Visual object tracking based on foreground segmentation and adaptive feature space selection
Ping Sheng · Kongzhi yu juece · 2010
A novel algorithm of object model update for visual tracking is presented. Firstly,feature histograms combined with spatial information are used to model the object and background. Then,for each feature space,the cross entropy measure of information theoretic is applied to evaluate the divergence between the distributions of object and background,and the feature space with maximal divergence is selected for tracking. The new position of the object is calculated by mean shift under the selected feature space. To alleviate the tracking drift problem,the model updating process is incorporated with the figure/ground segmentation based on systematic integration of spatial and temporal data over time by using conditional random field (CRF). The results of the experiments performed on several sequences show the effectiveness of the proposed method.