Human motion recognition based on neural network

This paper proposes a method that can perform human motion pattern recognition using principal component analysis (PCA) and a neural network. The moving target is detected from a set of video image sequences, and the silhouette is extracted. The two-dimensional signal of the contour is converted into a one-dimensional signal through measuring the distance of pixels between centroid and boundaries of the silhouette. The feature of human motion is extracted by PCA, and then a neural network is employed to classify the motion pattern into three categories: walking, running, and other motions. Experimental results have shown that this method is capable of recognizing the motion pattern mentioned above effectively.