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:20260817T171530Z
LOCATION:Hall B7 (1)\, B Block\, Level 7
DTSTART;TZID=Asia/Tokyo:20241204T154300
DTEND;TZID=Asia/Tokyo:20241204T155400
UID:siggraphasia_SIGGRAPH Asia 2024_sess120_papers_664@linklings.com
SUMMARY:MaskedMimic: Unified Physics-Based Character Control Through Maske
 d Motion Inpainting
DESCRIPTION:Chen Tessler (NVIDIA Research), Yunrong Guo (NVIDIA), Ofir Nab
 ati and Gal Chechik (NVIDIA Research), and Xue Bin Peng (NVIDIA)\n\nCrafti
 ng a single, versatile physics-based controller that can breathe life into
  interactive characters across a wide spectrum of scenarios represents an 
 exciting frontier in character animation. An ideal controller should suppo
 rt diverse control modalities, such as sparse target keyframes, text instr
 uctions, and scene information. While previous works have proposed physica
 lly simulated, scene-aware control models, these systems have predominantl
 y focused on developing controllers that each specializes in a narrow set 
 of tasks and control modalities. This work presents MaskedMimic, a novel a
 pproach that formulates physics-based character control as a general motio
 n inpainting problem. Our key insight is to train a single unified model t
 o synthesize motions from partial (masked) motion descriptions, such as ma
 sked keyframes, objects, text descriptions, or any combination thereof. Th
 is is achieved by leveraging motion tracking data and designing a scalable
  training method that can effectively utilize diverse motion descriptions 
 to produce coherent animations. Through this process, our approach learns 
 a physics-based controller that provides an intuitive control interface wi
 thout requiring tedious reward engineering for all behaviors of interest. 
 The resulting controller supports a wide range of control modalities and e
 nables seamless transitions between disparate tasks. By unifying character
  control through motion inpainting, MaskedMimic creates versatile virtual 
 characters. These characters can dynamically adapt to complex scenes and c
 ompose diverse motions on demand, enabling more interactive and immersive 
 experiences.\n\nRegistration Category: Full Access, Full Access Supporter\
 n\nLanguage Format: English Language\n\nSession Chair: Hao (Richard) Zhang
  (Simon Fraser University, Augmenta)\n\n
URL:https://asia.siggraph.org/2024/program/?id=papers_664&sess=sess120
END:VEVENT
END:VCALENDAR