BEGIN:VCALENDAR
VERSION:2.0
PRODID:Linklings LLC
BEGIN:VTIMEZONE
TZID:Asia/Hong_Kong
X-LIC-LOCATION:Asia/Hong_Kong
BEGIN:STANDARD
TZOFFSETFROM:+0800
TZOFFSETTO:+0800
TZNAME:HKT
DTSTART:19911015T033000
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
DTSTAMP:20251218T030656Z
LOCATION:Meeting Room S426+S427\, Level 4
DTSTART;TZID=Asia/Hong_Kong:20251218T105000
DTEND;TZID=Asia/Hong_Kong:20251218T110100
UID:siggraphasia_SIGGRAPH Asia 2025_sess152_papers_1846@linklings.com
SUMMARY:Rigidity-Aware 3D Gaussian Deformation from a Single Image
DESCRIPTION:Jinhyeok Kim, Jaehun Bang, Seunghyun Seo, and Kyungdon Joo (Ul
 san National Institute of Science and Technology)\n\nReconstructing object
  deformation from a single image remains a significant challenge in comput
 er vision and graphics. Existing methods typically rely on multi-view vide
 o to recover deformation, limiting their applicability under constrained s
 cenarios. To address this, we propose DeformSplat, a novel framework that 
 effectively guides 3D Gaussian deformation from only a single image. Our m
 ethod introduces two main technical contributions. First, we present a Gau
 ssian-to-Pixel Matching that bridges the domain gap between 3D Gaussian re
 presentations and 2D pixel observations. This enables robust deformation g
 uidance from sparse visual cues. Second, we propose novel Rigid Part Segme
 ntation consisting of initialization and refinement steps. This segmentati
 on explicitly identifies rigid regions, crucial for maintaining geometric 
 coherence during deformation. By combining these two techniques, our appro
 ach can reconstruct consistent deformations from limited input. Extensive 
 experiments demonstrate that our approach significantly outperforms existi
 ng methods, particularly in deformation accuracy and geometric preservatio
 n. Furthermore, our framework naturally extends to various applications, s
 uch as frame interpolation and interactive object manipulation.\n\nRegistr
 ation Category: Full Access, Full Access Supporter\n\nSession Chair: Lin G
 ao (University of Chinese Academy of Sciences)\n\n
END:VEVENT
END:VCALENDAR