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:20250110T023312Z
LOCATION:Hall B7 (1)\, B Block\, Level 7
DTSTART;TZID=Asia/Tokyo:20241205T113100
DTEND;TZID=Asia/Tokyo:20241205T114300
UID:siggraphasia_SIGGRAPH Asia 2024_sess129_papers_363@linklings.com
SUMMARY:EgoHDM: An Online Egocentric-Inertial Human Motion Capture, Locali
 zation, and Dense Mapping System
DESCRIPTION:Technical Papers\n\nHandi Yin and Bonan Liu (Hong Kong Univers
 ity of Science and Technology, Guangzhou); Manuel Kaufmann (ETH Zürich); J
 inhao He (Hong Kong University of Science and Technology, Guangzhou); Samm
 y Christen (ETH Zürich); and Jie Song and Pan Hui (Hong Kong University of
  Science and Technology, Guangzhou; Hong Kong University of Science and Te
 chnology)\n\nWe present EgoHDM, an online egocentric-inertial human motion
  capture (mocap), localization, and dense mapping system. Our system uses 
 6 inertial measurement units (IMUs) and a commodity head-mounted RGB camer
 a. EgoHDM is the first human mocap system that offers dense scene mapping 
 in near real-time. Further, it is fast and robust to initialize and fully 
 closes the loop between physically plausible map-aware global human motion
  estimation and mocap-aware 3D scene reconstruction. Our key idea is integ
 rating camera localization and mapping information with inertial human mot
 ion capture bidirectionally in our system. To achieve this, we design a ti
 ghtly coupled mocap-aware dense bundle adjustment and physics-based body p
 ose correction module leveraging a local body-centric elevation map. The l
 atter introduces a novel terrain-aware contact PD controller, which enable
 s characters to physically contact the given local elevation map thereby r
 educing human floating or penetration. We demonstrate the performance of o
 ur system on established synthetic and real-world benchmarks. The results 
 show that our method reduces human localization, camera pose, and mapping 
 accuracy error by 41%, 71%, 46%, respectively, compared to the state of th
 e art. Our qualitative evaluations on newly captured data further demonstr
 ate that EgoHDM can cover challenging scenarios in non-flat terrain includ
 ing stepping over stairs and outdoor scenes in the wild.\n\nRegistration C
 ategory: Full Access, Full Access Supporter\n\nLanguage Format: English La
 nguage\n\nSession Chair: Yuting Ye (Reality Labs Research, Meta; Meta)
URL:https://asia.siggraph.org/2024/program/?id=papers_363&sess=sess129
END:VEVENT
END:VCALENDAR