A target tracking algorithm based on mean shift with feature fusion

Abstract

Classic Mean Shift tracking algorithm always suffers from large position errors, which may lead to the failure on tracking target in complex environment. To handle this problem, an improved Mean Shift tracking algorithm based on texture and color feature fusion is proposed. The histograms of improved Local Binary Patterns (LBP) texture and color features are calculated with the algorithm. Then, along with their similarity measuring functions, the tracking results of both LBP and color features are used to achieve the optimal target position. To solve the problem of full occlusion, Kalman filter is introduced. Experimental results show that the proposed algorithm is more robust and more adaptable than the classic Mean Shift and Particle Filter methods in complex environment, such as the similar background colors, rapid illumination changes and full occlusion.