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 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:Handi Yin and Bonan Liu (Hong Kong University of Science and T
 echnology, Guangzhou); Manuel Kaufmann (ETH Zürich); Jinhao He (Hong Kong 
 University of Science and Technology, Guangzhou); Sammy Christen (ETH Züri
 ch); and Jie Song and Pan Hui (Hong Kong University of Science and Technol
 ogy, Guangzhou; Hong Kong University of Science and Technology)\n\nWe pres
 ent EgoHDM, an online egocentric-inertial human motion capture (mocap), lo
 calization, and dense mapping system. Our system uses 6 inertial measureme
 nt units (IMUs) and a commodity head-mounted RGB camera. EgoHDM is the fir
 st human mocap system that offers dense scene mapping in near real-time. F
 urther, it is fast and robust to initialize and fully closes the loop betw
 een physically plausible map-aware global human motion estimation and moca
 p-aware 3D scene reconstruction. Our key idea is integrating camera locali
 zation and mapping information with inertial human motion capture bidirect
 ionally in our system. To achieve this, we design a tightly coupled mocap-
 aware dense bundle adjustment and physics-based body pose correction modul
 e leveraging a local body-centric elevation map. The latter introduces a n
 ovel terrain-aware contact PD controller, which enables characters to phys
 ically contact the given local elevation map thereby reducing human floati
 ng or penetration. We demonstrate the performance of our system on establi
 shed 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 the art. Our qualitati
 ve evaluations on newly captured data further demonstrate that EgoHDM can 
 cover challenging scenarios in non-flat terrain including stepping over st
 airs and outdoor scenes in the wild.\n\nRegistration Category: Full Access
 , Full Access Supporter\n\nLanguage Format: English Language\n\nSession Ch
 air: Yuting Ye (Reality Labs Research, Meta; Meta)\n\n
URL:https://asia.siggraph.org/2024/program/?id=papers_363&sess=sess129
END:VEVENT
END:VCALENDAR