Abstract

In this paper we propose the use of quantum genetic algorithm to optimize thesupport vector machine for human action recognition. The Microsoft Kinectsensor can be used for skeleton tracking, which provides the joints' positiondata. However, how to extract the motion features for representing the dynamicsof a human skeleton is still a challenge due to the complexity of human motion.We present a highly efficient features extraction method for actionclassification, that is, using the joint angles to represent a human skeletonand calculating the variance of each angle during an action time window. Usingthe proposed representation, we compared the human action classificationaccuracy of two approaches, inclduing the optimized SVM based on quantumgenetic algorithm and the conventional SVM with cross validation. Experiementalresults on the MSR-12 data show a higher accuracy in quantum genetic algorithmoptimized support vector machine.

 

Introduction (beta)

None

 

Conclusion (beta)

None