Learning action recognition model from depth and skeleton videos
- Submitting institution
-
The University of Lancaster
- Unit of assessment
- 11 - Computer Science and Informatics
- Output identifier
- 282587912
- Type
- E - Conference contribution
- DOI
-
10.1109/ICCV.2017.621
- Title of conference / published proceedings
- Proceedings of the IEEE International Conference on Computer Vision
- First page
- 5833
- Volume
- 2017-October
- Issue
- -
- ISSN
- -
- Open access status
- -
- Month of publication
- December
- Year of publication
- 2017
- URL
-
-
- Supplementary information
-
-
- Request cross-referral to
- -
- Output has been delayed by COVID-19
- No
- COVID-19 affected output statement
- -
- Forensic science
- No
- Criminology
- No
- Interdisciplinary
- No
- Number of additional authors
-
1
- Research group(s)
-
B - Data Science
- Citation count
- 41
- Proposed double-weighted
- No
- Reserve for an output with double weighting
- No
- Additional information
- This paper presents the first deep multi-modal fusion model for recognizing human activities from depth and skeleton videos. This paper is significant because it takes advantage of the complementary strengths of both depth images and skeleton data by developing an end-to-end learning framework, which is able to effectively learn the interactions between the human body-parts and the environmental objects. The experimental results demonstrate the reliability of the fusion model in the presence of significant changes in viewpoint and enhanced human action recognition accuracy. The work received follow-on funding from an Innovate UK project (£98K) with Digital Rail Ltd.
- Author contribution statement
- -
- Non-English
- No
- English abstract
- -
