The effect of depth data and upper limb impairment on lightweight monocular RGB human pose estimation models
Including depth data using an early fusion architecture improves the accuracy-efficiency trade-off of the HPE model. HPE accuracy is not affected by the presence of physical impairments. These results suggest that using depth data with RGB data is beneficial to HPE, and that models trained with data collected from uninjured individuals can generalize to persons with physical impairments.
Authors
- Xilin Liu (ORCID: https://orcid.org/0000-0002-1136-6783)
- José Zariffa (ORCID: https://orcid.org/0000-0002-8842-745X)
- César Márquez-Chin (ORCID: https://orcid.org/0000-0002-9916-4122)
- Gloria-Edith Boudreault-Morales
Institutions
- University Health Network (CA)
- University of Toronto (CA)
- Toronto Rehabilitation Institute (CA)
Publication Details
- Journal
- BioMedical Engineering OnLine
- Published
- 2025-02-07
- DOI
- https://doi.org/10.1186/s12938-025-01347-y
- Citations
- 8
- Primary Topic
- Human Pose and Action Recognition
- Type
- article
- Field-Weighted Citation Impact
- 10.21
Funders
- University of Toronto