Action Recognition with Dynamic Image Networks
- Submitting institution
-
University of Edinburgh
- Unit of assessment
- 11 - Computer Science and Informatics
- Output identifier
- 58706229
- Type
- D - Journal article
- DOI
-
10.1109/TPAMI.2017.2769085
- Title of journal
- IEEE Transactions on Pattern Analysis and Machine Intelligence
- Article number
- -
- First page
- 2799
- Volume
- 40
- Issue
- 12
- ISSN
- 0162-8828
- Open access status
- Compliant
- Month of publication
- November
- 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
-
3
- Research group(s)
-
D - Language, Interaction and Robotics
- Citation count
- 42
- Proposed double-weighted
- No
- Reserve for an output with double weighting
- No
- Additional information
- First paper that can compress and learn to classify arbitrary length videos simultaneously with a deep neural network. It enables use of deep neural networks in action recognition problems such as surveillance, navigation where storing video data is prohibitive and there are real-time constraints on prediction. It is an extended version of highly cited earlier work (oral presentation in CVPR’16 with 3.9% acceptance rate). The algorithm has been shown to outperform the state-of-the-art in international benchmarks. The method has been widely adopted internationally for biometric matching (Nagrani 2018), human action (Cherian 2017, Wang 2018, Liu 2018), gesture (Wan 2017) recognition.
- Author contribution statement
- -
- Non-English
- No
- English abstract
- -