Interframe_Variation_Vector Based Gait Recognition

Biometrics is a technology that uses the physiological or behavioral characteristics of people to authenticate their identities. Different from the first-generation biological features, such as face, fingerprint and iris, which are restricted to close distance detection, gait can be detected and measured under low-resolution at a distance. Also gait is hard to disguise and conceal, and it is non-invasive. All the above advantages make gait being the greatest potential biological feature in video surveillance field. With a growing demand for extensive visual surveillance and monitoring systems in security-sensitive environments, such as banks, airports, supermarkets and military bases, personal identification at a distance has attracted the growing interests of computer vision researchers.Gait recognition research is currently in its beginning. The term gait recognition is typically used to signify the identification of people in image sequences by the way they walk. The goal of this thesis is to investigate the information contained in the video sequences of human gait, and to perform personal identification based on the information. Focusing on this topic, this dissertation mainly includes the following issues:1) The status quo of gait recognition is summarized; the advantages and disadvantages of the main gait recognition algorithms are analyzed; the factors affecting gait are classified and the trends of gait recognition are discussed.2) Classical moving object detection approaches are outlined; Gaussian Mixture Model-based and Bayes model-based moving object detection algorithms are implemented and the comparative analysis of their performance is carried out; the seeded region growing technology is introduced to make up the holes caused by the updating strategies of the Bayes model-based algorithm and the algorithm's performance is improved greatly. A good foundation is constructed for the gait recognition of next stage.3) The framework of gait recognition is proposed and the improved detection approach described in chapter 3 is employed to extract gait features. The contours of the human body in a cycle are unwrapped and transformed into one-dimensional distance signals. By performing Principal Component Analysis technology, the human body contour's data can be successively reduced from 300-dimentional to ten-dimensional data, and IVVs(Interframe Variation Vector) are further extracted and used for gait recognition. The experiments show that the performance is very encouraging.