Object tracking across non-overlapping views by learning inter-camera transfer models.pdfVIP

Pattern Recognition 47 (2014) 1126–1137 Contents lists available at ScienceDirect Pattern Recognition journal homepage: /locate/pr Object tracking across non-overlapping views by learning inter-camera transfer models$ Xiaotang Chen n, Kaiqi Huang, Tieniu Tan Center for Research on Intelligent Perception and Computing, National Laboratory of Pattern Recognition, Institute of Automation, Chinese Academy of Sciences, No.95 ZhongGuanCun East Street, HaiDian District, P.O. Box 2778, Beijing, PR China article info Available online 24 June 2013 Keywords: Object tracking Transfer models Color transfer Camera network Non-overlapping views abstract In this paper, we introduce a novel algorithm to solve the problem of object tracking across multiple nonoverlapping cameras by learning inter-camera transfer models. The transfer models are divided into two parts according to different kinds of cues, i.e. spatio-temporal cues and appearance cues. To learn spatiotemporal transfer models across cameras, an unsupervised topology recovering approach based on N-neighbor accumulated cross-correlations is proposed, which estimates the topology of a non-overlapping multi-camera network. Different from previous methods, the proposed topology recovering method can deal with large amounts of data without considering the size of time window. To learn inter-camera appearance transfer models, a color transfer method is used to model the changes of color characteristics across cameras, which has an advantage of low requirements to training samples, making update ef?cient when illumination conditions change. The experiments are performed on different datasets. Experimental results demonstrate the effectiveness of the proposed algorithm. 2013 Elsevier Ltd. All rights reserved. 1. Introduction As the number of cameras used in the wide area video surveillance increases, multi-camera object tracking plays a more important role in understanding and analyzing the scenes. It is a challenging problem. E