Multi-person gait recognition system based on Kinect

Gait is a state of the art biometric modality for recognizing humans. Unlike other biometric modality, gait information can be captured in a large distance, and it's impossible to hide or imitate. This work presented a research which the objective was to develop a prototype biometric system to use gait recognition to recognize human, for identification scenarios, using a depth camera such as the Kinect Camera of Microsoft. Proposed approach used PCA algorithm to extract gait features and categorized these features to identify humans by SVM and it achieves recognition performance of above 90% in an unrestricted indoor ambience.