Trainable Pedestrian Detection

Robust, fast object detection systems are critical to the success of next-generation automotive vision systems. An importantcriteria is that the detection systembe easily con- figurableto a new domain or environment. In this paper, we present work on a general object detection system that can be trained to detect different types of objects; we will focus on the task of pedestrian detection. This paradigm of learning from examples allows us to avoid the need for a hand-craftedsolution. Unlike many pedestrian detection systems, the core technique does not rely on motion infor- mationand makesno assumptionson the scene structure or the number of objects present. We discuss an extension to the system that takes advantage of dynamical information when processing video sequencesto enhanceaccuracy. We also describe a real, real-time version of the system that has been integrated into a DaimlerChryslertest vehicle.