Content area
Full Text
Multimed Tools Appl (2015) 74:729742
DOI 10.1007/s11042-014-2177-x
Ruiyue Xu & Yepeng Guan & Yizhen Huang
Received: 4 November 2013 /Revised: 28 April 2014 /Accepted: 23 June 2014 / Published online: 5 October 2014 # Springer Science+Business Media New York 2014
Abstract Multiple human detection and tracking is a very important and active research topic in computer vision. At present, the recognition performance is not satisfactory, which is mainly due to the fact that the full-body of human cannot be captured efficiently by cameras. In this paper, an improved method is developed to detect and track multiple heads by considering them as rigid body parts. The appearance model of human heads is updated according to fusion of color histogram and oriented gradients. An associative mechanism of detection and tracking has been developed to recover transient missed detections and suppress transient false detections. The object identity can be kept invariant during tracking even if unavoidable occlusion occurs. Besides, the proposed method is fast to detect and track multiple human in a dynamic scene without any hypothesis for the scenario contents in advance. Comparisons with state-of-the-arts have indicated the superiority and good performance of the proposed method.
Keywords Human detection and tracking . Real-time . Associative mechanism
1 Introduction
Intelligent visual surveillance has been gaining more attention from the community due to the increasing importance and needs of crime prevention and anti-terrorist applications. Human detection and tracking from video sequences is naturally a key issue for intelligent visual surveillance. Many methods have been proposed for human detection and tracking so far: Zhao et al. [20] integrated fast gradient Hough transform, hair-color distribution model and circle existence model to detect human heads. Yang et al. [18] used skin color and head shape model to achieve this task. The color used in the method mentioned above may be confused with other objects in complex background. Besides, the above method is not suitable for complex situations such as several people moving in the scene with partial occlusion. Yuk et al. [19] proposed a probabilistic model based shape contour matching algorithm to detect
R. Xu : Y. Guan (*) : Y. Huang
School of Communication and Information Engineering, Shanghai University, Shanghai, China e-mail: [email protected]
Y. Guan
Key Laboratory of Advanced Displays and...