Gait extraction and recognition based on lower leg and ankle

摘要

This paper presents gait extraction and recognition method based on a regional information of the lower leg and ankle. First, according to the knowledge of human anatomy, extract lower leg and foot area where the contours of R3, and thin this region. Then the position of ankle is the intersection curve of lower leg and foot, and use least-square method to fit angle sequence of lower leg, and normalize these information and use Discrete Cosine Transform to transform the amplitude Angle sequences. Match two kinds of optimal characteristic, and process result by using feature fusion strategies. The results that experament in the NLPR gait database show: the lower leg and ankle is significant and effective in the gait feature.