Computer and Modernization ›› 2014, Vol. 0 ›› Issue (4): 64-68.

Previous Articles     Next Articles

 Human Posture Recognition Method Based on Multiple Classifiers

  

  1. School of Information Science and Engineering, Xiamen University, Xiamen 361005, China
  • Received:2014-03-03 Online:2014-04-17 Published:2014-04-23

Abstract:  

Abstract:  To get an accurate body posture recognition result and meet the needs of intelligent video surveillance, we present a fusion of Hu invariant moments and Fourier descriptors characteristics of human posture recognition algorithm. The ReliefF feature selection algorithm is introduced into the process, for distinguishing the importance of characteristics. Voting mechanism of using multiple classifier is constructed with large numbers to vote, plays good advantages of the various classifiers to improve the recognition accuracy. Finally, experimental results show that the proposed algorithm achieves good classification results for various postures.

Key words:

CLC Number: