A new method of pedestrian gait classification

Abstract

Gait classification is one of the hottest but most difficult subjects in computer vision. In order to identify pedestrian movement in an Intelligent Security Monitoring System, moving body is detected and the boundary is extracted. The paper proposes a complex number notation based on centroid in order to indicate a pedestrian's postures. And according to the different sorts of gaits, a set of different standard pedestrian posture contours is made. Different gait matrices based on spatio-temporal are acquired through Hidden Markov Models (HMM). A Procrustes distance analysis method is presented in order to get the degree to which two contours are resembled. Finally Fuzzy Associative Memory (FAM) is proposed to infer behavior classification of a walker. In this paper, an evaluation of ten kinds of different gaits is given with a 76.7% recognition rate.