Improved Camshift with adaptive searching window

Camshift is widely used real-time algorithm in video target tracking field. The size of searching window (SW) is a key factor of Camshift, and bigger or smaller size of SW will both decrease the real-time feature of Camshift. In this paper, a accelerated Camshift with adaptive searching window (ACASW) was proposed. Firstly the meanshift process and computational cost (CC) were modeled, and the relationship between the size of SW and CC was analyzed quantificationally, then the optimized size of SW was deduced, which was used in the proposed algorithm. From the experimental results it can be seen that, compared to the traditional Camshift with fixed EW, the proposed algorithm can reduce the computing time effectively, which improve the real-time feature of the algorithm.