Adaptive Tracking of Objects in Image Sequences Using Improved Mean Shift Algorithm
Zhao Rongchun · Jisuanji yingyong yanjiu · 2005
Current Mean Shift tracking algorithm uses a single radius parameter to describe the scale of the object. This is not suitable for complex movements of objects. Provides a new algorithm in which bandwidth matrix is employed to describe the objects with two directions determined independently. The strict proof of the algorithm's convergence is given. Experimental results show the better adaptability of the new algorithm to object's movements without additional computational cost.