Encoding The Regional Features For Person Re-identification Using Locality-constrained Linear Coding

With the development of science and technology,large-scale distributed multi-camera surveillance network plays an important role in people's life,ensuring the public security.Person re-identification,which refers to recognizing the same person who moves cross the non-overlapping fields of different cameras,is one of the latest topics in surveillance system and can be used in applications such as suspect retrieval.However,due to the variability of environmental conditions,person re-identification is still a challenging problem.In this paper,technologies of image processing and computer vision are used to solve the person re-identification problem: As for pedestrian detection,HOG and SVM is employed to find pedestrian in videos;In terms of feature extraction,an image is described by color histogram and texture with the help of image segmentation,kernel function and spatial pyramid;In case of classification and matching,an effective feature coding algorithm is implemented to build the corresponding relations between probe image and gallery set.In practical application,real-time response is as important as re-identification rate.Nonetheless,most person re-identification researches care less about computation time cost in their work.This paper proposes an algorithm not only achieves more accurate re-identification rate but also lower computation.In order to prove the effectiveness of the proposed method,two kinds of experiments have been conducted: one is on four public image datasets;another is on a self-collected video dataset.The results confirmed that this work has practical application value.Moreover,a visualization software,which can be provided to customer,is developed using C++.