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:20251218T030653Z
LOCATION:Meeting Room S421\, Level 4
DTSTART;TZID=Asia/Hong_Kong:20251218T093200
DTEND;TZID=Asia/Hong_Kong:20251218T094300
UID:siggraphasia_SIGGRAPH Asia 2025_sess146_papers_1790@linklings.com
SUMMARY:ArtiLatent: Realistic Articulated 3D Object Generation via Structu
 red Latents
DESCRIPTION:Honghua Chen, Yushi Lan, Yongwei Chen, and Xingang Pan (Nanyan
 g Technological University (NTU))\n\nWe propose ArtiLatent, a generative f
 ramework that synthesizes human-made 3D objects with fine-grained geometry
 , accurate articulation, and realistic appearance. Our approach jointly mo
 dels part geometry and articulation dynamics by embedding sparse voxel rep
 resentations and associated articulation properties—including joint type, 
 axis, origin, range, and part category—into a unified latent space via a v
 ariational autoencoder. A latent diffusion model is then trained over this
  space to enable diverse yet physically plausible sampling.\nTo reconstruc
 t photorealistic 3D shapes, we introduce an articulation-aware Gaussian de
 coder that accounts for articulation-dependent visibility changes (e.g., r
 evealing the interior of a drawer when opened). By conditioning appearance
  decoding on articulation state, our method assigns plausible texture feat
 ures to regions that are typically occluded in static poses, significantly
  improving visual realism across articulation configurations.\nExtensive e
 xperiments on furniture-like objects from PartNet-Mobility and ACD dataset
 s demonstrate that ArtiLatent outperforms existing approaches in geometric
  consistency and appearance fidelity. Our framework provides a scalable so
 lution for articulated 3D object synthesis and manipulation.\n\nRegistrati
 on Category: Full Access, Full Access Supporter\n\nSession Chair: Ziqi Wan
 g (HKUST, EPFL)\n\n
END:VEVENT
END:VCALENDAR