The Study Of Gait Recognition Based On Image Sequence And Pressure

Over the past decade the security situation has become ever more complicated. With the enhancement of safety protection consciousness in mind, gait recognition is being taken into consideration by the researchers and important progress has been made in the field. Gait recognition has important significance in theory and practice. For the advantages of being non- cooperate, easily gained at a distance and difficult to conceal, gait recognition can be effectively used in many security-sensitive environments, such as airports, banks, car parks and so on.According to the features which are used for recognition, gait recognition can be roughly divided into two categories: vision-based gait approaches and tactility-based gait approaches. The studies on the vision-based gait approaches were carried out early, and many approaches have been proposed. Vision-based gait features are easily affected by the pedestrian and the environment because gait is a course of action, so it is hard to gain high performance in real world. Compared with vision-based gait features, tactility-based gait features are insusceptible to influences of external factors. We can gain the structural parameters effectively. The study of tactility-based gait approaches is in its beginning stage, there are many problems need to be further studied.The research of this article mainly focuses on the recognition algorithm based on gait image sequences and the fusion method to consider the vision-based information and tactility-based information.We design and develop a gait data acquisition system for collecting foot pressure and gait image sequences. We establish a new bimodal gait database which lays the foundation for next work.In this article, a new gait recognition based on Hough transform is proposed. We construct the gait templates in accumulator space which take account of more straight lines information in the gait intensity image.A cascade of classifiers was used for gait recognition to fuse the gait pressure information and the gait image sequences information. We adopt dynamic time warping to measure similarity between gait pressure curves.