Model-based validation approaches and matching techniques for automotive vision based pedestrian detection

Pedestrian detection is a challenging vision task, especially applied to the automotive field where the background changes as the vehicle moves. This paper presents an extensive study upon human body models and the techniques suitable for being used in a pedestrian detection system. Several different approaches for building model sets, such as synthetic, real, and dynamic sets are presented and discussed. Comparative results are reported with reference to a case study of a real system. Preliminary results of current research status are shown together with further developments.