Unsupervised action classification using space-time link analysis
(2010) In Eurasip Journal on Image and Video Processing 2010. p.1-10- Abstract
We address the problem of unsupervised discovery of action classes in video data. Different from all existing methods thus far proposed for this task, we present a space-time link analysis approach which consistently matches or exceeds the performance of traditional unsupervised action categorization methods in various datasets. Our method is inspired by the recent success of link analysis techniques in the image domain. By applying these techniques in the space-time domain, we are able to naturally take into account the spatiotemporal relationships between the video features, while leveraging the power of graph matching for action classification. We present a comprehensive set of experiments demonstrating that our approach is capable... (More)
We address the problem of unsupervised discovery of action classes in video data. Different from all existing methods thus far proposed for this task, we present a space-time link analysis approach which consistently matches or exceeds the performance of traditional unsupervised action categorization methods in various datasets. Our method is inspired by the recent success of link analysis techniques in the image domain. By applying these techniques in the space-time domain, we are able to naturally take into account the spatiotemporal relationships between the video features, while leveraging the power of graph matching for action classification. We present a comprehensive set of experiments demonstrating that our approach is capable of handling cluttered backgrounds, activities with subtle movements, and video data from moving cameras. State-of-the-art results are reported on standard datasets. We also demonstrate our method in a compelling surveillance application with the goal of avoiding fraud in retail stores.
(Less)
- author
- Liu, Haowei
; Feris, Rogerio
; Krueger, Volker
LU
and Sun, Ming Ting LU
- publishing date
- 2010-07-20
- type
- Contribution to journal
- publication status
- published
- in
- Eurasip Journal on Image and Video Processing
- volume
- 2010
- article number
- 626324
- pages
- 10 pages
- publisher
- Springer
- external identifiers
-
- scopus:77954610621
- ISSN
- 1687-5176
- DOI
- 10.1155/2010/626324
- language
- English
- LU publication?
- no
- id
- 08450c06-6cb2-45ec-8965-f4c0e8e10e59
- date added to LUP
- 2019-06-28 09:23:34
- date last changed
- 2022-01-31 22:45:08
@article{08450c06-6cb2-45ec-8965-f4c0e8e10e59, abstract = {{<p>We address the problem of unsupervised discovery of action classes in video data. Different from all existing methods thus far proposed for this task, we present a space-time link analysis approach which consistently matches or exceeds the performance of traditional unsupervised action categorization methods in various datasets. Our method is inspired by the recent success of link analysis techniques in the image domain. By applying these techniques in the space-time domain, we are able to naturally take into account the spatiotemporal relationships between the video features, while leveraging the power of graph matching for action classification. We present a comprehensive set of experiments demonstrating that our approach is capable of handling cluttered backgrounds, activities with subtle movements, and video data from moving cameras. State-of-the-art results are reported on standard datasets. We also demonstrate our method in a compelling surveillance application with the goal of avoiding fraud in retail stores.</p>}}, author = {{Liu, Haowei and Feris, Rogerio and Krueger, Volker and Sun, Ming Ting}}, issn = {{1687-5176}}, language = {{eng}}, month = {{07}}, pages = {{1--10}}, publisher = {{Springer}}, series = {{Eurasip Journal on Image and Video Processing}}, title = {{Unsupervised action classification using space-time link analysis}}, url = {{http://dx.doi.org/10.1155/2010/626324}}, doi = {{10.1155/2010/626324}}, volume = {{2010}}, year = {{2010}}, }