Statistical Learning of Multi-view Face Detection
ECCV '02 Proceedings of the 7th European Conference on Computer Vision-Part IV
A Bayesian Framework for Semantic Content Characterization
CVPR '98 Proceedings of the IEEE Computer Society Conference on Computer Vision and Pattern Recognition
Semantic Video Indexing Using a Probabilistic Framework
ICPR '00 Proceedings of the International Conference on Pattern Recognition - Volume 3
Rapid estimation of camera motion from compressed video with application to video annotation
IEEE Transactions on Circuits and Systems for Video Technology
Multimedia Tools and Applications
Hi-index | 0.00 |
Sports video is characterized with strict game rules, numerable events and well defined structures. In this paper, we proposed a generic framework for spatio-temporal pattern mining in sports video. Specifically, the periodicities in sports video are identified using unsupervised clustering and data mining method. In this way sports video analysis never needs priori domain knowledge about video genres, producers or predefined models. Therefore, such framework is easier to apply to various sports than supervised learning based approaches. In this framework, a hierarchical spatial pattern clustering routine, including scene-level clustering, field-level clustering and motion pattern clustering from top to bottom, is designed to label each subshot with coherent dominant motion. Then the temporal patterns are identified from such label sequence using data mining method. These mined probabilistic patterns are presented as basic structural elements of sports video.