BEGIN:VCALENDAR
VERSION:2.0
PRODID:Linklings LLC
BEGIN:VTIMEZONE
TZID:Asia/Seoul
X-LIC-LOCATION:Asia/Seoul
BEGIN:STANDARD
TZOFFSETFROM:+0900
TZOFFSETTO:+0900
TZNAME:KST
DTSTART:18871231T000000
DTSTART:19881009T020000
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
DTSTAMP:20230103T035307Z
LOCATION:Room 324\, Level 3\, West Wing
DTSTART;TZID=Asia/Seoul:20221206T140000
DTEND;TZID=Asia/Seoul:20221206T153000
UID:siggraphasia_SIGGRAPH Asia 2022_sess154_papers_431@linklings.com
SUMMARY:QuestSim: Human Motion Tracking from Sparse Sensors with Simulated
  Avatars
DESCRIPTION:Technical Communications, Technical Papers\n\nQuestSim: Human 
 Motion Tracking from Sparse Sensors with Simulated Avatars\n\nWinkler, Won
 , Ye\n\nReal-time tracking of human body motion is crucial for interactive
  and\nimmersive experiences in AR/VR. However, very limited sensor data ab
 out\nthe body is available from standalone wearable devices such as HMDs (
 Head\nMounted Devices) or AR glasses. In this work, we present a reinforce
 ment\nlearning framework that takes in sparse signals from an HMD and two\
 ncontrollers, and simulates plausible and physically valid full body motio
 ns.\nUsing high quality full body motion as dense supervision during train
 ing,\na simple policy network can learn to output appropriate torques for 
 the\ncharacter to balance, walk, and jog, while closely following the inpu
 t signals.\nOur results demonstrate surprisingly similar leg motions to gr
 ound truth\nwithout any observations of the lower body, even when the inpu
 t is only\nthe 6D transformations of the HMD. We also show that a single p
 olicy\ncan be robust to diverse locomotion styles, different body sizes, a
 nd novel\nenvironments.\n\nRegistration Category: FULL ACCESS, ON-DEMAND A
 CCESS\n\nLanguage: ENGLISH\n\nFormat: IN-PERSON, ON-DEMAND
URL:https://sa2022.siggraph.org/en/full-program/?id=papers_431&sess=sess15
 4
END:VEVENT
END:VCALENDAR