Jointly learning heterogeneous features for RGB-D activity recognition
- Submitting institution
-
University of Dundee
- Unit of assessment
- 11 - Computer Science and Informatics
- Output identifier
- 28402993
- Type
- D - Journal article
- DOI
-
10.1109/TPAMI.2016.2640292
- Title of journal
- IEEE Transactions on Pattern Analysis and Machine Intelligence
- Article number
- -
- First page
- 2186
- Volume
- 39
- Issue
- 11
- ISSN
- 0162-8828
- Open access status
- Compliant
- Month of publication
- December
- Year of publication
- 2016
- 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
-
3
- Research group(s)
-
-
- Citation count
- 46
- Proposed double-weighted
- No
- Reserve for an output with double weighting
- No
- Additional information
- This paper proposes a learning model that simultaneously explores shared and feature-specific components by jointly mining a set of subspaces for RGB-D activity recognition. This work has inspired, for example, the design of a new variant of LSTM for action recognition (Liu et al, IEEE Trans. PAMI, 2018) involving Alibaba Group, new developments in multimodal analysis (Shahroudy, et al, IEEE Trans. PAMI, 2018), and a sparse representation-based tracker (Lan et al, IEEE Trans IP, 2018). The contributed RGB-D, human-object interaction dataset presents novel challenges to the research community and has been used in over sixty papers for benchmarking.
- Author contribution statement
- -
- Non-English
- No
- English abstract
- -