Abstract:

In order to track a target in space and scale in a complex background, a target tracking algorithm based on the nonparametric clustering and multi-scale images is presented. In this algorithm, first, a modified nonparametric color-clustering method is employed to adaptively partition the color space of a tracked object, and the Gaussian function is used to model the spatial information of each bin of the color histogram. Next, the Bhattacharyya coefficient is adopted to derive a function describing the similarity between the target model and the target candidate. Then, a coarse-to-fine approach of multi-scale images is employed to implement the spatial location of the tracked object. Finally, the derived automatic bandwidth selection method of kernel function is applied to obtain the scale of the tracked object. Experimental results show that the proposed algorithm outperforms the classical mean shift tracker.

Key words: target tracking, mean shift, clustering algorithm, bandwidth allocation, multi-scale image, spatial localization, scale localization