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:20250110T023312Z
LOCATION:Hall B5 (2)\, B Block\, Level 5
DTSTART;TZID=Asia/Tokyo:20241204T170500
DTEND;TZID=Asia/Tokyo:20241204T171600
UID:siggraphasia_SIGGRAPH Asia 2024_sess122_papers_369@linklings.com
SUMMARY:TEXGen: a Generative Diffusion Model for Mesh Textures
DESCRIPTION:Technical Papers\n\nXin Yu (University of Hong Kong); Ze Yuan 
 (Beihang University); Yuan-Chen Guo (VAST); Ying-Tian Liu (Tsinghua Univer
 sity); Jianhui Liu (University of Hong Kong); Yangguang Li, Yan-Pei Cao, a
 nd Ding Liang (VAST); and Xiaojuan Qi (University of Hong Kong)\n\nWhile h
 igh-quality texture maps are essential for realistic 3D asset rendering, f
 ew studies have explored learning directly in the texture space, especiall
 y on large-scale datasets. In this work, we depart from the conventional a
 pproach of relying on pre-trained 2D diffusion models for test-time optimi
 zation of 3D textures. Instead, we focus on the fundamental problem of lea
 rning in the UV texture space itself. For the first time, we train a large
  diffusion model capable of directly generating high-resolution texture ma
 ps in a feed-forward manner.\nTo facilitate efficient learning in high-res
 olution UV spaces, we propose a scalable network architecture that interle
 aves convolutions on UV maps with attention layers on point clouds. Levera
 ging this architectural design, we train a 700 million parameter diffusion
  model that can generate UV texture maps guided by text prompts and single
 -view images. Once trained, our model naturally supports various extended 
 applications, including text-guided texture inpainting, sparse-view textur
 e completion, and text-driven texture synthesis.\n\nRegistration Category:
  Full Access, Full Access Supporter\n\nLanguage Format: English Language\n
 \nSession Chair: Minhyuk Sung (Korea Advanced Institute of Science and Tec
 hnology (KAIST))
URL:https://asia.siggraph.org/2024/program/?id=papers_369&sess=sess122
END:VEVENT
END:VCALENDAR