Saliency Detection by Superpixel Ranking for Person Re-identification

Abstract

The research described in this paper consists in developing a person re-identification framework for multiple non-overlapping camera system. The proposed approach consists of three main steps. Firstly, human images are segmented into atomic regions using the concept of the superpixel. A saliency detection framework based on the manifold ranking is then carried out to estimate a saliency score map, which emphasizes the perceived important regions of an image. Finally, a flexible matching procedure is introduced to estimate the similarity between two images and to make the final decision of person re-identification. The performance of our system is evaluated on the well-known VIPeR dataset. The experimental results show that the proposed system leads to satisfactory results.