BEGIN:VCALENDAR
VERSION:2.0
PRODID:Linklings LLC
BEGIN:VTIMEZONE
TZID:Asia/Hong_Kong
X-LIC-LOCATION:Asia/Hong_Kong
BEGIN:STANDARD
TZOFFSETFROM:+0800
TZOFFSETTO:+0800
TZNAME:HKT
DTSTART:19911015T033000
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
DTSTAMP:20251218T030656Z
LOCATION:Meeting Room S423+S424\, Level 4
DTSTART;TZID=Asia/Hong_Kong:20251216T140400
DTEND;TZID=Asia/Hong_Kong:20251216T141500
UID:siggraphasia_SIGGRAPH Asia 2025_sess121_tog_122@linklings.com
SUMMARY:Self-supervised Texture Filtering
DESCRIPTION:Hao Jiang and Rongjia Zheng (Sun Yat-sen University), Yongwei 
 Nie (South China University of Technology), Chunxia Xiao (Wuhan University
 ), and Qing Zhang (Sun Yat-sen University)\n\nDecomposing an image Iinto t
 he combination of structure S and texture T components is an important pro
 blem in computational photography and image analysis. Traditional solution
 s are basically non-learning based, because it is difficult to construct d
 atasets containing ground-truth decompositions or find effective structure
 /texture supervisions. In this article, we present a self-supervised frame
 work for smoothing out textures while maintaining the image structures. At
  the core of our method is a texture-inversion observation - if structure 
 S and texture T are well disentangled, then S-T will produce a texture-inv
 erted image that is symmetric to the input image I= S + T and the two will
  be visually highly similar, while for other conditions that structure and
  texture are not effectively separated, the generated texture-inverted ima
 ges will be less similar to the input. Based on the observation, we propos
 e to learn texture fltering from unlabeled data by encouraging the texture
  inverted image generated from the fltering output to be visually more sim
 ilar to the input via contrastive learning. Experiments show that our meth
 od can robustly produce high-quality texture smoothing results, and also e
 nables various applications.\n\nRegistration Category: Full Access, Full A
 ccess Supporter\n\nSession Chair: Paul Debevec (Eyeline, USC Institute for
  Creative Technologies (ICT))\n\n
END:VEVENT
END:VCALENDAR