BEGIN:VCALENDAR
VERSION:2.0
PRODID:Linklings LLC
BEGIN:VTIMEZONE
TZID:America/Los_Angeles
X-LIC-LOCATION:America/Los_Angeles
BEGIN:DAYLIGHT
TZOFFSETFROM:-0800
TZOFFSETTO:-0700
TZNAME:PDT
DTSTART:19700308T020000
RRULE:FREQ=YEARLY;BYMONTH=3;BYDAY=2SU
END:DAYLIGHT
BEGIN:STANDARD
TZOFFSETFROM:-0700
TZOFFSETTO:-0800
TZNAME:PST
DTSTART:19701101T020000
RRULE:FREQ=YEARLY;BYMONTH=11;BYDAY=1SU
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
DTSTAMP:20260417T190058Z
LOCATION:West Building\, Rooms 211-214
DTSTART;TZID=America/Los_Angeles:20250811T140000
DTEND;TZID=America/Los_Angeles:20250811T153000
UID:siggraph_SIGGRAPH 2025_sess126@linklings.com
SUMMARY:Learning & Shapes
DESCRIPTION:NAM: Neural Adjoint Maps for refinement of shape correspondenc
 es\n\nWe introduce Neural Adjoint Maps, a novel representation for corresp
 ondences between 3D shapes. Built on and extending the functional map fram
 ework, our approach enables accurate, non-linear refinement of shape match
 ing across meshes and point clouds, setting a new standard in diverse scen
 arios and ...\n\n\nGiulio Viganò (Università di Milano Bicocca), Maks Ovsj
 anikov (Centre National de la Recherche Scientifique - Laboratoire d'infor
 matique de l'École Polytechnique (LIX)), and Simone Melzi (Università di M
 ilano Bicocca)\n---------------------\nBANG: Dividing 3D Assets via Genera
 tive Exploded Dynamics\n\nBANG introduces Generative Exploded Dynamics, a 
 novel method that dynamically decomposes 3D objects into meaningful, volum
 etric parts through smooth, controllable exploded views. Bridging intuitiv
 e human understanding and generative AI, it enables precise part-level man
 ipulation, semantic comprehens...\n\n\nLongwen Zhang, Qixuan Zhang, and Ha
 oran Jiang (ShanghaiTech University, Deemos Technology); Yinuo Bai (Shangh
 aiTech University); Wei Yang (Huazhong University of Science and Technolog
 y); and Lan Xu and Jingyi Yu (ShanghaiTech University)\n------------------
 ---\nOctGPT: Octree-based Multiscale Autoregressive Models for 3D Shape Ge
 neration\n\nOctGPT is a novel multiscale autoregressive model for 3D shape
  generation. It introduces hierarchical serialized octree representation, 
 octree-based transformer with 3D RoPE and token-parallel generation scheme
 s. OctGPT significantly accelerates convergence, achieves performance riva
 ling or surpassi...\n\n\nSi-Tong Wei, Rui-Huan Wang, Chuan-Zhi Zhou, Baoqu
 an Chen, and Peng-Shuai Wang (Peking University)\n---------------------\nL
 earning & Shapes - Interactive Discussion\n\nAfter the summary presentatio
 ns, attendees will participate in an interactive discussion. Outside the r
 oom will be a series of poster boards for authors to gather around with th
 e audience. Authors are invited to bring any material related to their pap
 er that could instigate further conversation such...\n\n------------------
 ---\nGenAnalysis: Joint Shape Analysis by Learning Man-Made Shape Generato
 rs with Deformation Regularizations\n\nWe present GenAnalysis, an implicit
  shape generation framework enabling joint shape matching and consistent s
 egmentation by enforcing as-affine-as-possible (AAAP) deformations via reg
 ularization loss in latent space. It enables shape analysis via extracting
  and analysing shape variations in the tang...\n\n\nYuezhi Yang and Haitao
  Yang (University of Texas at Austin), George Kiyohiro Nakayama (Stanford 
 University), Xiangru Huang (Westlake University), Leonidas Guibas (Stanfor
 d University), and Qixing Huang (University of Texas at Austin)\n---------
 ------------\nUnsupervised Decomposition of 3D Shapes into Expressive and 
 Editable Extruded Profile Primitives\n\n3D2EP transforms 3D shapes into ex
 pressive, editable primitives by extruding 2D profiles along 3D curves. Th
 is approach creates compact, interpretable representations that support in
 tuitive editing and flexible redesign. It delivers high fidelity and effic
 iency, outperforming existing methods across...\n\n\nChunyi Sun (Australia
 n National University); Junlin Han and Runjia Li (University of Oxford); a
 nd Weijian Deng, Dylan Campbell, and Stephen Gould (Australian National Un
 iversity)\n---------------------\nMASH: Masked Anchored SpHerical Distance
 s for 3D Shape Representation and Generation\n\nWe introduce Masked Anchor
 ed SpHerical Distances (MASH), a novel multi-view and parametrized represe
 ntation of 3D shapes. MASH is versatilefor multiple applications including
  surface reconstruction, shape generation, completion, and blending, achie
 ving superior performance thanks to its unique repre...\n\n\nChanghao Li a
 nd Yu Xin (University of Science and Technology of China); Xiaowei Zhou (S
 tate Key Laboratory of CAD & CG, Zhejiang University); Ariel Shamir (Reich
 man University); Hao Zhang (Simon Fraser University); Ligang Liu (Universi
 ty of Science and Technology of China); and Ruizhen Hu (Shenzhen Universit
 y)\n\nInterest Area: Research & Education\n\nRecording: Livestreamed, Not 
 Livestreamed, Recorded, Not Recorded\n\nRegistration Category: Full Confer
 ence, Virtual Access, Monday\n\nSession Chair: Nicholas Sharp (NVIDIA)
END:VEVENT
END:VCALENDAR