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:20250110T023309Z
LOCATION:Hall B7 (1)\, B Block\, Level 7
DTSTART;TZID=Asia/Tokyo:20241203T135800
DTEND;TZID=Asia/Tokyo:20241203T140900
UID:siggraphasia_SIGGRAPH Asia 2024_sess105_tog_107@linklings.com
SUMMARY:Identity-Preserving Face Swapping via Dual Surrogate Generative Mo
 dels
DESCRIPTION:Technical Papers\n\nZiyao Huang and Fan Tang (Institute of Com
 puting Technology, Chinese Academy of Sciences); Yong Zhang (Tencent); Jua
 n Cao, Chengyu Li, Sheng Tang, and Jintao Li (Institute of Computing Techn
 ology, Chinese Academy of Sciences); and Tong-Yee Lee (National Cheng Kung
  University)\n\nIn this study, we revisit the fundamental setting of face-
 swapping models and reveal that only using implicit supervision for traini
 ng leads to the difficulty of advanced methods to preserve the source iden
 tity. We propose a novel reverse pseudo-input generation approach to offer
  supplemental data for training face-swapping models, which addresses the 
 aforementioned issue. Unlike the traditional pseudo-label-based training s
 trategy, we assume that arbitrary real facial images could serve as the gr
 ound-truth outputs for the face-swapping network and try to generate corre
 sponding input <source, target> pair data.  Specifically, we involve a sou
 rce-creating surrogate that alters the attributes of the real image while 
 keeping the identity, and a target-creating surrogate intends to synthesiz
 e attribute-preserved target images with different identities. Our framewo
 rk, which utilizes proxy-paired data as explicit supervision to direct the
  face-swapping training process, partially fulfills a credible and effecti
 ve optimization direction to boost the identity-preserving capability. We 
 design explicit and implicit adaption strategies to better approximate the
  explicit supervision for face swapping.\nQuantitative and qualitative exp
 eriments on FF++, FFHQ, and wild images show that our framework could impr
 ove the performance of various face-swapping pipelines in terms of visual 
 fidelity and ID preserving. Furthermore, we display applications with our 
 method on re-aging, swappable attribute customization, cross-domain, and v
 ideo face swapping. Code is available under https://github.com/ICTMCG/CSCS
 .\n\nRegistration Category: Full Access, Full Access Supporter\n\nLanguage
  Format: English Language\n\nSession Chair: Kfir Aberman (Snap)
URL:https://asia.siggraph.org/2024/program/?id=tog_107&sess=sess105
END:VEVENT
END:VCALENDAR