Automatic Gait Recognition Using Kernel Principal Component Analysis

Gait is one of the biometric technologies which can be identified as an individual by his/her walking style. This paper proposes an effective gait recognition method based on mean gait energy image(MGEI) which utilizes kernel principal component analysis(KPCA). KPCA is capable of capturing part of the high-order statistics which are particularly important for MGEI structure. MGEI is calculated from gait cycle. KPCA can make use of the high correlation between the training samples and test samples for feature extraction by selecting the proper kernel function. And Euclidean distance of covariance weighted reciprocal is designed as the classifier. Comprehensive experiments are carried out on CASIA gait database and USF HumanID database. The experimental results demonstrate that the proposed approach has an encouraging recognition performance.