normal University; Shi-Sheng Huang, Beijing Normal University; Hua Huang, Beijing Normal University
This paper presents Multi-Plane Gaussian Splatting (MPGS), a novel method for high-fidelity rendering of heterogeneous scenes with compact reconstruction. MPGS enhances 3D Gaussian Splatting (3DGS) by optimizing Gaussian distributions for both rich and weak-textured regions. A multi-scale geometric correction mechanism refines the 3D Gaussian distribution for efficient scene representation. Regularizing with normal information further improves rendering quality. MPGS efficient reduces storage, and boosts real-time performance. Experimental results demonstrate that MPGS outperforms existing methods, particularly in weak-textured areas, establishing it as a state-of-the-art approach for scene reconstruction.
Silo: Half-Gigapixel Cylindrical Stereoscopic Immersive Display
Saeed Boorboor, Stony Brook University; Doris Gutierrez-Rosales, Stony Brook University; Aahmed Shoaib, Stony Brook University; Chahat Kalsi, Stony Brook University; Yue Wang, Stony Brook University; Yuyang Cao, Stony Brook University; Xianfeng David Gu, Stony Brook University; Arie Kaufman, Stony Brook University
We present the Silo, an immersive stereoscopic cylindrical tiled-display visualization facility that delivers an ultra-high-resolution image of 619 million pixels and a nearly 360-degree horizontal field of regard. In this article, we cover the design principles and the hardware and software systems used in constructing the Silo. To compensate for missing visual information due to the absence of a ceiling and floor, we have designed a mapping to reproject the entire 360 virtual scene to the available display real estate. Our user studies indicate that the Silo enhances natural exploration and visualization compared to traditional curved and flat powerwall displays.
Session: Presence, body ownership, and agency (9 - D)
Am I (Not) a Ghost? Leveraging Affordances to Study the Impact of Avatar/Interaction Coherence on Embodiment and Plausibility in Virtual Reality
Florian Dufresne, Arts et Métiers Institute of Technology; Charlotte Dubosc, Arts et Métiers Institute of Technology; Titouan LEFROU, Arts et Métiers Institute of Technology; Geoffrey Gorisse, Arts et Métiers Institute of Technology; Olivier Christmann, Arts et Métiers Institute of Technology
Understanding how avatars suggest possibilities for action, namely affordances, has attracted attention in the human-computer interaction community. Indeed, avatars' aesthetic features may signify false affordances conflicting with users' expectations and impacting avatar plausibility. The proposed research initially aimed at exploring the use of such affordances to investigate the impact of congruence manipulations on the sense of embodiment. However, it appeared participants' behavior was driven by other processes priming over the perception of the affordances. We unexpectedly manipulated the internal congruence following repeated exposures, causing a rupture in plausibility and significantly lowering embodiment scores and performance.
Illuminating the Scene: How Virtual Environments and Learning Modes Shape Film Lighting Mastery in Virtual Reality
Zheng Wei, The Hong Kong University of Science and Technology; Jia Sun, The Hong Kong University of Science and Technology (Guangzhou); Junxiang LIAO, The Hong Kong University of Science and Technology (Guangzhou); Lik-Hang Lee, Hong Kong Polytechnic University; Chan In Devin SIO, The Hong Kong Polytechnic University; Pan Hui, The Hong Kong University of Science and Technology; Huamin Qu, The Hong Kong University of Science and Technology; Wai Tong, Texas A&M University; Xian Xu, The Hong Kong University of Science and Technology
This study explores the impact of virtual environments and collaboration on learning film lighting techniques in VR education. A 3ž2 factorial experiment with 36 participants examined three environments' baseline, dynamic beach, and familiar office. Results revealed the beach heightened engagement but increased frustration for individuals, while team learning in the office reduced frustration and improved collaboration. Team-based learning excelled in the baseline, while individuals performed better in challenging settings. These findings offer insights into optimizing VR environments to enhance learning outcomes in both individual and collaborative settings.
Tap into Reality: Understanding the Impact of Interactions on Presence and Reaction Time in Mixed Reality
Yasra Chandio, University of Massachusetts; Amherst; Victoria Interrante, University of Minnesota; Fatima Muhammad Anwar, UMASS Amherst
Enhancing presence in MR requires precise measurement. While traditionally assessed through questionnaires, recent work links presence to objective metrics like reaction time. However, the effect of interaction type remains unclear. To investigate, we conducted a within-subjects study with 50 participants, comparing presence and reaction time across direct and symbolic interactions in selection and manipulation tasks. Our results show a correlation of -0.54 between presence and reaction time, indicating that interaction type influences response speed and presence similarly.
Reaction Time as a Proxy for Presence in Mixed Reality Environments with Break-In Presence
Yasra Chandio, University of Massachusetts; Amherst; Victoria Interrante, University of Minnesota; Fatima Muhammad Anwar, UMASS Amherst
Distractions in mixed reality (MR) environments affect presence, reaction time, cognitive load, and Break in Presence (BIP), where attention shifts from the virtual to the real world. While prior work has examined these factors individually, their relationship remains underexplored in MR, where users engage with real and virtual stimuli. We present a model examining how congruent and incongruent distractions affect these constructs. In a within-subject study (N=54), participants performed image-sorting tasks under different distraction conditions. Our results show that incongruent distractions increase cognitive load, slow reaction times, and elevate BIP frequency, with presence mediating these effects.
One Body, But Four Hands: Exploring the Role of Virtual Hands in Virtual Co-embodiment
Jingjing Zhang, Hong Kong University of Science and Technology (Guangzhou); Xiyao Jin, University of Tsukuba; Han Tu, Massachusetts Institute of Technology; Hai-Ning Liang, The Hong Kong University of Science and Technology (Guangzhou); Zhuying Li, Southeast University; Xin Tong, Hong Kong University of Science and Technology (Guangzhou)
Virtual co-embodiment allows two users to share a virtual avatar in Virtual Reality (VR), facilitating skill acquisition by enabling teachers to share hand movements with learners while influencing Sense of Agency (SoA), Sense of Ownership (SoO) and social presence. However, mismatches between real and displayed movements of users in VR can negatively impact these perceptions. Through two user studies, we explored the role of self-hands and the visual factors (transparency and congruency) of self-hands and co-embodied hands in enhancing these perceptions. Results highlight that maintaining self-hands and both hands with identical transparency improves SoO. These findings provide insights for designing virtual hands in VR co-embodiment.
Session: Immersive Applications for Training, Shopping, and Social Interaction (10 - A)
AR Fitness Dog: The Effects of a User-Mimicking Interactive Virtual Pet on User Experience and Social Presence in Physical Exercise
Hyeongil Nam, University of Calgary; Kisub Lee, Hanyang University; Jong-Il Park, Hanyang University; Kangsoo Kim, University of Calgary
This paper explores the impact of an augmented reality (AR) virtual dog that physically presents and mimics user behavior on exercise experiences in solo and group (pair) settings. A human-subject experiment compared three conditions: a mimicking virtual dog, a randomly behaving virtual dog, and no virtual dog. Results revealed that the mimicking virtual dog significantly enhanced the solo exercise experience by fostering companionship and improved group cohesion in pair settings by acting as a social facilitator. These findings underscore the potential of behavior-mimicking virtual pets to enhance exercise experiences and inform the development of AR-based fitness applications.
PaRUS: A Virtual Reality Shopping Method Focusing on Contextual Information between Products and Real Usage Scenes
Yinyu Lu, Zhejiang University; Weitao You, Zhejiang University; Ziqing Zheng, Zhejiang University; Yizhan Shao, Zhejiang University; Changyuan Yang, NO.699 West Wenyi Road Yuhang District; Zhibin Zhou, The Hong Kong Polytechnic University
Current VR shopping applications lack integration with real usage contexts, limiting users' ability to evaluate products in their intended environments. We present PaRUS, a VR shopping approach that reconstructs users' real usage scenes through a semantic scene reconstruction pipeline (with human involvement). By enabling users to visualize products in their actual intended environments, PaRUS bridges the gap between virtual shopping and real-world usage. Two user studies were conducted and the results demonstrate that PaRUS significantly reduces perceived performance risk and enhances purchase satisfaction and trust.
Enhancing Patient Acceptance of Robotic Ultrasound through Conversational Virtual Agent and Immersive Visualizations
Tianyu Song, Technical University of Munich; Felix Pabst, Technical University of Munich; Ulrich Eck, Technische Universitaet Muenchen; Nassir Navab, Technische Universität München
Robotic ultrasound systems can enhance medical diagnostics, but patient acceptance is a challenge. We propose a system combining an AI-powered conversational virtual agent with three mixed reality visualizations to improve trust and comfort. The virtual agent, powered by a large language model, engages in natural conversations and guides the ultrasound robot, enhancing interaction reliability. The visualizations include augmented reality, augmented virtuality, and fully immersive virtual reality, each designed to create patient-friendly experiences. A user study demonstrated significant improvements in trust and acceptance, offering valuable insights for designing mixed reality and virtual agents in autonomous medical procedures.
Enhancing consumer insights through VR metaphor elicitation
Sai Priya Jyothula, University of Illinois Chicago; Andrew E Johnson, University of Illinois Chicago
Our research introduces a novel system for performing metaphor elicitation interviews (consumer research) in VR, leveraging metaphors and non-verbal communication. Participants use 3D objects to express thoughts and emotions, instead of traditional pictures. The system features an asymmetric setup, with participants immersed in VR through an HMD, while the interviewers view the participant's perspective on a monitor. We discuss the technical and design aspects of the VR system, validating the effectiveness of this approach through a user study (N=17). Results show that immersion, presence and embodied interaction in VR positively affect and aid in sense-making and deeper expression of the participants' thoughts, perspectives and emotions.
OPVSim: Applying a Graph-Based Methodology for VR Training in Guided Learning of Emergency Procedures in a Ship's Engineering Room
Vivian Gómez, Universidad de los Andes; Pablo Figueroa, Universidad de los Andes; Aldo Francisco Lovo Capit n de Corbeta, Colombian Navy
Virtual Reality offers scalable training for step-by-step procedures in machinery use, like those in engineering rooms. Many current training methods lack interaction with real machinery and allow no human errors. To bridge this gap, we used a graph-based methodology (called ProtoColVR) to design a VR training simulation for emergency procedures in a ship's engineering room. The simulator includes a tutorial, step-by-step training, feedback, and alternative paths to the goal. A study with 32 participants (cadets and experienced officers) showed increased knowledge, improved usability, presence, and workload outcomes. We discuss our findings, limitations, and the implications for designing VR training systems for guided procedural learning.
Session: Enhancing Interaction and Feedback in Virtual and Cross-Reality Systems (10 - B)
Presenting Tingling Aftereffects Using Vibro-Thermal Feedback to Enhance Impact Sensation in Virtual Reality
Hikaru Fujita, The University of Tokyo; Juro Hosoi, The University of Tokyo; Yuki Ban, The University of Tokyo; Shinichi Warisawa, The University of Tokyo
Impact sensation is crucial for immersive VR, yet many systems struggle to produce strong, realistic impacts, constrained by device size, complexity, and safety. Most solutions focus solely on the moment of collision, this study investigates overlooked aftereffects, tingling sensations that emerge within the body after a strong impact. We designed vibrotactile and thermal feedback to replicate those aftereffects, which significantly enhanced perceived impact realism, intensity, and tingling in a user study. High-frequency vibrotactile stimuli particularly enhanced tingling sensation and realism. These results highlight the potential for aftereffects feedback to enhance impact sensations for entertainment and safety education purposes.
Optimizing Moving Target Selection in VR by Integrating Proximity-Based Feedback Types and Modalities
Xuning Hu, Xi'an Jiaotong-Liverpool University; Wenxuan Xu, Xi'an Jiaotong-Liverpool University; Yushi Wei, The Hong Kong University of Science and Technology (Guangzhou); Hao Zhang, Chinese Academy of Sciences; Jin Huang, Chinese Academy of Sciences; Hai-Ning Liang, The Hong Kong University of Science and Technology (Guangzhou)
Proximity-based feedback offers users real-time guidance toward an interaction goal, especially in tasks like selecting moving targets. This work explores proximity-based feedback types and modalities to improve the selection of moving targets in VR by leveraging three feedback types that combine visual, auditory, and haptic modalities. We evaluated the performance of these mechanisms through two user studies to explore the characteristics of feedback types across different modalities and to examine the roles of various modalities within multimodal combinations. Our findings suggest optimal selection mechanisms for developers and should be tailored to different goals.
Predicting Ray Pointer Landing Poses in VR Using Multimodal LSTM-Based Neural Networks
Wenxuan Xu, Xi'an Jiaotong-Liverpool University; Yushi Wei, The Hong Kong University of Science and Technology (Guangzhou); Xuning Hu, Xi'an Jiaotong-Liverpool University; Wolfgang Stuerzlinger, Simon Fraser University; Yuntao Wang, Tsinghua University; Hai-Ning Liang, The Hong Kong University of Science and Technology (Guangzhou)
Target selection is crucial in VR interaction. Prediction heuristics enhance user experience. We aim to predict ray landing poses for hand-based raycasting in VR using an LSTM neural network with time-series data from three channels: hand, HMD, and eye. A study gathered motion data from these inputs and analyzed behaviors to identify the best combination of modalities. A second study validated raycasting across various distances, angles, and target sizes. Our technique predicted landing poses within 4.6° accuracy and, compared to baseline and prior kinematic methods, it improved prediction accuracy by 3.5 and 1.9 times, respectively.
Comparison of Cross-Reality Transition Techniques Between 3D and 2D Display Spaces in Desktop-AR Systems
Robbe Cools, KU Leuven; Inne Maerevoet, KU Leuven; Matt Gottsacker, University of Central Florida; Adalberto L. Simeone, KU Leuven
The aim of this study is to develop an understanding of how virtual objects can be transitioned between 3D Augmented Reality and 2D standard monitor display spaces. The increased availability of Augmented Reality devices, gives rise to hybrid setups in which users may need to transition virtual objects between display spaces. We developed three virtual object transition techniques: Mouse-based, Hand-based, and a Modality Switch. The three techniques were evaluated in a user study (N=24) alongside a fourth condition in which participants could freely switch between Hand- and Mouse-based techniques. Participants preferred the mouse-based technique, which allowed them to transition objects faster.
From Display to Interaction: Design Patterns for Cross-Reality Systems
Cross-reality is an emerging research area concerned with systems operating across realities, which are often complex, and thus there is a need for an overarching design framework. This paper addresses this need by presenting eleven design patterns for cross-reality applications across the following four categories: fundamental, origin, display, and interaction patterns.
Session: Interactive Experience and System Usability (10 - C)
May The Force be With You: Cloning Distant Objects to Improve Medium-Field Interactions in Augmented Reality
Danish Nisar Ahmed Tamboli, University of Florida; Rohith Venkatakrishnan, University of Florida; Roshan Venkatakrishnan, Clemson University; Balagopal Raveendranath, Clemson University; Julia Woodward, University of South Florida; Isaac Wang, James Madison University; Jesse D Smith, University of Florida; Jaime Ruiz, University of Florida
Augmented Reality (AR) features virtual objects registered in the physical world. In this work, we conceptualized a medium-field interaction technique called The Force, which allows users to clone distant objects and manipulate their replicas instead. We compared it against controller-based ray-casting and gaze-based pinching in terms of a pick-and-place task, employing a within-subjects design, studying both objective performance and subjective user experience. Results suggest that The Force can allow for higher accuracy and efficiency in tasks requiring precision and fine motor control. We detail these results, also noting areas that warrant refinement, attempting to improve AR interactions in the future.
Mixed Reality and Real-Time X-Ray Simulation in Vet Radiography Training: A User-Centered Comparative Study
Xuanhui Xu, University College Dublin; Antonella Puggioni, University College Dublin; David Kilroy, University College Dublin; Abraham G. Campbell, University College Dublin
Proficient training in Diagnostic Imaging (DI) is essential for veterinary students but it faces challenges such as limited animal access and strict radiation safety regulations. Two collaborative Mixed Reality (MR) systems simulated the horse DI procedure with real-time X-rays, one using real X-ray equipment and the other virtual. A user-centered comparative study assessed 22 veterinary students and 4 DI experts, focusing on radiograph quality and time to obtain X-rays. Results indicate that the MR system, combined with real-time X-ray synthesis, shows significant potential to overcome challenges in veterinary DI training and serve as a tool for both training and assessment.
Don't They Really Hear Us? A Design Space for Private Conversations in Social Virtual Reality
Josephus Jasper Limbago, Aalto University; Robin Welsch, Aalto University; Florian Müller, TU Darmstadt; Mario Di Francesco, Aalto University
Seamless transition between public dialogue and private talks is essential in everyday conversations. However, existing Social Virtual Reality (VR) platforms are not successful in supporting it and providing safety, thereby hindering self-disclosure. We approach this problem by exploring the factors affecting private conversations in social VR applications, including the usability of different interaction methods and the awareness with respect to the virtual world. We conduct both expert interviews and a controlled experiment with a social VR prototype we realized. We then leverage their outcomes to establish a design space that considers diverse dimensions, laying the groundwork for more intuitive and meaningful experiences in social VR.
Spatial Bar: Exploring Window Switching Techniques for Large Virtual Displays
Leonardo Pavanatto, Virginia Tech; Jens Grubert, Coburg University; Doug Bowman, Virginia Tech
Virtual displays provided through head-worn displays (HWDs) offer users large screen space for productivity, but managing this space effectively presents challenges. This paper explores how to enhance window-switching strategies for virtual displays by leveraging eye tracking provided by HWDs and underutilized spaces around the main display area. We investigate the efficiency and usability of different cursor behaviors and selection modes in a Spatial Bar interface for window-switching tasks in augmented reality environments. Results show gaze coupled with teleport led to the quickest window-switching times, particularly in tasks where the original cursor position or the target window was far from the Spatial Bar.
Shaping the Future of VR Hand Interactions: Lessons Learned from Modern Methods
Byungmin Kim, Korea University; DongHeun Han, Kyung Hee University; HyeongYeop Kang, Korea University
In virtual reality (VR), increased realism in hand-object interactions is often assumed to enhance immersion and usability, but recent studies suggest this is not always the case. This study addresses the lack of clarity in selecting optimal hand interaction methods by evaluating three approaches -Attachment, Penetration, and Torque- across simple and precision-driven tasks. Our findings highlight that while Attachment offers simplified controls ideal for commercial use, Penetration and Torque hold potential for next-generation interactions. By examining technical features and trade-offs, we provide insights to help developers balance realism, usability, and satisfaction in VR applications.
Session: 3D authoring (10 - D)
Personalized Dual-Level Color Grading for 360-degree Images in Virtual Reality
Lin-Ping Yuan, The Hong Kong University of Science and Technology; John J Dudley, University of Cambridge; Per Ola Kristensson, University of Cambridge; Huamin Qu, The Hong Kong University of Science and Technology
The rising popularity of 360-degree images and VR has spurred a growing interest among creators in producing visually appealing content through effective color grading processes. We propose a dual-level Preferential Bayesian Optimization framework that integrates global and local color adjustments tailored for VR environments. We design and evaluate a novel context-aware preferential Gaussian Process (GP) to learn contextual preferences for local colors, taking into account the dynamic contexts of previously established global colors. Additionally, we design and evaluate three VR interfaces for color comparison.
GO-NeRF: Generating Objects in Neural Radiance Fields for Virtual Reality Content Creation
Peng Dai, The University of Hong Kong; Feitong Tan, Google; Xin Yu, The University of Hong Kong; Yifan (Evan) Peng, The University of Hong Kong; Yinda Zhang, Google; xiaojuan qi, The University of Hong Kong
We propose a novel pipeline featuring an intuitive interface, dubbed GO-NeRF. Our approach takes text prompts and user-specified regions as inputs and leverages the scene context to generate 3D objects within the scene. We employ a compositional rendering formulation that effectively integrates the generated 3D objects into the scene, utilizing optimized 3D-aware opacity maps to avoid unintended modifications to the original scene. Furthermore, we develop tailored optimization objectives and training strategies to enhance the model's ability to capture scene context and mitigate artifacts, such as floaters, that may occur while optimizing 3D objects within the scene.
Which Side is the Top? A User Study to Compare Visual Assets for Component Orientation in Assembly with Augmented Reality
Enricoandrea Laviola, Polytechnic University of Bari; Michele Gattullo, Polytechnic University of Bari; Sara Romano, Polytechnic University of Bari; Antonio E. Uva, Polytechnic Institute of Bari
This study explores the use of Augmented Reality (AR) visual assets to convey information about component orientation. We focused on assembly scenarios where no affordance is provided for orientation while maintaining a high affordance for how components are mounted. We evaluated 6 AR presentation modes: image, video, static/animated side-by-side product model, static/animated in-situ product model. We conducted a user study with 36 participants, measuring completion time, accuracy, cognitive load. We also analyzed users' interaction and feedback. Our findings revealed that the animated side-by-side product model ensures better completion time, demanding less cognitive load and being favored by users.
Usability Evaluation of Integrated and Separated Interfaces in an Immersive Authoring Tool based on Panoramic Videos
Daniel Xuan Hien Mai, Université Paris-Saclay; Guillaume Loup, Paris-Saclay; Jean-Yves Pascal Didier, Université Paris-Saclay
This study evaluates the usability of an immersive authoring tool for Virtual Reality Learning Systems (VRLS) using panoramic videos. Two interface designs were compared: one integrating storyboarding and rendering visualization, and another separating these tasks. Results from 24 participants show that the separated interface outperforms the integrated one in effectiveness, efficiency, and user satisfaction, especially in complex scenarios. However, the integrated interface caused more cybersickness symptoms. These findings underline the importance of tailored interface design for immersive tools, offering valuable insights for creating user-friendly VRLS authoring systems that meet diverse design needs.
Virtual Reality Impacts on Novice Programmers’ Self-efficacy
Nanlin Sun, Virginia Tech; Wallace Santos Lages, Northeastern University
Virtual Reality has been used to improve motivation and help in the visualization of complex computing topics. However, few studies directly compared immersive and non-immersive environments. To address this limitation, we developed Abacus, a programming environment that can run in both immersive and non-immersive modes. Participants used a block-based editor to complete two programming tasks: a non-spatial procedural task, and a spatial 3D math task. We found that VR led to higher gains in self-efficacy and that the gain was significant for participants with lower initial levels of experience and spatial skills.
Session: Perception, Embodiment, and Spatial Awareness (11 - A)
A Comparison of the Effects of Older Age on Homing Performance in Real and Virtual Environments
Maggie K McCracken, University of Utah; Corey Shayman, University of Utah; Peter C Fino, University of Utah; Jeanine Stefanucci, University of Utah; Sarah Creem-Regehr, University of Utah
To validate the use of virtual reality (VR) for age-related navigation research, we compared young (18-30) and older adults (60+) performing a homing task in a closely matched real and virtual environment under three conditions: vision only, self-motion only, and combined cues. Both environments revealed similar between-age-group differences. However, young adults performed less accurately with only self-motion cues in VR compared to real world, while older adults showed deficits in VR across both single-cue conditions. Both age-groups performed consistently across environments with multiple cues. These findings establish VR as a valuable tool for studying age-related navigation, especially with multiple sensory cues.
Sensitivity to Redirected Walking Considering Gaze, Posture, and Luminance
Niall L. Williams, University of Maryland; College Park; Logan C. Stevens, University of Maryland; Aniket Bera, Purdue University; Dinesh Manocha, University of Maryland
We study the effects of luminance on users' sensitivity to rotation gains in virtual reality, and study the relationship between rotation gain strength and users' gaze behavior and postural stability. To do this, we conducted a psychophysical experiment to measure users' sensitivity to RDW rotation gains and collect gaze and posture data during the experiment. Our results showed that physiological signals were significantly positively correlated with the strength of redirection gain, the duration of trials, and the trial number. Gaze velocity was negatively correlated with trial duration. Results also showed that there were no significant differences in RDW detection thresholds between the photopic and mesopic luminance conditions.
When Flames Feel Real in Augmented Reality: Effects of Plausibility and Placement of Virtual Flames on the Burning Hand Illusion and Physiological Responses
Daniel Eckhoff, City University of Hong Kong; Jan Schnupp, The Chinese University of Hong Kong; Wong Pui Chung, City University of Hong Kong; Alvaro Cassinelli, City University; Hong Kong
During the Burning Hand Illusion users can perceive illusory heat when observing virtual flames on the hands in Augmented Reality. This study investigates the role of stimulus location within peri-hand space and its resemblance to realistic fire in the emergence of the illusion, alongside autonomic nervous responses measured via skin conductance and pupil dilation. Realistic flames displayed on the hand elicited the strongest heat illusion and physiological responses, whereas green flames or flames positioned away from the hand diminished the illusion's intensity. These findings suggest that both flame location and plausibility significantly influence the BHI, eliciting physiological effects akin to those of real thermal stimuli.
An embodied body morphology task for investigating self-avatar proportions perception in Virtual Reality
Loën Boban, EPFL; Ronan Boulic, EPFL; Bruno Herbelin, EPFL
Body perception is prone to distortions and can be influenced by visual stimuli of altered body shapes. In VR, users can embody avatars with dissimilar appearances. This study examines body perception in immersive VR and the impact of virtual human co-presence. Participants compared avatars with varying body proportions to their perceived body morphology, while being co-present with a virtual agent that could have various appearances. Results show an overestimation of the leg length and highlight how embodiment can affect body judgment
How Collaboration Context and Personality Traits Shape the Social Norms of Human-to-Avatar Identity Representation
Seoyoung Kang, KAIST; Boram Yoon, KI-ITC ARRC; KAIST; Kangsoo Kim, University of Calgary; Jonathan Gratch, University of Southern California; Woontack Woo, KAIST
As avatars evolve from digital representations into identity extensions, they enable unprecedented self-expression beyond physical limitations. Our survey of 150 participants investigated social norms surrounding avatar modifications in various contexts. Results show modifications are viewed more favorably from a partner's perspective, especially for changeable attributes, but are less acceptable in professional settings. Individuals with high self-monitoring are more resistant to changes, while those with higher Machiavellian traits show greater acceptance. These findings suggest creating context-sensitive customization options that balance core identity elements with personality-driven preferences while respecting social norms.
Session: Accessibility of immersive interfaces (11 - B)
VF-Lens: Enhancing Visual Perception of Visually Impaired Users in VR via Adversarial Learning with Visual Field Attention
Xiaoming Chen, The University of Sydney; Han Dehao, Beijing Technology and Business University; Qiang Qu, The University of Sydney; Yiran Shen, Shandong University
This research proposes VF-Lens, a generative adversarial model designed to enhance image perception for visually impaired users in VR. Through adversarial training, VF-Lens adaptively compensates for light sensitivity and generates hyperimages tailored to the user's visual field parameters, enabling perceptions closer to those of normal-vision users. VF-Lens is applicable to various types of visual impairments while bypassing engineering complexities. The proposed 'visual field attention' mechanisms further prioritize critical visual information for improved performance. Extensive evaluations demonstrate the effectiveness of VF-Lens, while a standardized evaluation process ensures reusability and comparability for future research.
BoundaryScreen: Summoning the Home Screen in VR via Walking Outward
Yang Tian, Guangxi University; Xingjia Hao, Guangxi University; Jianchun Su, Guangxi Huasheng New Materials Co.; Ltd; Wei Sun, Institute of Software Chinese Academy of Sciences; Yangjian Pan, Guangxi University; Yunhai Wang, Renmin University of China; Minghui Sun, Jilin University; Teng Han, Institute of Software; Chinese Academy of Sciences; Ningjiang Chen, Guangxi University
Existing implementations of a safety boundary wall neglect to utilize the safety boundary wall's large surface for displaying interactive information. We propose the BoundaryScreen technique based on the walking outward metaphor to add interactivity to the safety boundary wall. Specifically, we augment the safety boundary wall by placing the home screen on it. To summon the home screen, the user only needs to walk outward until it appears. Results showed that (i) participants significantly preferred BoundaryScreen in the outermost two-step-wide ring-shaped section of a circular safety area, and (ii) participants exhibited strong -behavioral inertia- for walking.
Comparing Vibrotactile and Skin-Stretch Haptic Feedback for Conveying Spatial Information of Virtual Objects to Blind VR Users
Jiasheng Li, University of Maryland; Zining Zhang, University of Maryland; Zeyu Yan, University Of Maryland; Yuhang Zhao, University of Wisconsin-Madison; Huaishu Peng, University of Maryland
Perceiving spatial information of a virtual object is critical yet challenging for blind users. To facilitate VR accessibility for blind users, we investigate the effectiveness of two types of haptic cues---vibrotactile and skin-stretch---in conveying the spatial information of a virtual object when applied to the dorsal side of a blind user's hand. We conducted a user study with 10 blind users to investigate how they perceive static and moving objects. Our results reveal that blind users can more accurately understand an object's location and movement when receiving skin-stretch cues, as opposed to vibrotactile cues. We discuss the pros and cons of both haptic cues and conclude with design recommendations for future haptic solutions.
Exploring Aiming Techniques for Blind People in Virtual Reality
Joaeo Mendes, Universidade de Lisboa; Manuel Pi‡arra, Faculdade de Ciências; Universidade de Lisboa; Inês Gon‡alves, Faculdade de Ciências; Universidade de Lisboa; André Rodrigues, Universidade de Lisboa; Joaeo Guerreiro, Universidade de Lisboa
We explore how to support blind people in aiming tasks in VR using an archery scenario. We implemented three techniques: 1) Spatialized Audio, a baseline where the target emits a 3D sound to convey its location, 2) Target Confirmation, where the previous condition is augmented with secondary Beep sounds to indicate proximity to the target, and 3) Reticle-Target Perspective, where the auditory feedback conveys the relation between the target and the user's aiming reticle. In a study with 15 blind participants, Target Confirmation and Reticle-Target Perspective clearly outperformed Spatialized Audio. We discuss how our findings may support the development of VR experiences that are more accessible and enjoyable to a broader range of users.
SummonBrush: Enhancing Touch Interaction on Large XR User Interfaces by Augmenting Users Hands with Virtual Brushes
Yang Tian, Guangxi University; Zhao Su, Guangxi University; Tianren Luo, Institute of Software; Teng Han, Institute of Software; Chinese Academy of Sciences; Shengdong Zhao, City University of Hong Kong; Youpeng Zhang, City University of Hong Kong; Yixin Wang, Guangxi University; BoYu Gao, Jinan University; Dangxiao Wang, Beihang University
Users typically need to perform extensive arm movements for engaging with XR user interfaces much larger than mobile device touchscreens. We propose the SummonBrush technique to facilitate easy access to hidden windows while interacting with large XR user interfaces, requiring minimal arm movements. The SummonBrush technique adds a virtual brush to the index fingertip of a user's hand. The user can summon hidden windows or background applications in situ via pressing the brush against the user interface to make ink fully fill the brush and then performing swipe gestures. Ecological studies showed that SummonBrush significantly reduced arm movement time, leading to a significant decrease in reported physical demand.
Session: Embodiment, Agency, and Interaction (11 - C)
Utilizing Gaze-Contingent Rendering to Maintain Visual Attention in Educational VR
Yue Liu
In educational Virtual Reality (VR) environments, objects irrelevant to learning can lead to students' inattention, which adversely affects learning. However, removing these objects from virtual scenes is not feasible, as they are crucial for creating a realistic and immersive experience. Balancing the need to maintain students' attention while preserving the integrity of scenarios is a challenging task. In this paper, we introduce a gaze-contingent rendering (GCR) technique to address such an issue, which is independent of specific elements or configurations in virtual scenes and adaptable across various contexts. Specifically, we utilize gaze-aware rendering adjustments to adaptively reduce the visibility of objects irrelevant to learning while highlighting relevant ones. We develop three GCR strategies (i.e., blur, pixelation, and underexposure) and investigate how these strategies affect students' visual attention, academic achievement, and perceptions of the learning activity across different scenarios. Our findings indicate that the proposed rendering strategies effectively achieve the goals of sustaining visual attention and improving academic achievement without significantly impacting immersion or engagement. As an initial exploration of GCR for maintaining attention within educational VR, this study may inspire new directions in future research on GCR and visual attention maintenance in immersive VR.
The Hidden Face of the Proteus Effect: Deindividuation, Embodiment and Identification
Anna Martin Coesel, CESI LINEACT; Beatrice Biancardi, CESI LINEACT; Mukesh Barange, CESI LINEACT; Stéphanie Buisine, CESI LINEACT
The Proteus effect describes how users in virtual environments align their attitudes with stereotypes associated with their avatars. This study examines the role of deindividuation in the effect, hypothesizing it reinforces the phenomenon alongside embodiment and identification processes. Two experiments were conducted: the first showed engineering students performed better on a statistics task using an Albert Einstein avatar. Using the same task and avatar, the second study manipulated identity cues to alter deindividuation. While no differences were found between deindividuation conditions, identification and embodiment were found to be predictors of performance, suggesting potential avatar-task relationships within the Proteus effect.
Immersive Tailoring of Embodied Agents Using Large Language Models
Andrea Bellucci, Universidad Carlos III de Madrid; Giulio Jacucci, Helsinki Institute for InformationÿTechnology HIIT; University of Helsinki; Kien Duong, Univerisity of Helsinki; Pritom K Das, Univerisity of Helsinki; Sergei Smirnov, Univerisity of Helsinki; Imtiaj Ahmed, University of Helsinki; Jean-Luc Lugrin, Department of Computer Science; HCI Group
LLM-based embodied agents are emerging in VR, enabling dynamic interactions in scenarios like teaching, companionship, and gaming. We present an architecture integrating LLM modules to support conversational tailoring, allowing users to adjust an agent's behavior, personality, and appearance directly in the virtual space. This approach eliminates the need for pre-development modifications or external interfaces. Performance evaluation highlights latency challenges in the speech-to-text-to-speech pipeline, alongside insights from a six-week user study. While promising, future work focuses on reducing latency and addressing the black-box limitations of LLMs to improve system reliability and responsiveness.
Effects of Embodiment and Personality in LLM-based Conversational Agents
Sinan Sonlu, Bilkent University; Bennie Bendiksen, The University of Massachusetts; Funda Durupinar Babur, University of Massachusetts Boston; Ugur Gudukbay, Bilkent University
This work investigates the effects of personality expression and embodiment in conversational agents. We extend an existing framework by integrating LLM-based conversation support to provide information about contemporary scientific topics. We describe a user study to evaluate two opposing personality styles using a dialogue-only model that conveys personality verbally, an animated human model that expresses personality only through dialogue, and an animated human model expressing personality through dialogue and expressive animations. The users perceive all models positively regarding personality and learning outcomes, however, models with high personality traits are perceived as more engaging than those with low personality traits.
The Impact of Avatar Retargeting on Pointing and Conversational Communication
Simbarashe Nyatsanga, University of California; Davis; Doug Roble, Meta Reality Labs; Michael Neff, Meta Reality Labs
Retargeting user motion to avatars of different proportions changes the motion. This paper presents three studies exploring how these motion changes impact functional and social communication for baseline forward (pose-preserving) and inverse (position-preserving) kinematics retargeting. Pointing accuracy that can be maintained with full constraints on the orientation and position of the hand. Error increases for pose preserving retargeting, but can be minimized by a Semantic IK approach that offers a more natural pose. Similar results are shown for gestures displaying object size. Perception of avatar personality is altered in a speech giving task, but not the object size task.
Session: Social Interaction and Collaboration (11 - D)
Exploring the Influence of Interpersonal Relationships on Gamification Preferences in Collaborative IVR Environments
Shaoteng Ke, University of Nottingham Ningbo China; Lijie ZHENG, University of Nottingham Ningbo China; Boon Giin Lee, University of Nottingham Ningbo China
This study explores the impact of interpersonal relationships on preferences for gamification in collaborative immersive virtual reality (IVR) environments. A multiplayer avatar-based IVR game that features a maze with three puzzles was designed to improve teamwork and communication. The study involved 44 participants, exploring how gamification elements influenced groups of peers versus strangers. The results showed significant improvements in communication, collaboration, motivation, and engagement in both types of groups. Game elements preferences varied based on social relationships and individual characteristics. These findings provide insights into designing IVR collaborative learning environments that emphasize social interaction.
Enhancing Social Experiences in Immersive Virtual Reality with Artificial Facial Mimicry
Alessandro Visconti, Politecnico di Torino; Davide Calandra, Politecnico di Torino; Federica Giorgione, Politecnico di Torino; Fabrizio Lamberti, Politecnico di Torino
The availability of affordable VR hardware and growing interest in the Metaverse are driving the expansion of Social VR (SVR) platforms. These platforms enable real-time avatar-mediated interactions, letting users engage in augmented social dynamics that overcome physical limitations. Artificial Facial Mimicry (AFM), one of such augmentations, can potentially enhance social experiences but remains underexplored in avatar-mediated interactions. To assess AFM's potential, an automated system providing AFM, nodding, and eye contact was developed, integrated into an SVR application, and tested in a controlled user study. Results showed that AFM can improve social closeness, attraction, trust, presence, without impacting avatar naturalness.
PASCAL - A Collaboration Technique Between Non-Collocated Avatars in Large Collaborative Virtual Environments
David Gilbert, RWTH Aachen University; Abhirup Bose, RWTH Aachen University; Tim Weissker, RWTH Aachen University; Torsten Wolfgang Kuhlen, RWTH Aachen University
Collaborative work in large collaborative virtual environments (CVEs) often transitions from loosely- to tightly-coupled collaboration. We present two novel interaction techniques designed to share spatial context over large distances. The first method replicates video conferencing by providing users with a virtual tablet to share video feeds. The second, PASCAL (Parallel Avatars in a Shared Collaborative Aura Link), allows users to share surroundings by creating synchronized copies at collaborators' locations. In a study, PASCAL significantly improved task completion time, communication, understanding, and co-presence. Our insights contribute to successful interaction techniques to mediate between loosely- and tightly-coupled work in CVEs.
Does Hand Size Matter? The Effect of Avatar Hand Size on Non-verbal Communication in Virtual Reality
Jackson Henry, Clemson University; Ryan Canales, Clemson University; Catherine Stolarski, Clemson University; Alex Adkins, Clemson University; Rohith Venkatakrishnan, University of Florida; Roshan Venkatakrishnan, University of Florida; Sophie Jörg, University of Bamberg
Virtual Reality (VR) is becoming a popular platform for remote socializing and collaboration using embodied avatars. However, avatars often do not match the user's proportions, which may affect nonverbal communication. This paper explores the impact of disproportionate avatar hands-on nonverbal communication in VR. We conducted an experiment with 72 participants (36 dyads) in a -Charades- game, where avatars' hands were scaled to 25% smaller, same size, or 25% larger than the user's hands. Results show no significant effects on task performance, communication effectiveness, or user experience, although participants who noticed hand size changes showed slight effects on workload measures.
Session: Cybersickness (12 - A)
A Testbed for Studying Cybersickness and its Mitigation in Immersive Virtual Reality
Davide Calandra
Cybersickness (CS) represents one of the oldest problems affecting Virtual Reality (VR) technology. In an attempt to resolve or at least limit this form of discomfort, an increasing number of mitigation techniques have been proposed by academic and industrial researchers. However, the validation of such techniques is often carried out without grounding on a common methodology, making the comparison between the various works in the state of the art difficult. To address this issue, the present article proposes a novel testbed for studying CS in immersive VR and, in particular, methods to mitigate it. The testbed consists of four virtual scenarios, which have been designed to elicit CS in a targeted and predictable manner. The scenarios, grounded on available literature, support the extraction of objective metrics about user's performance. The testbed additionally integrates an experimental protocol that employs standard questionnaires as well as measurements typically adopted in state-of-the-art practice to assess levels of CS and other subjective aspects regarding User Experience. The article shows a possible use case of the testbed, concerning the evaluation of a CS mitigation technique that is compared with the absence of mitigation as baseline condition.
Beyond Subjectivity: Continuous Cybersickness Detection Using EEG-based Multitaper Spectrum Estimation
Berken Utku Demirel, ETH Zurich; Adnan Harun Dogan, ETH Zurich; Juliete Rossie, ETHZ; Max Möbus, ETH Zürich; Christian Holz, ETH Zürich
Virtual reality (VR) offers immersive experiences but poses a risk of cybersickness, marked by symptoms such as dizziness and nausea. Traditional assessments rely on post-immersion questionnaires in controlled setups. This paper explores cybersickness's dynamic nature during interactive VR use. We proposed a method to continuously monitor cybersickness using passive electroencephalography (EEG) and motion signals. By estimating the multitaper spectrum, the method tracks cybersickness in real time without user-specific calibration. In addition to our method's implementation, we release our dataset of 16 participants to spur future work in this domain.
To use or not to use viewpoint oscillations when walking in VR ? State of the art and perspectives
Yann Moullec
Viewpoint oscillations are periodic changes in the position and/or orientation of the point of view in a virtual environment. They can be implemented in Virtual Reality (VR) walking simulations to make them feel closer to real walking. This is especially useful in simulations where users remain in place because of space or hardware constraints. As for today, it remains unclear what exact benefit they bring to user experience during walking simulations, and with what characteristics they should be implemented. To answer these questions, we conduct a systematic literature review focusing on five main dimensions of user experience (walking sensation, vection, cybersickness, presence and embodiment) and discuss 44 articles from the fields of VR, Vision, and Human-Computer Interaction. Overall, the literature suggests that viewpoint oscillations benefit vection, and with less evidence, walking sensation and presence. As for cybersickness, the literature contains contrasted results. Based on these results, we recommend using viewpoint oscillations in applications that require accurate distance or speed perception, or that aim to provide compelling walking simulations without a walking avatar, and a particular attention should be paid to cybersickness. Taken together, this work gives recommendations for enhancing walking simulations in VR, which may be applied to entertainment, virtual visits, and medical rehabilitation.
Multiple Self-Avatar Effect: Effects of Using Diverse Self-Avatars on Memory Acquisition and Retention of Sign-Language Gestures
Takato Mizuho
This study proposes a new learning method that employs multiple embodied self-avatars during learning, to use the potential benefit of virtual reality (VR) for effective learning and training. In this study, by taking advantage of the benefit of virtual reality (VR), we propose a new learning method that employs multiple embodied self-avatars during learning. Based on the multiple-context effect, which posits that learning in diverse situations can prevent forgetting and enhance memory retention, we conducted a between-participants study under two conditions: the varied avatar condition, in which participants learned sign languages with different self-avatars in six iterations, and the constant avatar condition, in which the same self-avatar was used consistently. We employed sign language as a learning material that naturally draws attention to self-avatars and is suitable for investigating the effects of varying self-avatars. Initially, the varied avatar condition performed worse than the constant avatar condition. However, in a test conducted after one week in the real world, the varied avatar condition showed significantly less forgetting and better retention than the constant avatar condition. Furthermore, our results suggested a positive correlation between the degree of embodiment toward the avatars and the effectiveness of the proposed method. This study presents an innovative design approach for the use of self-avatars in VR-based education.
The Least Increasing Aversion (LIA) Protocol: Illustration on Identifying Individual Susceptibility to Cybersickness Triggers
Nana Tian
This paper introduces the Least Increase aversion (LIA) protocol to investigate the relative impact of factors that may trigger cybersickness. The protocol is inspired by the Subjective Matching methodology (SMT) from which it borrows the incremental construction of a richer VR experience, except that the full-blown target experience may cause undesired discomfort. In the first session, the participant briefly encounter all factors at the maximum level. Then in the second session they start with the minimum level of all factors as a Baseline. Subsequently, we expect the participant to minimize their exposure to the most adverse factors. This approach ranks the factors from mildest to worst and helps detect individual susceptibility to cybersickness triggers.To validate the applicability of LIA protocol, we further evaluate it with an experiment to identify individual susceptibility to three rotational axes (Yaw, Pitch, and Roll). The findings not only confirm the protocol's capability to accurately discern individual rankings of various factors to cybersickness but also indicate that individual susceptibility is more intricate and multifaceted than initially anticipated
Session: Locomotion and Spatial Awareness (12 - B)
Development and Evaluation of a Treadmill-based Video-see-through and Optical-see-through Mixed Reality Systems for Obstacle Negotiation Training
Tamon Miyake
Mixed reality (MR) technologies have a high potential to enhance obstacle negotiation training beyond the capabilities of existing physical systems. Despite such potential, the feasibility of using MR for obstacle negotiation on typical training treadmill systems and its effects on obstacle negotiation performance remains largely unknown. This research bridges this gap by developing an MR obstacle negotiation training system deployed on a treadmill, and implementing two MR systems with a video see-through (VST) and an optical see-through (OST) Head Mounted Displays (HMDs). We investigated the obstacle negotiation performance with virtual and real obstacles. The main outcomes show that the VST MR system significantly changed the parameters of the leading foot in cases of Box obstacle (approximately 22 cm to 30 cm for stepping over 7cm-box), which we believe was mainly attributed to the latency difference between the HMDs. In the condition of OST MR HMD, users tended to not lift their trailing foot for virtual obstacles (approximately 30 cm to 25 cm for stepping over 7cm-box). Our findings indicate that the low-latency visual contact with the world and the user's body is a critical factor for visuo-motor integration to elicit obstacle negotiation.
Preview Teleport: An Occlusion-Free Point-and-Teleport Technique Enhanced with an Augmented Preview
Yen-Ming Huang, National Yang Ming Chiao Tung University; Tzu-Wei Mi, National Yang Ming Chiao Tung University; Liwei Chan, National Chiao Tung University
The Preview Teleport technique in VR integrates a preview window into the point-and-teleport locomotion system, enabling users to see and target destinations obscured by obstacles in one seamless action. Our study assessed the influence of window position, arc type, camera pose, and environment familiarity on user performance. Findings indicate that Preview Teleport enhances efficiency and spatial awareness while reducing fatigue, though it compromises accuracy slightly compared to non-preview methods, which allow for closer adjustments after initial teleportation. Recommendations include attaching the window to the controller, utilizing a user-height camera pose, and employing a standard parabolic arc for optimal performance.
Measuring Human Perception of Airflow for Natural Motion Simulation in Virtual Reality
Yu Cai, Zhejiang University; Sanyi Jin, Zhejiang University; Zihan Chen, ZheJiang University; Daiwei Yang, Zhejiang University; Han Tu, Zhejiang University; Preben Hansen, Stockholm University; Lingyun Sun, Zhejiang University; Liuqing Chen, Zhejiang University
Airflow effectively induces self-motion sensations (vection) and reduces motion sickness in VR. However, the quantitative relationship between airflow stimuli and the user's virtual motion sensations remains unclear. Our experiments revealed a strong linear correlation between airflow speed and virtual speed during linear motion, as well as a relationship similar to centripetal acceleration between the left-right airflow speed difference and the radius and angular velocity of virtual curvilinear motion. We also demonstrated that a dynamic airflow scheme, based on these relationships, provides more effective motion sickness relief and better consistency between visual and bodily vection compared to constant airflow.
Effects of viewpoint oscillations and gaze-based stabilization on walking sensation, embodiment and cybersickness in immersive VR
Yann Moullec, Univ Rennes; Justine Saint-Aubert, CNRS; Inria; IRISA Bretagne-Atlantique; Mélanie Cogne, Inria Bretagne-Atlantique; Anatole Lécuyer, Inria
The head's natural swaying motion when walking can be simulated as 'viewpoint oscillations' to enhance VR walking simulations. In this paper, we provide new experimental work on their effects on walking sensation, cybersickness, and virtual embodiment, and we propose a technical improvement that uses a gaze tracker to add rotations that stabilize the target of the gaze in the users' field of view. We conducted a user study that showed that viewpoint oscillations increased walking sensation, did not impact cybersickness or agency, and increased ownership when stabilized, compared to a linear motion without oscillations. Taken together, our results promote the use of viewpoint oscillations during walking simulations in embodied VR.
Comparative Evaluation of Differing Levels of Information Presentation in 3D Mini-Maps on Spatial Knowledge Acquisition in VR
Pei-Chin Hsu, National Yang Ming Chiao Tung University; Sabarish V. Babu, Clemson University; Jung-Hong Chuang, National Yang Ming Chiao Tung University
Maps are essential for navigation in both physical and VEs, with varying content and appearance. This paper examines three mini-maps: WiM-3DMap, a standard WiM, UC-3DMap, which features landmarks and nearby buildings, and LM-3DMap, which presents only landmarks, alongside a NoMap baseline condition. A comparative study evaluated their effectiveness in spatial updating, landmark recall, landmark placement, and route recall. Results showed that LM-3DMap and UC-3DMap outperformed WiM-3DMap, while NoMap yielded the lowest performance. UC-3DMap excelled in landmark placement due to its inclusion of local building context. These findings highlight the impact of building detail on mini-map effectiveness and the benefits of local context.
Session: User experience and usability (12 - C)
Exploring Multiscale Navigation of Homogeneous and Dense Objects with Progressive Refinement in Virtual Reality
Leonardo Pavanatto, Virginia Tech; Alexander Giovannelli, Virginia Tech; Brian Giera, Lawrence Livermore National Laboratory; Peer-Timo Bremer, Lawrence Livermore National Laboratory; Haichao Miao, Lawrence Livermore National Laboratory; Doug Bowman, Virginia Tech
Locating small features in a large, dense object in virtual reality (VR) poses a significant interaction challenge. Existing multiscale techniques are not focused on handling dense, homogeneous objects with hidden features. We propose an approach that applies progressive refinement to VR navigation, enabling focused inspections. We conducted a user study to better understand the effects of navigation style and display mode on efficiency and awareness during multiscale navigation. Results showed that unstructured navigation can be faster than structured and that displaying only the selection can be faster than displaying the entire object. However, displaying everything can support better location awareness and object understanding.
Single Actor Controlling Multiple Avatars for Social Virtual Realities
Jingyi Zhang, University College London; Anthony Steed, University College London
Social virtual reality applications aim to provide immersive, interactive experiences with virtual characters. However, developing characters capable of natural verbal and non-verbal interactions remains a significant challenge, particularly in managing complex and unexpected user interactions. To address this, we present a system supporting full-body avatars with six-point tracking and a streamlined switch control procedure, enabling one actor to control multiple virtual humans and interact seamlessly with users. The system supports both verbal and non-verbal interactions. In an experiment, we showed that our system enhances co-presence, creating the feeling of multiple distinct, human-controlled characters are present in the scene.
Effects of Proprioceptive Attenuation with Noisy Tendon Electrical Stimulation on Adaptation to Beyond-Real Interaction
Maki Ogawa, the University of Tokyo; Keigo Matsumoto, The University of Tokyo; Kazuma Aoyama, The University of Tokyo; Takuji Narumi, the University of Tokyo
Beyond-real interactions (BRI) offer effective user experiences in VR. However, adapting to novel BRI mappings often reduces performance and embodiment. To enhance adaptation, we propose attenuating proprioception with noisy tendon electrical stimulation (n-TES). A user study evaluated the effects of go-go interaction (a BRI technique) and n-TES in a 2x2 within-participants design. Results from 24 participants showed that go-go interaction reduced the quality of VR experience, but n-TES did not significantly mitigate these negative effects, except that perceived agency was higher with n-TES. The limited impact of n-TES may stem from adaptation during the tasks. Future studies should explore adaptation processes and diverse BRI scenarios.
ReLive: Walking into Virtual Reality Spaces from Video Recordings of One's Past Can Increase the Experiential Detail and Affect of Autobiographical Memories
Valdemar Danry, MIT; Eli Villa, MIT; Samantha W. T. Chan, MIT Media Lab; Pattie Maes, MIT Media Lab
With the rapid development of advanced machine learning methods for spatial reconstruction, it becomes important to understand the psychological and emotional impacts of such technologies on autobiographical memories. In a within-subjects study, we found that allowing users to walk through old spaces reconstructed from their videos significantly enhances their sense of traveling into past memories, increases the vividness of those memories, and boosts their emotional intensity compared to simply viewing videos of the same past events. These findings highlight that, regardless of the technological advancements, the immersive experience of VR can profoundly affect memory phenomenology and emotional engagement.
Analyzing Multimodal Interaction Strategies for LLM-Assisted Manipulation of 3D Scenes
Junlong Chen, University of Cambridge; Jens Grubert, Coburg University of Applied Sciences and Arts; Per Ola Kristensson, University of Cambridge
As more applications of large language models (LLMs) for 3D content in immersive environments emerge, it is crucial to study user behaviour to identify interaction patterns and potential barriers to guide the future design of immersive systems. In an empirical user study with 12 participants, we combine quantitative usage data with post-experience questionnaire feedback to reveal common interaction patterns and key barriers in LLM-assisted 3D scene editing systems. We identify opportunities for improving natural language interfaces in 3D design tools and propose design recommendations. Through an empirical study, we demonstrate that LLM-assisted interactive systems can be used productively in immersive environments.
Session: Teleoperation and telepresence (12 - D)
Intuitive Visual Feedback in Virtual Reality for the Teleoperation of Robots
Paul Christopher Gloumeau, University of Manchester; Stephen Robert Pettifer, University of Manchester
Teleoperation often requires significant skill, limiting its accessibility to average users. Challenges include delays between commands and robot responses, as well as frustration caused by rejected commands due to unattainable goals. Advances in virtual reality (VR) and inverse kinematics are improving usability but leave gaps in understanding robot capabilities and constraints. VR's customisable environments offer unique opportunities for enhanced feedback. This paper introduces three visual feedback techniques to inform users about their commands. A user study found that higher levels of visual feedback reduces task completion time, lowers mental workload, and improves user experience, making teleoperation more effective and accessible.
MRUnion: Asymmetric Task-Aware 3D Mutual Scene Generation of Dissimilar Spaces for Mixed Reality Telepresence
Michael Pabst, TUM; Linda Rudolph, Technical University of Munich; Nikolas Brasch, Technical University of Munich; Verena Biener, University of Stuttgart; Chloe Eghtebas, TUM; Ulrich Eck, Technische Universitaet Muenchen; Dieter Schmalstieg, University of Stuttgart; Gudrun Klinker, Technical University of Munich
In MR telepresence applications, the differences between participants' physical environments can interfere with effective collaboration. For asymmetric tasks, users might need to access different resources distributed throughout their room. Existing intersection methods do not support such interactions, because a large portion of the telepresence participants' rooms become inaccessible, along with the relevant task resources. We propose MRUnion, a Mixed Reality Telepresence pipeline for asymmetric task-aware 3D mutual scene generation. The key concept of our approach is to enable a user in an asymmetric telecollaboration scenario to access the entire room, while still being able to communicate with remote users in a shared space.
Influence of Audiovisual Realism on Communication Behaviour in Group-to-Group Telepresence
Gareth Rendle, Bauhaus-Universität Weimar; Felix Immohr, Technische Universität Ilmenau; Christian Kehling, TU Ilmenau; Anton Benjamin Lammert, Bauhaus-Universität Weimar; Adrian Kreskowski, Bauhaus-Universität Weimar; Karlheinz Brandenburg, TU Ilmenau; Alexander Raake, TU Ilmenau; Bernd Froehlich, Bauhaus-Universität Weimar
Group-to-group telepresence systems allow users to interact with collocated and remote interlocutors, enabling direct comparison of virtual and real user representations. We investigate how visual and aural realism affect communication in such systems and examine how interaction differs between collocated and remote pairs. The results from our experiment, in which groups perform a conversation task, indicate that avatar realism improves perceived message understanding and group cohesion, and yields behavioural differences that indicate more engagement. Comparisons between collocated and remote communication found that collocated interaction was perceived as more effective, but that remote participants received more visual attention.
VASA-Rig: Audio-Driven 3D Facial Animation with `Live' Mood Dynamics in Virtual Reality
Ye Pan, Shanghai Jiaotong University; Chang Liu, Shanghai Jiao Tong University; Sicheng Xu, Microsoft Research Asia; Shuai Tan, Shanghai Jiao Tong University; Jiaolong Yang, Microsoft Research Asia
Audio-driven 3D facial animation is key to enhancing realism and interactivity in the metaverse. While existing methods excel at 2D talking head videos, they lack adaptability to 3D environments. We present VASA-Rig, which advances lip-audio synchronization, facial dynamics, and head movements. Using a novel rig parameter-based emotional talking face dataset, our Latents2Rig model transforms 2D facial animations into 3D. Unlike mesh-based models, VASA-Rig outputs 174 Metahuman rig parameters, ensuring compatibility with industry-standard pipelines. Experiments show that VASA-Rig surpasses state-of-the-art methods in realism and accuracy, offering a robust solution for 3D animation in interactive virtual environments.
SensARy Substitution: Augmented Reality Techniques to Enhance Force Perception in Touchless Robot Control
Tonia Mielke, Otto-von-Guericke University Magdeburg; Florian Heinrich, Otto von Guericke University Magdeburg; Christian Hansen, Faculty of Computer Science
The lack of haptic feedback in touchless human-robot interaction is critical in applications like robotic ultrasound, where force perception is essential. Augmented reality (AR) can address this by providing sensory substitution through visual or vibrotactile feedback. We implemented two visualization types at three positions and investigated the effects of vibrotactile feedback. Additionally, we compared multimodal feedback with visual or vibrotactile feedback alone. Our results show that sensory substitution improves interaction, with visual support reducing force errors and being preferred by participants. A 2D bar visualization outperformed a 3D arrow, and visualizations on the ultrasound screen were favored.
Session: Immersive applications and 3DUI (13 - B)
Investigating the Impact of Voice-only and Embodied Conversational Virtual Agents on Mixed Reality Puzzle Solving
Shirin Hajahmadi, Universita di Bologna; Pasquale Cascarano, Universita di Bologna; Fariba Mostajeran, Universität Hamburg; Kevin Heuer, Universität Hamburg; Anton Lux, Universität Hamburg; Gil Otis Mends-Cole, Universität Hamburg; Frank Steinicke, Universität Hamburg; Gustavo Marfia, Universita di Bologna
This paper explores the integration of a Conversational Virtual Agent (CVA) into a Mixed Reality application designed to assist in solving a 2D physical puzzle. The agent works as a state-aware assistant to guide users on demand. We conducted a study with 34 participants to investigate the influence of Voice-only and Embodied CVAs on puzzle-solving performance and user experience. Both modalities showed equivalent outcomes in terms of number of queries. However, results revealed that voice-only CVA may be more effective since it significantly enhanced task efficiency and reported lower effort than embodied condition.
ResponsiveView: Enhancing 3D Artifact Viewing Experience in VR Museums
Xueqi Wang, Xi'an Jiaotong-Liverpool University; Yue Li, Xi'an Jiaotong-Liverpool University; Boge Ling, University of British Columbia; Han-Mei Chen, University of Liverpool; Hai-Ning Liang, The Hong Kong University of Science and Technology (Guangzhou)
We collect empirical data on user-determined parameters for the optimal viewing experience in VR museums. By modeling users' viewing behaviors, we derive predictive functions that configure the pedestal height, calculate the optimal viewing distance, and adjust the appropriate handheld scale for the optimal viewing experience. This led to our novel 3D responsive design, ResponsiveView. Similar to the responsive web design that automatically adjusts for different screen sizes, ResponsiveView automatically adjusts the parameters in the VR environment to facilitate users' viewing experience. The design has been validated with popular inputs available in current commercial VR devices, demonstrating enhanced viewing experience in VR museums.
Examining the Design Process for 3D Interactions in Performing Arts : A Spatial Augmented Reality Cyber-Opera Case Study
While 3D user interfaces are often designed with efficiency and accuracy in mind, artistic performances have their own very specific constraints and criteria for a successful interaction in mixed or virtual reality, which have yet to be fully understood. In this paper, we study the design of 3D interactions for a Spatial Augmented-Reality display in the context of a cyber-opera. We perform an analysis of design decisions taken during the
Daddy Long Legs: A Scale and Speed Up Virtual Reality Locomotion Technique for Medium-Scale Scenarios
Yue Zhao, Human Interface Technology Laboratory New Zealand; Robert W. Lindeman, University of Canterbury; Thammathip Piumsomboon, University of Canterbury
Locomotion techniques are essential to enhance the sense of presence or -being there- in Virtual Reality (VR). This research proposes a novel technique, -Daddy Long Legs- (DLL), for navigating medium-scale virtual environments (VE). DLL builds on principles from Seven-League Boots (7LB) and Ground-Level Scaling (GLS), aiming to enhance positional accuracy by offering an elevated perspective while preserving immersion through the maintenance of the user's interpupillary distance. A user study (n = 24) compared DLL, 7LB, and GLS in comfort, embodiment, workload, and walking behavior. DLL outperformed 7LB across measures, while GLS, noted for near-natural walking behavior, received the most favorable feedback and was the preferred method.
Scaling Techniques for Exocentric Navigation Interfaces in Multiscale Virtual Environments
Jong-In Lee, Texas A&M University; Wolfgang Stuerzlinger, Simon Fraser University
We present a scroll-based scaling technique for navigating multiscale virtual environments across different levels of scale (LoS). While previous work has explored various scaling techniques, our approach focuses on optimizing exocentric navigation where precise, continuous scaling is essential. Through user testing, we examined two key design components of scaling techniques: input method and scaling center. Results show our scroll-based input method significantly outperforms traditional pinch/grab based bi-manual techniques in accuracy, efficiency, and task load. Additionally, we discovered that scaling center selection impacts usability, particularly when combined with specific input methods.
Session: Computer graphics techniques (13 - D)
360o 3D Photos from a Single 360o Input Image
Manuel Rey-Area, University of Bath; Christian Richardt, Meta Reality Labs
360° images are a popular medium for bringing photography into virtual reality. While users can look in any direction by rotating their heads, 360° images ultimately look flat. That is because they lack depth and thus cannot create motion parallax when translating the head. To achieve a fully immersive VR experience from a single 360° image, we introduce a novel method to upgrade 360° images to free-viewpoint renderings with 6 degrees of freedom based on 3D Gaussian splatting. To fill plausible content in previously unseen regions, we introduce a novel combination of latent diffusion inpainting and monocular depth estimation with Poisson-based blending. Our results demonstrate state-of-the-art visual quality and real-time rendering speed.
Fov-GS: Foveated 3D Guassian Splatting for Dynamic Scene
Runze Fan, Beihang University; Jian Wu, Beihang University; Xuehuai Shi, Nanjing University of Posts and Telecommunications; Lizhi Zhao, Beihang University; Qixiang Ma, Beihang University; Lili Wang, Beihang University
3D Gaussian Splatting-based methods can achieve photo-realistic rendering with speeds of over 100 fps in static scenes, but the speed drops below 10 fps in monocular dynamic scenes. Foveated rendering provides a possible solution to accelerate rendering without compromising visual perceptual quality. However, 3DGS and foveated rendering are not compatible. In this paper, we propose Fov-GS, a foveated 3D Gaussian splatting method for rendering dynamic scenes in real time. Experiments demonstrate that our method not only achieves higher rendering quality in the foveal and salient regions compared to the SOTA methods but also dramatically improves rendering performance, achieving up to $11.33X$ speedup.
Accelerating Stereo Rendering via Image Reprojection and Spatio-Temporal Supersampling
Sipeng Yang, State Key Lab of CAD&CG; Junhao Zhuge, State Key Lab of CAD&CG; Jiayu Ji, State Key Lab of CAD&CG; Qingchuan Zhu, State Key Lab of CAD&CG; Xiaogang Jin, State Key Lab of CAD&CG
This paper introduces a lightweight supersampling framework that integrates image projection with spatio-temporal supersampling to accelerate stereo rendering in VR. Leveraging temporal and spatial redundancies in stereo videos, our approach rapidly generates high-resolution (HR) frames by blending rendered low-resolution (LR) frames with temporal samples. These HR frames are then reprojected to synthesize new images for opposite viewpoints. To address disocclusions, we utilize history data and low-pass filtering, ensuring high-quality, resolution-enhanced, and anti-aliased images with minimal delay. Extensive evaluations confirm the framework's efficiency on both PC and standalone VR devices, making it suitable for various VR platforms.
Multi-Layer Gaussian Splatting for Immersive Anatomy Visualization
Constantin Kleinbeck, Technical University of Munich; Hannah Schieber, Technical University of Munich (TUM); Klaus Engel, Siemens Healthineers; Ralf Gutjahr, Siemens Healthineers; Daniel Roth, Technical University of Munich
Path tracing of volumetric medical data produces lifelike visualizations. VR enhances this rendering approach by enabling spatial exploration beyond traditional 2D diagnostics. We propose a Gaussian Splatting (GS)-based method to create an efficient intermediate representations, delivering path tracing quality to real-time immersive displays. Our layered GS structure incrementally includes anatomical features, reducing storage and computational demands while preserving some interactivity in otherwise static GS models.
Saliency-aware Foveated Path Tracing for Virtual Reality Rendering
Yang Gao, Beihang university; Wencan Li, Beihang University; Shiyu Liang, Beihang University; Aimin Hao, Beihang University; Xiaohui Tan, Capital Normal University
Foveated rendering reduces computational load by distributing resources based on the human visual system. However, traditional foveation methods based on eccentricity cannot account for the complex behavior of visual attention. This is one of the reasons for lower perceived quality. In this study, we introduce a pipeline that incorporates ocular attention through visual saliency. Based on saliency, our approach facilitates the real-time production of high-quality images utilizing path tracing. To further augment image quality, an adaptive filtering process is employed to reduce artifacts in non-foveal regions. Our experiments prove that our approach has superior performance both in terms of quantitative metrics and perceived quality.
Conference Sponsors
Supporters
©IEEE VR Conference 2025, Sponsored by the IEEE Computer Society and the Visualization and Graphics Technical Committee