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:20260817T171535Z
LOCATION:Hall B5 (1)\, B Block\, Level 5
DTSTART;TZID=Asia/Tokyo:20241205T135800
DTEND;TZID=Asia/Tokyo:20241205T140900
UID:siggraphasia_SIGGRAPH Asia 2024_sess130_papers_199@linklings.com
SUMMARY:SRIF: Semantic Shape Registration Empowered by Diffusion-based Ima
 ge Morphing and Flow Estimation
DESCRIPTION:Mingze Sun (Tsinghua shenzhen international graduate school); 
 Chen Guo and Puhua Jiang (Tsinghua shenzhen international graduate school,
  Pengcheng Lab); and Shiwei Mao, Yurun Chen, and Ruqi Huang (Tsinghua shen
 zhen international graduate school)\n\nIn this paper, we propose \textbf{S
 RIF}, a novel \textbf{S}emantic shape \textbf{R}egistration framework base
 d on diffusion-based \textbf{I}mage morphing and \textbf{F}low Estimation.
  \nMore concretely, given a pair of extrinsically aligned shapes, we first
  render them from multi-views, and then we utilize an image interpolation 
 framework tailored for diffusion models to generate sequences of intermedi
 ate images between them. The images are later fed into a dynamic 3D Gaussi
 an splatting framework, with which we reconstruct and post-process for int
 ermediate \emph{point clouds} respecting the image morphing processing. In
  the end, tailored for the above, we propose a novel registration module t
 o estimate continuous normalizing flow, which deforms source shape consist
 ently towards the target, with intermediate point clouds as weak guidance.
  Our key insight is to leverage LVMs to \emph{associate} shapes and theref
 ore obtain much richer semantic information on the relationship between sh
 apes than the ad-hoc independent semantic information extraction. As conse
 quence, \textbf{SRIF} achieves high-quality dense correspondences on chall
 enging shape pairs, but also delivers smooth, semantically\nmeaningful int
 erpolation in between. Empirical evidence justifies the effectiveness and 
 superiority of our method as well as specific design choices. The code wil
 l be made public upon acceptance.\n\nRegistration Category: Full Access, F
 ull Access Supporter\n\nLanguage Format: English Language\n\nSession Chair
 : Noam Aigerman (University of Montreal, Mila)\n\n
URL:https://asia.siggraph.org/2024/program/?id=papers_199&sess=sess130
END:VEVENT
END:VCALENDAR