Level set tracking with fusion of color and dynamical shape priors

Level set tracking has been widely used for tracking contours of objects. However, traditional methods are sensitive to noise, partial occlusions, background disturbance and some other factors, especially there exist serious problems for non-rigid objects tracking. With respect to this point, we propose a level set-based tracking framework in which color and dynamical shape priors are fused together. The framework consists of two stages- off-line training and on-line tracking. During the off-line training stage, a graph-based dominant set clustering (DSC) method is applied to learn a shape codebook with each codeword representing a certain shape mode. Then a codeword transition matrix is learnt to characterize the temporal correlations of contours of an object. During the on-line tracking stage, firstly, the initial contour is evolved only with the color feature, a Markov random field (MRF)-like model measuring the correlations between neighboring pixels is added to improve the general region-based level set speed model; secondly, we fuse the knowledge of shape priors and the results obtained with the color feature, and adopt maximum a posteriori (MAP) estimation to predict the current shape mode; and finally the contours are evolved with the shape constraint continuously. The experimental results on synthetic and real video sequences demonstrate the effectiveness of our method.