BEGIN:VCALENDAR
VERSION:2.0
PRODID:Linklings LLC
BEGIN:VTIMEZONE
TZID:Australia/Melbourne
X-LIC-LOCATION:Australia/Melbourne
BEGIN:DAYLIGHT
TZOFFSETFROM:+1000
TZOFFSETTO:+1100
TZNAME:AEDT
DTSTART:19721003T020000
RRULE:FREQ=YEARLY;BYMONTH=4;BYDAY=1SU
END:DAYLIGHT
BEGIN:STANDARD
DTSTART:19721003T020000
TZOFFSETFROM:+1100
TZOFFSETTO:+1000
TZNAME:AEST
RRULE:FREQ=YEARLY;BYMONTH=10;BYDAY=1SU
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
DTSTAMP:20240214T070244Z
LOCATION:Meeting Room C4.9+C4.10\, Level 4 (Convention Centre)
DTSTART;TZID=Australia/Melbourne:20231213T140000
DTEND;TZID=Australia/Melbourne:20231213T141000
UID:siggraphasia_SIGGRAPH Asia 2023_sess164_papers_183@linklings.com
SUMMARY:Towards Practical Capture of High-Fidelity Relightable Avatars
DESCRIPTION:Technical Papers, TOG\n\nHaotian Yang, Mingwu Zheng, Wanquan F
 eng, and Haibin Huang (Kuaishou Technology); Yu-Kun Lai (Cardiff Universit
 y); and Pengfei Wan, Zhongyuan Wang, and Chongyang Ma (Kuaishou Technology
 )\n\nIn this paper, we propose a novel framework, Tracking-free Relightabl
 e Avatar (TRAvatar), for capturing and reconstructing high-fidelity 3D ava
 tars. Compared to previous methods, TRAvatar works in a more practical and
  efficient setting. Specifically, TRAvatar is trained with dynamic image s
 equences captured in a Light Stage under varying lighting conditions, enab
 ling realistic relighting and real-time animation for avatars in diverse s
 cenes. Additionally, TRAvatar allows for tracking-free avatar capture and 
 obviates the need for accurate surface tracking under varying illumination
  conditions. Our contributions are two-fold: First, we propose a novel net
 work architecture that explicitly builds on and ensures the satisfaction o
 f the linear nature of lighting. Trained on simple group light captures, T
 RAvatar can predict the appearance in real-time with a single forward pass
 , achieving high-quality relighting effects under illuminations of arbitra
 ry environment maps. Second, we jointly optimize the facial geometry and r
 elightable appearance from scratch based on image sequences, where the tra
 cking is implicitly learned. This tracking-free approach brings robustness
  for establishing temporal correspondences between frames under different 
 lighting conditions. Extensive qualitative and quantitative experiments de
 monstrate that our framework achieves superior performance for photorealis
 tic avatar animation and relighting.\n\nRegistration Category: Full Access
 \n\nSession Chair: Parag Chaudhuri (Indian Institute of Technology Bombay)
URL:https://asia.siggraph.org/2023/full-program?id=papers_183&sess=sess164
END:VEVENT
END:VCALENDAR