Output details
13 - Electrical and Electronic Engineering, Metallurgy and Materials
University of Surrey
Article title
Action Recognition Using Mined Hierarchical Compound Features
Type
D - Journal article
Title of journal
IEEE Transactions on Pattern Analysis and Machine Intelligence
Article number
-
Volume number
33
Issue number
5
First page of article
883
ISSN of journal
2160-9292
Year of publication
2011
URL
-
Number of additional authors
-
Additional information
This paper introduces a new approach to categorising/recognising actions in video data. Microsoft Kinect and the Video Google search engine are two examples of the potential value of methods for categorising and recognising actions in video data. This proposed work utilises machine-learning techniques and applies these to simple features to build larger distinctive structures. This work is evaluated on large scale, real-world datasets and out-perform state-of-the-art methods by a significant margin. The research is in the highest-ranking journal in the field and is becoming highly cited.
Interdisciplinary
-
Cross-referral requested
-
Research group
None
Proposed double-weighted
No
Double-weighted statement
-
Reserve for a double-weighted output
No
Non-English
No
English abstract
-