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:20260720T003345Z
LOCATION:Room 408 B
DTSTART;TZID=America/Los_Angeles:20260720T104500
DTEND;TZID=America/Los_Angeles:20260720T123500
UID:siggraph_SIGGRAPH 2026_sess151@linklings.com
SUMMARY:Generative 3D & Editing
DESCRIPTION:VecSet-Edit: Unleashing Pre-trained LRM for Mesh Editing from 
 Single Image\n\nWe propose VecSet-Edit, a novel pipeline leveraging a pre-
 trained Large Reconstruction Model for high-fidelity 3D mesh editing. Usin
 g only 2D image conditions and strategic token manipulation based on our e
 xploration of VecSet features, it enables precise localized edits while pr
 eserving original de...\n\n\nTeng-Fang Hsiao, Bo-Kai Ruan, Yu-Lun Liu, and
  Hong-Han Shuai (National Yang Ming Chiao Tung University)\n--------------
 -------\nCubePart: An Open-Vocabulary Part-Controllable 3D Generator\n\nCu
 bePart is an open-vocabulary 3D generative framework offering explicit sem
 antic part-based control over mesh structure. Guided by a text prompt and 
 custom part schema, it synthesizes coherent multi-part meshes, which are r
 eady for direct integration into game engines and animation pipelines with
 ou...\n\n\nYiheng Zhu, Kangle Deng, Jean-Philippe Fauconnier, Inaki Navarr
 o Oiza, Daiqing Li, Ava Pun, Yinan Zhang, Peiye Zhuang, and Xiaoxia Sun (R
 oblox); Maneesh Agrawala (Roblox, Stanford University); and Kiran Bhat and
  Tinghui Zhou (Roblox)\n---------------------\nIris3D: 3D Generation via S
 ynchronized Diffusion Distillation\n\nIris3D is a novel 3D generation syst
 em that combines feedforward reconstruction models with a new Synchronized
  Diffusion Distillation (SDD) method to produce detailed geometry and vivi
 d textures. By addressing view inconsistency and over-smoothing issues in 
 traditional Score Distillation Sampling (S...\n\n\nYixun Liang, Weiyu Li, 
 and Rui Chen (The Hong Kong University of Science and Technology); Fei-Pen
 g Tian (Light Illusions); Jiarui Liu, Ying-Cong Chen, and Ping Tan (The Ho
 ng Kong University of Science and Technology); Xiao-Xiao Long (The Hong Ko
 ng University of Science and Technology, Nanjing University); and Yixun Li
 ang\n---------------------\nInvSculpt: Inverse Sculpting Modeling via Cont
 rolled 3D Generation and a Vector Displacement Field\n\nWe propose InvScul
 pt, which decomposes a mesh into a base shape and reusable details represe
 nted by a vector displacement field. It uses text-guided 2D priors and a 3
 D flow model for mask-free detail removal while preserving identity. The m
 ethod enables accurate decomposition and high-fidelity detai...\n\n\nHengy
 u Meng and Lanjiong Li (The Hong Kong University of Science and Technology
  (Guangzhou), Tencent LightSpeed Studio); Zhijing Shao (The Hong Kong Univ
 ersity of Science and Technology (Guangzhou)); Yingda Yin, Lingting Zhu, Z
 eyu Hu, and Xin Wang (Tencent LightSpeed Studio); Ligang Liu (Laoshan Labo
 ratory); and Zeyu Wang (The Hong Kong University of Science and Technology
  (Guangzhou), The Hong Kong University of Science and Technology)\n-------
 --------------\nPixal3D: Pixel-Aligned 3D Generation from Images\n\nPixal3
 D is a pixel-aligned paradigm that solves the fidelity bottleneck in image
 -to-3D synthesis. By utilizing a back-projection conditioning scheme to ex
 plicitly lift 2D features into view-aligned 3D volumes, it eliminates ambi
 guity, enabling scalable generative reconstruction of highly accurate 3D..
 .\n\n\nDong-Yang Li (Tsinghua University); Wang Zhao, Yuxin Chen, and Wenb
 o Hu (Tencent); Meng-Hao Guo (Tsinghua University); Fang-Lue Zhang (Victor
 ia University of Wellington); Ying Shan (Tencent); and Shi-Min Hu (Tsinghu
 a University)\n---------------------\nGenerative 3D & Editing - Interactiv
 e Discussion\n\nAfter the summary presentations, attendees will participat
 e in an interactive discussion. Outside the room will be a series of poste
 r boards for authors to gather around with the audience. Authors are invit
 ed to bring any material related to their paper that could instigate furth
 er conversation such...\n\n---------------------\nProx-E: Fine-Grained 3D 
 Shape Editing via Primitive-Based Abstractions\n\nProx-E is a training-fre
 e framework for fine-grained 3D shape editing that uses a primitive-based 
 geometric proxy to explicitly control structural changes. By editing this 
 proxy with a VLM and guiding a 3D generative model accordingly, it enables
  localized, identity-preserving edits that outperform ...\n\n\nEtai Sella 
 (Tel Aviv University, Snap); Hao Phung (Cornell Tech); Nitay Amiel (Techni
 on - Israel Institute of Technology); Or Litany (Technion - Israel Institu
 te of Technology, NVIDIA); Or Patashnik (Tel Aviv University, Snap); and H
 adar Averbuch-Elor (Cornell Tech)\n---------------------\nShapeUP: Scalabl
 e Image-Conditioned 3D Editing\n\nShapeUp introduces a scalable framework 
 for high-fidelity 3D asset editing. By casting editing as an image-conditi
 oned, latent-to-latent translation, it enables precise local and global tr
 ansformations while maintaining native 3D consistency. Our approach ensure
 s seamless identity preservation witho...\n\n\nInbar Gat (Aigency.ai, Tel 
 Aviv University); Dana Cohen Bar and Guy Levy (Tel Aviv University); Elad 
 Richardson (Runway); and Daniel Cohen-Or (Tel Aviv University)\n\nInterest
  Area: Research & Education\n\nKeyword: Artificial Intelligence/Machine Le
 arning, Geometry, Modeling\n\nRegistration Category: Full Conference Suppo
 rter, Full Conference\n\nSession Chair: Sunghyun Cho (POSTECH)
END:VEVENT
END:VCALENDAR