BEGIN:VCALENDAR
VERSION:2.0
PRODID:Linklings LLC
BEGIN:VTIMEZONE
TZID:Asia/Tokyo
X-LIC-LOCATION:Asia/Tokyo
BEGIN:STANDARD
TZOFFSETFROM:+0900
TZOFFSETTO:+0900
TZNAME:JST
DTSTART:18871231T000000
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
DTSTAMP:20260817T171533Z
LOCATION:Hall B5 (1)\, B Block\, Level 5
DTSTART;TZID=Asia/Tokyo:20241205T163000
DTEND;TZID=Asia/Tokyo:20241205T164100
UID:siggraphasia_SIGGRAPH Asia 2024_sess136_papers_924@linklings.com
SUMMARY:Frankenstein: Generating Semantic-Compositional 3D Scenes in One T
 ri-Plane
DESCRIPTION:Han Yan (Shanghai Jiao Tong University); Yang Li (Tencent); Zh
 ennan Wu (University of Tokyo); Shenzhou Chen, Weixuan Sun, Taizhang Shang
 , Weizhe Liu, Tian Chen, and Xiaqiang Dai (Tencent); Chao Ma (Shanghai Jia
 o Tong University); Hongdong Li (Australian National University); and Pan 
 Ji (Tencent)\n\nWe present Frankenstein, a diffusion-based framework that 
 can generate semantic-compositional 3D scenes in a single pass. Unlike exi
 sting methods that output a single, unified 3D shape, Frankenstein simulta
 neously generates multiple separated shapes, each corresponding to a seman
 tically meaningful part. The 3D scene information is encoded in one single
  tri-plane tensor, from which multiple Signed Distance Function (SDF) fiel
 ds can be decoded to represent the compositional shapes. During training, 
 an auto-encoder compresses tri-planes into a latent space, and then the de
 noising diffusion process is employed to approximate the distribution of t
 he compositional scenes. Frankenstein demonstrates promising results in ge
 nerating room interiors as well as human avatars with automatically separa
 ted parts. The generated scenes facilitate many downstream applications, s
 uch as part-wise re-texturing, object rearrangement in the room or avatar 
 cloth re-targeting.\n\nRegistration Category: Full Access, Full Access Sup
 porter\n\nLanguage Format: English Language\n\nSession Chair: Manolis Savv
 a (Simon Fraser University)\n\n
URL:https://asia.siggraph.org/2024/program/?id=papers_924&sess=sess136
END:VEVENT
END:VCALENDAR