제 목 Two studies on visual perception of VR users: Foveated rendering and light field displays
강 사 김주환 박사 (NVIDIA, USA)
시 간 2017년 9월 5일 (화요일) 오전 10시
장 소 제1신공학관 301동 1112호
Abstract
I would like to present two exciting findings on cutting-edge technologies for VR, namely foveated rendering and light field displays. The abstracts for the two studies are shown below:
1. Latency requirements for foveated rendering in virtual reality
Foveated rendering reduces rendering cost by omitting imperceptible details in the peripheral view. The technique will be particularly useful for virtual reality (VR) devices because the view provided by VR devices are mostly peripheral. However, it may require fast and accurate gaze tracking to ensure correct foveation of the scene even when the viewer makes fast and large eye movements such as saccades. On the other hand, the latency requirements may be relaxed due to the phenomenon called saccadic omission (the temporary masking of visual input around the time of making saccades). In this talk, we investigate the latency requirements for foveated rendering in VR. We conducted psyhophysical experiments where we measured the visibility of visual artifacts using three methods of foveation and six levels of gaze tracker latency. Our results show that 50~70 ms of end-to-end system latency could be tolerated regardless of foveation methods.
2. Requirements on spatial and angular resolution for light field displays to drive accommodation
Super-multi view (SMV) is a condition where a light field display provides two or more parallax views to a single pupil of the viewer. Several studies presented experimental results showing that a light field display can drive accommodation when the SMV condition is met. These findings are of critical importance for three dimensional (3D) displays because one can avoid vergence-accommodation conflict, which is a major cause of visual discomfort of 3D display viewers. However, several studies in vision science literature suggest that spatial resolution may also matter to the effectiveness of focal cue. Unfortunately, no study had systematically examined the effect of spatial resolution on the quality of focal cues provided by a light field display. In this talk, we will analyze the quality of focal cue of a light field display when both spatial and angular resolution changes. The analysis shows that both spatial and angular resolution plays important roles in creating an effective focal cue. Biography Joohwan Kim received his B.S. degree in electrical engineering from Seoul National University in 2003, and his Ph.D. degree in electrical engineering and computer science from Seoul National University in 2009. After working as a postdoctoral researcher from 2009 to 2015 at the University of California, Berkeley, he now works at Nvidia as a research scientist. His research interests include visual perception, visual discomfort, and novel displays.
댓글 0
번호 | 제목 | 글쓴이 | 날짜 | 조회 수 |
---|---|---|---|---|
3601 | 제 712차 그룹미팅 | 최태원 | 2023.12.14 | 38 |
3600 | 제 711차 그룹미팅 | 최태원 | 2023.12.06 | 33 |
3599 | 제 710차 그룹미팅 | 최태원 | 2023.11.29 | 30 |
3598 | 제 709차 그룹미팅 | 최태원 | 2023.11.15 | 40 |
3597 | 김동연 박사 청첩장(2023년 11월 25일 토요일 오후 2시 30분, 서울대학교 교수회관 2층) | 관리자 | 2023.11.15 | 426 |
3596 | 제 708차 그룹미팅 | 최태원 | 2023.11.09 | 31 |
3595 | 서울대학교 전기정보공학부 이병호홀 개관식 및 1주기 추모식 | 최태원 | 2023.11.06 | 176 |
3594 | 제 707차 그룹미팅 | 최태원 | 2023.11.01 | 30 |
3593 | 제 706차 그룹미팅 | 최태원 | 2023.10.18 | 36 |
3592 | 제 705차 그룹미팅 | 최태원 | 2023.10.04 | 34 |