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:20260817T171530Z
LOCATION:Hall B5 (2)\, B Block\, Level 5
DTSTART;TZID=Asia/Tokyo:20241204T164100
DTEND;TZID=Asia/Tokyo:20241204T165300
UID:siggraphasia_SIGGRAPH Asia 2024_sess122_papers_535@linklings.com
SUMMARY:Text-Guided Texturing by Synchronized Multi-View Diffusion
DESCRIPTION:Yuxin Liu and Minshan Xie (Chinese University of Hong Kong); H
 anyuan Liu (City University of Hong Kong); and Tien-Tsin Wong (Monash Univ
 ersity, Chinese University of Hong Kong)\n\nThis paper introduces a novel 
 approach to synthesize texture to dress up a given 3D object, given a text
  prompt. \nBased on the pretrained text-to-image (T2I) diffusion model, ex
 isting methods usually employ a project-and-inpaint approach, in which a v
 iew of the given object is first generated and warped to another view for 
 inpainting. But it tends to generate inconsistent texture due to the async
 hronous diffusion of multiple views. We believe such asynchronous diffusio
 n and insufficient information sharing among views are the root causes of 
 the inconsistent artifact. \nIn this paper, we propose a synchronized mult
 i-view diffusion approach that allows the diffusion processes from differe
 nt views to reach a consensus on the generated content early in the proces
 s, and hence ensures the texture consistency. To synchronize the diffusion
 , we share the denoised content among different views in each denoising st
 ep, specifically blending the latent content in the texture domain from vi
 ews with overlap. Our method demonstrates superior performance in generati
 ng consistent, seamless, highly detailed textures, comparing to state-of-t
 he-art methods. Code of the paper will be release upon acceptance of the p
 aper.\n\nRegistration Category: Full Access, Full Access Supporter\n\nLang
 uage Format: English Language\n\nSession Chair: Minhyuk Sung (Korea Advanc
 ed Institute of Science and Technology (KAIST))\n\n
URL:https://asia.siggraph.org/2024/program/?id=papers_535&sess=sess122
END:VEVENT
END:VCALENDAR