BEGIN:VCALENDAR
VERSION:2.0
PRODID:Linklings LLC
BEGIN:VTIMEZONE
TZID:Asia/Tokyo
X-LIC-LOCATION:Asia/Tokyo
BEGIN:STANDARD
TZOFFSETFROM:+0900
TZOFFSETTO:+0900
TZNAME:JST
DTSTART:18871231T000000
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
DTSTAMP:20250110T023313Z
LOCATION:Hall B5 (2)\, B Block\, Level 5
DTSTART;TZID=Asia/Tokyo:20241206T114300
DTEND;TZID=Asia/Tokyo:20241206T115400
UID:siggraphasia_SIGGRAPH Asia 2024_sess143_papers_796@linklings.com
SUMMARY:URAvatar: Universal Relightable Gaussian Codec Avatars
DESCRIPTION:Technical Papers\n\nJunxuan Li, Chen Cao, Gabriel Schwartz, Ra
 wal Khirodkar, Christian Richardt, Tomas Simon, Yaser Sheikh, and Shunsuke
  Saito (Reality Labs Research)\n\nWe present a new approach to creating ph
 otorealistic and relightable head avatars from a phone scan with unknown i
 llumination. The reconstructed avatars can be animated and relit in real t
 ime with the global illumination of diverse environments. Unlike existing 
 approaches that estimate parametric reflectance parameters via inverse ren
 dering, our approach directly models learnable radiance transfer that inco
 rporates global light transport in an efficient manner for real-time rende
 ring. However, learning such a complex light transport that can generalize
  across identities is non-trivial. A phone scan in a single environment la
 cks sufficient information to infer how the head would appear in general e
 nvironments. To address this, we build a universal relightable avatar mode
 l represented by 3D Gaussians. We train on hundreds of high-quality multi-
 view human scans with controllable point lights.\nHigh-resolution geometri
 c guidance further enhances the reconstruction accuracy and generalization
 . Once trained, we finetune the pretrained model on a phone scan using inv
 erse rendering to obtain a personalized relightable avatar. Our experiment
 s establish the efficacy of our design, outperforming existing approaches 
 while retaining real-time rendering capability.\n\nRegistration Category: 
 Full Access, Full Access Supporter\n\nLanguage Format: English Language\n\
 nSession Chair: Iain Matthews (Epic Games, Carnegie Mellon University)
URL:https://asia.siggraph.org/2024/program/?id=papers_796&sess=sess143
END:VEVENT
END:VCALENDAR