Biometric authorization system by video analysis of human gait in controlled environments

Using biometrics in developing new tools for security enhancement is one of the most important research topic in the computer vision community. This paper shows how human gait is used to recognize individuals when they enter a surveillance area, by a novel hybrid holistic approach. Video sequences captured by a static camera deployed in surveillance area are used to model the background, which is a key step for segmenting foreground moving objects. Then gait representing spatial, temporal, and wavelet features are extracted and fused for training and testing the multi class Probabilistic Neural Network model (PNN). The proposed system is evaluated using side view videos of Chinese National Laboratory of Pattern Recognition (NLPR) gait database and experimental results demonstrate the effectiveness of our approach. At the background modelling and foreground extraction the reliability and computational cost of the algorithm presented are comparatively discussed with several algorithms.