A Determined Binary Level Set Method Based on Mean Shift for Contour Tracking

Traditional mean shift method has the limitation that could not effectively adjust kernel bandwidth to represent object accurately. To address this problem, in this paper, we propose a novel contour tracking algorithm using a determined binary level set model (DBLSM) based on mean shift procedure. In contrast with other previous work, the computational efficiency is greatly improved due to the simple form of the level set function and the efficient mean shift search. The DBLSM add prior knowledge of the target model to the implementation of curve evolution and ensure a more accurate convergence to the target. Then we use the energy function to measure weight for samples in mean shift framework. Experiment results on several challenging video sequences have verified the proposed algorithm is efficient and effective in many complicated scenes.