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:20241205T105600
DTEND;TZID=Asia/Tokyo:20241205T110800
UID:siggraphasia_SIGGRAPH Asia 2024_sess129_papers_497@linklings.com
SUMMARY:RoMo: A Robust Solver for Full-body Unlabeled Optical Motion Captu
 re
DESCRIPTION:Technical Papers\n\nXiaoyu Pan and Bowen Zheng (State Key Labo
 ratory of CAD&CG, Zhejiang University); Xinwei Jiang, Zijiao Zeng, and Qil
 ong Kou (Tencent Games Digital Content Technology Center); He Wang (Depart
 ment of Computer Science and UCL Centre for Artificial Intelligence, Unive
 rsity College London); and Xiaogang Jin (State Key Laboratory of CAD&CG, Z
 hejiang University)\n\nOptical motion capture (MoCap) is the "gold standar
 d" for accurately capturing full-body motions. To make use of raw MoCap po
 int data, the system labels the points with corresponding body part locati
 ons and solves the full-body motions. However, MoCap data often contains m
 islabeling, occlusion and positional errors, requiring extensive manual co
 rrection. To alleviate this burden, we introduce RoMo, an automatic learni
 ng-based framework for robustly labeling and solving raw optical motion ca
 pture data. In the labeling stage, RoMo employs a divide-and-conquer strat
 egy to break down the complex full-body labeling challenge into manageable
  subtasks: full-body segmentation and part-specific labeling. To utilize t
 he temporal continuity of markers, RoMo generates marker tracklets using a
  K-partite graph-based clustering algorithm, where markers serve as nodes 
 and edges are formed based on positional and feature similarities. For mot
 ion solving, to prevent error accumulation along the kinematic chain, we i
 ntroduce a hybrid inverse kinematic solver that utilizes joint positions a
 s intermediate representations and adjusts the template skeleton to match 
 estimated joint rotations. We demonstrate that RoMo achieves high labeling
  and solving accuracy across multiple metrics and various datasets. Extens
 ive comparisons show that our method outperforms state-of-the-art research
  methods. On a real dataset, RoMo improves the F1 score of hand labeling f
 rom 0.94 to 0.98, and reduces the position error of body motion solving by
  25%. Furthermore, RoMo can be applied in scenarios where commercial syste
 ms are inadequate.\n\nRegistration Category: Full Access, Full Access Supp
 orter\n\nLanguage Format: English Language\n\nSession Chair: Yuting Ye (Re
 ality Labs Research, Meta; Meta)
URL:https://asia.siggraph.org/2024/program/?id=papers_497&sess=sess129
END:VEVENT
END:VCALENDAR