A Dual-Masked Auto-Encoder for Robust Motion Capture with Spatial-Temporal Skeletal Token Completion

Junkun Jiang, Jie Chen*, Yi-Ke Guo

*Corresponding author for this work

Research output: Chapter in book/report/conference proceedingConference proceedingpeer-review

1 Citation (Scopus)

Abstract

Multi-person motion capture can be challenging due to ambiguities caused by severe occlusion, fast body movement, and complex interactions. Existing frameworks build on 2D pose estimations and triangulate to 3D coordinates via reasoning the appearance, trajectory, and geometric consistencies among multi-camera observations. However, 2D joint detection is usually incomplete and with wrong identity assignments due to limited observation angle, which leads to noisy 3D triangulation results. To overcome this issue, we propose to explore the short-range autoregressive characteristics of skeletal motion using transformer. First, we propose an adaptive, identity-aware triangulation module to reconstruct 3D joints and identify the missing joints for each identity. To generate complete 3D skeletal motion, we then propose a Dual-Masked Auto-Encoder (D-MAE) which encodes the joint status with both skeletal-structural and temporal position encoding for trajectory completion. D-MAE's flexible masking and encoding mechanism enable arbitrary skeleton definitions to be conveniently deployed under the same framework. In order to demonstrate the proposed model's capability in dealing with severe data loss scenarios, we contribute a high-accuracy and challenging motion capture dataset of multi-person interactions with severe occlusion. Evaluations on both benchmark and our new dataset demonstrate the efficiency of our proposed model, as well as its advantage against the other state-of-the-art methods.
Original languageEnglish
Title of host publicationMM '22: Proceedings of the 30th ACM International Conference on Multimedia
PublisherAssociation for Computing Machinery (ACM)
Pages5123–5131
Number of pages9
ISBN (Print)9781450392037
DOIs
Publication statusPublished - 10 Oct 2022
Event30th ACM International Conference on Multimedia, MM 2022 - Lisboa, Portugal
Duration: 10 Oct 202214 Oct 2022
https://dl.acm.org/doi/proceedings/10.1145/3503161

Conference

Conference30th ACM International Conference on Multimedia, MM 2022
Country/TerritoryPortugal
CityLisboa
Period10/10/2214/10/22
Internet address

User-Defined Keywords

  • 3D human pose estimation
  • motion capture
  • masked auto-encoder
  • transformer
  • spatial-temporal encoding

Fingerprint

Dive into the research topics of 'A Dual-Masked Auto-Encoder for Robust Motion Capture with Spatial-Temporal Skeletal Token Completion'. Together they form a unique fingerprint.

Cite this