Person Re-identification: System Design and Evaluation Overview [chapter]

Xiaogang Wang, Rui Zhao
2014 Person Re-Identification  
Person re-identification has important applications in video surveillance. It is particularly challenging because observed pedestrians undergo significant variations across camera views, and there are a large number of pedestrians to be distinguished given small pedestrian images from surveillance videos. This chapter discusses different approaches of improving the key components of a person reidentification system, including feature design, feature learning and metric learning, as well as
more » ... strength and weakness. It provides an overview of various person reidentification systems and their evaluation on benchmark datasets. Mutliple benchmark datasets for person re-identification are summarized and discussed. The performance of some state-of-the-art person identification approaches on benchmark datasets is compared and analyzed. It also discusses a few future research directions on improving benchmark datasets, evaluation methodology and system desgin. Person re-identification is to match pedestrian images observed in different camera views with visual features. The task is to match one or one set of query images with images of a large number of candidate persons in the gallery in order to recognize the identity of the query image (set). It has important applications in video surveillance including pedestrian search, multi-camera tracking and behaviour analysis. Under the settings of multi-camera object tracking, matching of visual features can be integrated with spatial and temporal reasoning [29, 32, 8] . This chapter focuses
doi:10.1007/978-1-4471-6296-4_17 dblp:series/acvpr/WangZ14 fatcat:6b6uon4g6ncf3amubxlwh3ouxu