Boosting and Vision Helmut GRABNER 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 has several advantages and opens new areas of application for boosting in computer vision. In this talk, I first present the on-line AdaBoost for feature selection method, guiding you from off-line boosting to on-line boosting and finally show how boosting can be applied to perform efficient feature selection. In conjunction with fast feature calculation using efficient data structures the algorithm is real time capable. The second part of the talk demonstrates the multifariousness of the method on such diverse tasks as visual tracking, learning complex background models, and object detection.