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:20260114T163643Z
LOCATION:Meeting Room C4.11\, Level 4 (Convention Centre)
DTSTART;TZID=Australia/Melbourne:20231214T094000
DTEND;TZID=Australia/Melbourne:20231214T095000
UID:siggraphasia_SIGGRAPH Asia 2023_sess124_papers_381@linklings.com
SUMMARY:AniPortraitGAN: Animatable 3D Portrait Generation from 2D Image Co
 llections
DESCRIPTION:Yue Wu (Hong Kong University of Science and Technology), Siche
 ng Xu (Microsoft Research Asia), Jianfeng Xiang (Tsinghua University), Fan
 gyun Wei (Microsoft Research Asia), Qifeng Chen (Hong Kong University of S
 cience and Technology), and Jiaolong Yang and Xin Tong (Microsoft Research
  Asia)\n\nPrevious animatable 3D-aware GANs for human generation have prim
 arily focused on either the human head or full body. However, head-only vi
 deos are relatively uncommon in real life, and full body generation typica
 lly does not deal with facial expression control and still has challenges 
 in generating high-quality results. Towards applicable video avatars, we p
 resent an animatable 3D-aware GAN that generates portrait images with cont
 rollable facial expression, head pose, and shoulder movements. It is a gen
 erative model trained on unstructured 2D image collections without using 3
 D or video data. For the new task, we base our method on the generative ra
 diance manifold representation and equip it with learnable facial and head
 -shoulder deformations. A dual-camera rendering and adversarial learning s
 cheme is proposed to improve the quality of the generated faces, which is 
 critical for portrait images. A pose deformation processing network is dev
 eloped to generate plausible deformations for challenging regions such as 
 long hair. Experiments show that our method, trained on unstructured 2D im
 ages, can generate diverse and high-quality 3D portraits with desired cont
 rol over different properties.\n\nRegistration Category: Full Access\n\nSe
 ssion Chair: Lin Gao (University of Chinese Academy of Sciences)\n\n
URL:https://asia.siggraph.org/2023/full-program?id=papers_381&sess=sess124
END:VEVENT
END:VCALENDAR