Linear latent low dimensional space for online early action recognition and prediction
- Submitting institution
-
Kingston University
- Unit of assessment
- 11 - Computer Science and Informatics
- Output identifier
- 11-33-1365
- Type
- D - Journal article
- DOI
-
10.1016/j.patcog.2017.07.003
- Title of journal
- Pattern Recognition
- Article number
- -
- First page
- 532
- Volume
- 72
- Issue
- -
- ISSN
- 0031-3203
- Open access status
- Compliant
- Month of publication
- -
- 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
-
-
- Research group(s)
-
-
- Citation count
- 14
- Proposed double-weighted
- No
- Reserve for an output with double weighting
- No
- Additional information
- Most action recognition methods deal with manually segmented video clips, the processing of which takes place offline. However, in many real-time applications such as gaming and human-computer interfaces, actions should be recognised online in real time and even, ideally, before they have been completed. This paper is significant because it introduces one of the very first methods that addresses the challenges of online and early action recognition and prediction. This research was part of the ICBISP-2017 keynote presentation: “Compact Representation of Multivariate Sequences using Structural Laplacian Eigenmaps”, http://icbisp2017.events.theiet.org.cn/?nxtId=247790.
- Author contribution statement
- -
- Non-English
- No
- English abstract
- -
