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:20250110T023313Z
LOCATION:Hall B5 (2)\, B Block\, Level 5
DTSTART;TZID=Asia/Tokyo:20241206T130000
DTEND;TZID=Asia/Tokyo:20241206T131100
UID:siggraphasia_SIGGRAPH Asia 2024_sess146_papers_359@linklings.com
SUMMARY:Correlation-aware Encoder-Decoder with Adapters for SVBRDF Acquisi
 tion
DESCRIPTION:Technical Papers\n\nDi Luo and Hanxiao Sun (Nankai University)
 , Lei Ma (Peking University), Jian Yang (Nankai University), and Beibei Wa
 ng (Nanjing University)\n\nCapturing materials from the real world avoids 
 laborious manual material authoring. However, recovering high-fidelity Spa
 tially Varying Bidirectional Reflectance Distribution Function (SVBRDF) ma
 ps from a few captured images is challenging due to its ill-posed nature. 
 Existing approaches have made extensive efforts to alleviate this ambiguit
 y issue by leveraging generative models with latent space optimization or 
 extracting features with variant encoder-decoders. Albeit the rendered ima
 ges at input views can match input images, the problematic decomposition a
 mong maps leads to significant differences when rendered under novel views
 /lighting. We observe that for human eyes, besides individual images, the 
 correlation (or the highlights variation) among input images also serves a
 s an important hint to recognize the materials of objects. Hence, our key 
 insight is to explicitly model this correlation in the SVBRDF acquisition 
 network. To this end, we propose a correlation-aware encoder-decoder netwo
 rk to model the correlation features among the input images via a graph co
 nvolutional network by treating channel features from each image as a grap
 h node. This way, the ambiguity among the maps has been reduced significan
 tly. However, several SVBRDF maps still tend to be over-smooth, leading to
  a mismatch in the novel-view rendering. The main reason is the uneven upd
 ate of different maps caused by a single decoder for map interpretation. T
 o address this issue, we further design an adapter-equipped decoder consis
 ting of a main decoder and four tiny per-map adapters, where adapters are 
 employed for individual maps interpretation, together with fine-tuning, to
  enhance flexibility. As a result, our framework allows the optimization o
 f the latent space with the input image feature embeddings as the initial 
 latent vector and the fine-tuning of per-map adapters. Consequently, our m
 ethod can outperform existing approaches both visually and quantitatively 
 on synthetic and real data.\n\nRegistration Category: Full Access, Full Ac
 cess Supporter\n\nLanguage Format: English Language\n\nSession Chair: Vale
 ntin Deschaintre (Adobe Research)
URL:https://asia.siggraph.org/2024/program/?id=papers_359&sess=sess146
END:VEVENT
END:VCALENDAR