Abstract
We present a novel online tracking feature selection algorithm in this thesis. This algorithm performs more robust than the previous works by taking the correlation between features into consideration. Pixels of object/background regions are first treated as training samples. The feature selection problem is then modeled as finding a good subset of features and constructing a compound likelihood image with better discriminability for the tracking process. By adopting the AdaBoost algorithm, we iteratively select one best feature which compensate the previous selected features and linearly combine the set of corresponding likelihood images to obtain the compound likelihood image. Ellipse fitting is adopted to further improve the overall performance of tracking. We also propose an online feature validity test to monitor the features and only conduct feature selection when the selected features are not representative. We include the proposed algorithm into the mean shift based tracking system. Experimental results demonstrate that the proposed algorithm achieve very promising results.