Gait recognition based on gait energy image and linear discriminant analysis

Aiming at the low rate of human gait recognition and other issues, a novel gait recognition based on gait energy image and linear discriminant analysis is proposed in this paper. Firstly, using background subtraction to segment contours of human body, then a gait cycle is computed with the change of body silhouette width; Secondly, using the images in gait cycle sequences to calculate the gait energy image (GEI); Finally, principal component analysis (PCA) and linear discriminant analysis (LDA) are used to extract the main vector of gait feature, the nearest neighbor classifier is selected for classification and identification. The experimental results indicate that our method can improve the recognition rate effectively on CASIA gait database.