Soccer video event detection by fusing middle level visual semantics of an event clip

  • Authors:

  • Xueming Qian;Guizhong Liu;Huan Wang;Zhi Li;Zhe Wang

  • Affiliations:

  • Department of Information and Communication Engineering, Xi'an Jiaotong University, Xi'an, China;Department of Information and Communication Engineering, Xi'an Jiaotong University, Xi'an, China;Department of Information and Communication Engineering, Xi'an Jiaotong University, Xi'an, China;Department of Information and Communication Engineering, Xi'an Jiaotong University, Xi'an, China;Department of Information and Communication Engineering, Xi'an Jiaotong University, Xi'an, China

  • Venue:
  • PCM'10 Proceedings of the Advances in multimedia information processing, and 11th Pacific Rim conference on Multimedia: Part II
  • Year:
  • 2010

Quantified Score

Hi-index 0.00

Visualization

Abstract

Highlight event detection is a fundamental step of semantic based video retrieval and personalized sports video browsing. In this paper, an enhanced hidden Markov models (EHMM) based soccer video event detection method is proposed. Firstly, each soccer video shot is classified into one of the thirteen middle level semantics. Then the sequential soccer video sequence is segmented into event clips. Finally, HMMs are utilized to model the defined four highlights (goal, shoot, foul, and placed kick) and a normal kick. Not only the transitions of the middle level semantics and but also the overall features of an event clip are fused by HMMs to determine the event type. Comparisons are made with some existing soccer video event detection approaches. Experimental results show the effectiveness of the proposed EHMM based soccer video event detection approach. The influences of hidden state number and overall feature types to the event detection performances are discussed.