The gait is the emerging biometric technology, which is used for person authentication based on walking style of a person. Covariates play a very important role in gait recognition, which degrades the recognition accuracy. Covariates include View point, Clothes, Footwear, Surface type, Carried weight, Walk velocity, Time, Emotional state. Among these, we consider the variation of viewpoint and large intraclass variations like carrying and wearing conditions. Gait Energy Image (GEI) features are extracted from the binary silhouette images and perform the View Transformation Model (VTM), in order to recognize the person. In this paper, the experiments conducted on CASIA gait database, shows that the proposed algorithm is robust to view point and intraclass variations like carrying and wearing conditions.
展开▼
