Gait Recognition Based On Joint Points

With the rapid development of information technology, Biometrics has gradually goes into people's lives.It is not only simple and fast,but also safe and reliable. Gait refers to person's walking style.Compared to fingerprint identification, face recognitionand iris identification technology, gait characteristics, which captures the manner of human walking, is non-contact, difficult to camouflage and concealment, easily acquired at a distance.Therefore, from the view of video surveillance, gait recognition has been considered to be the most potential biometrics in the area of intelligent visual surveillance at a distance. It has gained increasing interest from computer vision researchers recently.The study of gait recognition technique could promote the development of the theory of computer vision and pattern recognition. It also has extensive applications such as the video surveillance in security-sensitive places, access control for special occasions and assistance to catch criminals, etc.The main purpose of this work is to improve both the identification performance and the verification performance of gait recognition algorithms, and also to furtherance the practical use of gait recognition techniques.Gait recognition is defined to identify the identity of the person as the way people walk and kinetic characteristics.It includes moving target detection,feature extraction,classification and recognition.Based on the study of the various of gait recognition algorithms, main research about the key stage- feature extraction has been done and a new algorithm has been proposed in this paper. In detail, the main contributions of this thesis are as follows:1. The existing methods of gait cycle estimation which are through achieving the cycle of autocorrelation function of aspect ratio are complex to obtain. Considering the real-time processing demand, this paper presents a simple and effective method of cycle estimation which is through calculating aspect ratio of human movement, furtherly, replaces key frames with Semi cycle long representative short sequences as the image sequence in the experiment. It expresses hundreds of frames gait sequence as half a cycle standard images of gait sequence concisely. It will not only standard the representation of data, but also eliminate the noise. Compared with the key frame method, it has a lot obvious advantages.2. Extraction of effective gait characteristics.Because of the impact of external factors, the recognition rate based on a single feature is unsatisfactory.In this paper, we have proposed a new algorithm for human gait recognition based on feature fusion to improve recognition rate. The joint points of the human body can be a good characterization of gait characteristics, for each sequence we describe the dynamic characteristics with joint angles of low limb and the static characteristics of their profile with the shape context descriptor in order to get better gait information. In the end, through the experiment, assign the weights corresponding to the different characteristics to carry out feature fusion and use K-nearest neighbor classifier to improve the recognition rate effectively. Large amount of experiments have shown that, both the identification performance and verification performance can be promoted to a certain degree by fusing, and the fused features outperform any single feature when it is used individually. The experimental results indicate that the classification performance of algorithm the can be further promoted by introducing multiple features.Experimental results show that the proposed algorithm has effective recognition performance.