TR2003-144

Generation of Sports Highlights Using a Combination of Supervised & Unsupervised Learning in Audio Domain


    •  Radhakrishan, R.; Xiong, Z.; Divakaran, A.; Ishikawa, Y., "Generation of Sports Highlights Using a Combination of Supervised & Unsupervised Learning in Audio Domain", IEEE Pacific-Rim Conference on Multimedia (PCM), December 2003, vol. 2, pp. 935-939.
      BibTeX Download PDF
      • @inproceedings{Radhakrishan2003dec,
      • author = {Radhakrishan, R. and Xiong, Z. and Divakaran, A. and Ishikawa, Y.},
      • title = {Generation of Sports Highlights Using a Combination of Supervised & Unsupervised Learning in Audio Domain},
      • booktitle = {IEEE Pacific-Rim Conference on Multimedia (PCM)},
      • year = 2003,
      • volume = 2,
      • pages = {935--939},
      • month = dec,
      • url = {http://www.merl.com/publications/TR2003-144}
      • }
  • Research Areas:

    Multimedia, Speech & Audio


In our past work we have used supervised audio classification to develop a common audio-based platform for highlight extraction that works across three different sports. We then use a heuristic to post-process the classification results to identify interesting events and also to adjust the summary length. In this paper, we propose a combination of unsupervised and supervised learning approaches to replace the heuristic. The proposed unsupervised framework mines the semantic audio-visual labels so as to detect "interesting" events. We then use a Hidden Markov Model based approach to control the length of the summary. Our experimental results show that the proposed techniques are promising.