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:20260817T171534Z
LOCATION:Hall B5 (1)\, B Block\, Level 5
DTSTART;TZID=Asia/Tokyo:20241206T111900
DTEND;TZID=Asia/Tokyo:20241206T113100
UID:siggraphasia_SIGGRAPH Asia 2024_sess142_papers_205@linklings.com
SUMMARY:DreamUDF: Generating Unsigned Distance Fields from A Single Image
DESCRIPTION:Yu-Tao Liu and Xuan Gao (Institute of Computing Technology, Ch
 inese Academy of Sciences; University of Chinese Academy of Sciences); Wei
 kai Chen (Tencent Games); Jie Yang (Institute of Computing Technology, Chi
 nese Academy of Sciences; University of Chinese Academy of Sciences); Xiao
 xu Meng and Bo Yang (Tencent Games); and Lin Gao (Institute of Computing T
 echnology, Chinese Academy of Sciences; University of Chinese Academy of S
 ciences)\n\nRecent advances in diffusion models and neural implicit surfac
 es have shown promising progress in generating 3D models. However, existin
 g generative frameworks are limited to closed surfaces, failing to cope wi
 th a wide range of commonly seen shapes that have open boundaries. In this
  work, we present DreamUDF, a novel framework for generating high-quality 
 3D objects with arbitrary topologies from a single image. To address the c
 hallenge of generating proper topology given sparse and ambiguous observat
 ions, we propose to incorporate both the data priors from a multi-view dif
 fusion model and the geometry priors brought by an unsiged distance field 
 (UDF) reconstructor. In particular, we leverage a joint framework that con
 sists of 1) a generative module that produces a neural radiance field that
  provides photo-realistic renderings from the arbitrary view; and 2) a rec
 onstructive module that distills the learnable radiance field into surface
 s with arbitrary topologies. We further introduce a field coupler that bri
 dges the radiance field and UDF under an novel optimization scheme. This a
 llows the two modules to mutually boost each other during training. Extens
 ive experiments and evaluations demonstrate that DreamUDF achieves high-qu
 ality reconstruction and robust 3D generation on both closed and open surf
 aces with arbitrary topologies, compared to the previous works.\n\nRegistr
 ation Category: Full Access, Full Access Supporter\n\nLanguage Format: Eng
 lish Language\n\nSession Chair: Maria Larsson (The University of Tokyo)\n\
 n
URL:https://asia.siggraph.org/2024/program/?id=papers_205&sess=sess142
END:VEVENT
END:VCALENDAR