基于Contourlet直方图的目标跟踪算法

Object tracking based on contourlet histogram

提出一种基于Contourlet直方图的目标跟踪算法。对图像先进行Contourlet变换,并利用变换后的Contourlet系数建立Contourlet直方图,将其作为meanshift算法的迭代参数来实现目标跟踪。实验结果表明,本算法具有较好的鲁棒性,能够在遮挡、小目标等情况下实现快速准确的跟踪。

This paper proposes a target tracking algorithm based on Contourlet histogram. Contourlet histogram, which uses the Contourlet transform to extract image texture feature information, combined with meanshift algorithm is pro posed for target tracking. Experimental results show that the proposed algorithm is robust for small targets, occlusion and can achieve fast and accurate tracking.