Object Tracking by Maximizing Classification Score of Detector Based on Rectangle Features

In this paper, we proposed a novel classifierbased object tracker which combined a rectangular features based adaboost detector with optical-flow based tracking method, Support Vector Tracker. We show that gradient of extended rectangular features can be calculated rapidly by using integral image method. The proposed tracker was tested on real video sequences. We applied our tracker for face tracking and car tracking experiments. Our tracker worked over 100fps while maintaining comparable accuracy to rectangle features based detector. The tracking routine without I/O process reaches 500 to 2500 fps with sufficient accuracy. key words: Object Detection, Tracking, Support Vector Tracker, Rectangle Features, Boosting