Chinese Journal of Computers   Full Text
Title	Errors of Bhattacharyya Coefficient and Its Reduction in Object Tracking
Authors	WEN Zhi-Qiang1),2) CAI Zi-Xing1)
Address	1)(School of Information Science & Engineering, Central South University, Changsha 410083) 2)(School of Computer & Communication, Hunan University of Technology, Zhuzhou, Hunan 412008)
Year	2008
Issue	No.7(1165—1174)
Abstract &   Background	Abstract Taylor approximate formula, which will bring some errors of Bhattacharyya coefficient in mean shift object tracking, will result in errors of object racking, so mean shift object tracking method do not fit for the case that object moves quickly. Errors of Bhattacharyya coefficient and its influence on object tracking are studied in this paper. Based the analysis on errors of Bhattacharyya coefficient, the optimization problem of Bhattacharyya coefficient is transformed into a constrained optimization problem for improving the precision of object tracking, so an improved object tracking method is presented in this paper. In addition, the convergence of the improved object tracking method is proved and more details about this method are given. Experiments on object tracking are done and correlative parameter is discussed by experiments. The experimental results show the improved method is good for fast moving object tracking. Keywords mean shift; object tracking; object model; Bhattacharyya coefficient Background The research in this paper is based on projects supported by National Natural Science Foundation of China under grant Nos.60234030, 60773110, and by the Fundamental Research Project of the ‘Eleventh Five-Year-Plan’ of China under Grant No.A1420060159. In field of robot navigation, there are two navigation methods as vision navigation and navigation based on distance measurement (such as laser radar, sonar, etc). The research of this paper focuses on vision navigation. As a nonparametric method mean shift is widely used in object tracking. But in object tracking method based on mean shift, Taylor approximate formula will bring some errors of Bhattacharyya coefficient, so this paper studies the errors of Bhattacharyya coefficient and its influence on object tracking, and presents an improved object tracking method for fast moving object tracking. The work of this paper is the significant part of research of dynamic object tracking in robotics field.