Bag-of-Words and Topic Modeling-Based Sport Video Analysis
This paper presents a method to perform team activity recognition in handball videos by using low level motion related features (position and direction of the motion), where a tracking process is not needed. Bag-of-words and topic modeling-based techniques have been used to characterize each video clip. Several parameter configurations have been tested to select the ones producing the best performance. An ensemble of selected classifiers has been constructed to obtain an overall accuracy rate of 98.38% in the recognition task among four different team activities.
KeywordsSport video analysis Team activity recognition Topic models Bag-of-words
Unable to display preview. Download preview PDF.
- 3.Blunsden, S., Fisher, B., Andrade, E.: Recognition of coordinated multi agent activities: the individual vs the group. In: Proc. of CVBASE 2006, pp. 61–70 (2006)Google Scholar
- 7.Fei-Fei, L., Perona, P.: A bayesian hierarchical model for learning natural scene categories. In: Proc. of IEEE CVPR 2005, vol. 2, pp. 524–531 (June 2005)Google Scholar
- 8.Sivic, J., Russell, B.C., Efros, A.A., Zisserman, A., Freeman, W.T.: Discovering object categories in image collections. In: Proc. of IEEE ICCV 2005 (2005)Google Scholar
- 12.Cristianini, N., Shawe-Taylor, J.: An Introduction to Support Vector Machines and Other Kernel-based Learning Methods, 1st edn. Cambridge University Press (2000)Google Scholar
- 13.Bradski, G., Kaehler, A.: Learning OpenCV: Computer Vision with the OpenCV Library. O’Reilly (2008)Google Scholar