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_214@linklings.com
SUMMARY:Stitch it in Time: GAN-Based Facial Editing of Real Videos
DESCRIPTION:Technical Communications, Technical Papers\n\nStitch it in Tim
 e: GAN-Based Facial Editing of Real Videos\n\nTzaban, Mokady, Gal, Bermano
 , Cohen-Or\n\nThe ability of Generative Adversarial Networks to encode ric
 h semantics within their latent space has been widely adopted for facial i
 mage editing. However, replicating their success with videos has proven ch
 allenging. Applying StyleGAN editing over real videos introduces two main 
 challenges: (i) StyleGAN operates over aligned crops. When editing videos,
  these crops need to be pasted back into the frame, resulting in a spatial
  inconsistency. (ii) Videos introduce a fundamental barrier to overcome - 
 temporal coherency. To address the first challenge, we propose a novel sti
 tching-tuning procedure. The generator is carefully tuned to overcome the 
 spatial artifacts at crop borders, resulting in a smooth transition even w
 hen difficult backgrounds are involved. Turning to temporal coherence, we 
 propose that this challenge is largely artificial. The source video is alr
 eady temporally coherent, and deviations arise in part due to the careless
  treatment of individual components in the editing pipeline. We leverage t
 he natural alignment of StyleGAN and the tendency of neural networks to le
 arn low-frequency functions, and demonstrate that they provide a strongly 
 consistent prior. These components are combined in an end-to-end framework
  for semantic editing of facial videos. We compare our pipeline to the cur
 rent state-of-the-art and demonstrate significant improvements. Our method
  produces various meaningful manipulations and maintain greater spatial an
 d temporal consistency, even in challenging talking head videos which curr
 ent methods struggle with. Our code will be made publicly available.\n\nRe
 gistration Category: FULL ACCESS, ON-DEMAND ACCESS\n\nLanguage: ENGLISH\n\
 nFormat: IN-PERSON, ON-DEMAND
URL:https://sa2022.siggraph.org/en/full-program/?id=papers_214&sess=sess16
 6
END:VEVENT
END:VCALENDAR