"Person re-identification refers to identify a particular person automatically by computer in the surveillance video who has occurred in the monitoring network,which is significantly important for the improvement of intelligence of video monitoring. But the research of person re-identification is not mature and facing many challenges. The following reasons may lead to a certain difference for the same person in different monitoring video images. For examples: the change of illumination in the monitoring environment, the different camera parameters in the monitoring network, the different shooting angle, and the different posture, etc.

These may lead to low recognition accuracy. This proposal launches an investigation on the following topics for improving the recognition accuracy: (1)Propose high-performance background modeling method by fusing pixel-based and region-based background modeling methods, and the time and space continuity between the video frames is considered. (2) Obtain reasonable and effective feature representation scheme which will start from both global and local features, and will fuse color, texture and shape features. (3) Design the proper similarity measure function based on the Markov distance model, which takes into account the characteristics of the fitness function. And the self-feedback evolutionary algorithm is proposed and applied to the feature matching search,which can assure the matching speed and reliability."