BEGIN:VCALENDAR
VERSION:2.0
PRODID:Linklings LLC
BEGIN:VTIMEZONE
TZID:Asia/Seoul
X-LIC-LOCATION:Asia/Seoul
BEGIN:STANDARD
TZOFFSETFROM:+0900
TZOFFSETTO:+0900
TZNAME:KST
DTSTART:18871231T000000
DTSTART:19881009T020000
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
DTSTAMP:20230103T035311Z
LOCATION:Room 324\, Level 3\, West Wing
DTSTART;TZID=Asia/Seoul:20221208T153000
DTEND;TZID=Asia/Seoul:20221208T170000
UID:siggraphasia_SIGGRAPH Asia 2022_sess166_papers_118@linklings.com
SUMMARY:NeRFFaceEditing: Disentangled Face Editing in Neural Radiance Fiel
 ds
DESCRIPTION:Technical Communications, Technical Papers\n\nNeRFFaceEditing:
  Disentangled Face Editing in Neural Radiance Fields\n\nJiang, Chen, Liu, 
 Fu, Gao\n\nRecent methods for synthesizing 3D-aware face images have achie
 ved rapid development thanks to neural radiance fields, allowing for high 
 quality and fast inference speed. However, existing solutions for editing 
 facial geometry and appearance independently usually require retraining an
 d are not optimized for the recent work of generation, thus tending to lag
  behind the generation process. To address these issues, we introduce NeRF
 FaceEditing, which enables editing and decoupling geometry and appearance 
 in the pretrained tri-plane-based neural radiance field while retaining it
 s high quality and fast inference speed. Our key idea for disentanglement 
 is to use the statistics of the tri-plane to represent the high-level appe
 arance of its corresponding facial volume. Moreover, we leverage a generat
 ed 3D-continuous semantic mask as an intermediary for geometry editing. We
  devise a geometry decoder (whose output is unchanged when the appearance 
 changes) and an appearance decoder. The geometry decoder aligns the origin
 al facial volume with the semantic mask volume. We also enhance the disent
 anglement by explicitly regularizing rendered images with the same appeara
 nce but different geometry to be similar in terms of color distribution fo
 r each facial component separately. Our method allows users to edit via se
 mantic masks with decoupled control of geometry and appearance. Both quali
 tative and quantitative evaluations show the superior geometry and appeara
 nce control abilities of our method compared to existing and alternative s
 olutions.\n\nRegistration Category: FULL ACCESS, ON-DEMAND ACCESS\n\nLangu
 age: ENGLISH\n\nFormat: IN-PERSON, ON-DEMAND
URL:https://sa2022.siggraph.org/en/full-program/?id=papers_118&sess=sess16
 6
END:VEVENT
END:VCALENDAR