Presentation

<
BlockFusion: Expandable 3D Scene Generation Using Latent Tri-plane Extrapolation
SessionGenerative 3D Geometry
DescriptionBlockFusion is a 3D diffusion-based model that generates and expands 3D scenes using unit blocks. The denoising diffusion process allows high-quality and diverse scene generation and auto-regressive scene expansion. BlockFusion generates large, high-quality 3D scenes for indoor and outdoor scenarios.
Authors
Zhennan Wu
University of Tokyo
Yang Li
Tencent
Han Yan
Shanghai Jiao Tong University
Taizhang Shang
Tencent
Weixuan Sun
Tencent
Senbo Wang
Tencent
Ruikai Cui
Australian National University
Weizhe Liu
Tencent
Hiroyuki Sato
National Institute of Informatics, Japan
University of Tokyo
Hongdong Li
Tencent
Pan Ji
Tencent
Event Type
Technical Paper
TimeMonday, 29 July 202410:55am - 11:05am MDT
LocationMile High 2A
ACM Digital Library Journal Papers' PDFs
Conference Papers' PDFs
Session Time & Location
Sunday, 28 July 20246:00pm - 8:45pm MDTBluebird Ballroom
Monday, 29 July 202410:45am - 12:15pm MDTMile High 2A
Interest Areas
Research & Education
Recordings
Livestreamed
Recorded
Keywords
AI
Geometry
Modeling
Registration Categories
Full Conference
Full Conference Supporter
Virtual Access
Exhibitor Full Conference
Monday
Next PresentationNext PresentationSketchDream: Sketch-based Text-to-3D Generation and Editing