Online Boosting For Car Detection

Boosting has become very popular in computer vision, showing impressive performance in detection and recognition tasks. Mainly off-line training methods have been used, which implies that all training data has to be a priori given; training and usage of the classifier are separate steps. Training the classifier on-line and incrementally as new data becomes available, which has several advantages and opens new areas of application for boosting in computer vision. In this paper we propose a novel on-line Boosting feature selection method. In conjunction with efficient feature extraction methods the method is real time capable. We demonstrate the method on such diverse tasks as learning complicated background models, visual tracking and object detection. All approaches benefit significantly by the on-line training. In contrast to related work, our framework does not rely on any priori knowledge of the image like model information, but if necessary this information can be incorporated.