Abstract:

For human action recognition,an algorithm for classifying human action based on back propagation (BP) neural network was proposed.Firstly,singular value decomposition (SVD) was used to extract the singular value of each frame of the video.Then each row of a matrix was composed of the singular value of each video.Every single row of the matrix is a sample of human action.Secondly,the principle component analysis (PCA) algorithm was proposed to remove correlation and reduce dimension.Then the linear discriminant analysis (LDA) algorithm was applied to matrix processed by PCA to make samples linearly separable.Finally,the back propagation neural network was used as a classifier.The experimental results show that the proposed method,compared with nearest neighbor classifier and K-nearest neighbor (kNN) classifier,has a higher recognition rate.

Key words: human action recognition, SVD, PCA, LDA, back propagation neural network