PEDESTRIAN DETECTION USING ADABOOST LEARNING OF FEATURES AND VEHICLE PITCH ESTIMATION

In this paper we propose a combination of different Haar filter sets and Edge Orientation Histograms (EOH) in or-der to learn a model for pedestrian detection. As we will show, with the addition of EOH we obtain better ROCs than using Haar filters alone. Hence, a model consist-ing of discriminant features, selected by AdaBoost, is ap-plied at pedestrian-sized image windows in order to per-form the classification. Additionally, taking into account the final application, a driver assistance system with real-time requirements, we propose a novel stereo-based cam-era pitch estimation to reduce the number of explored win-dows. With this approach, the system can work in urban roads, as will be illustrated by current results. KEY WORDS ADAS, pedestrian detection, adaboost learning, pitch esti-mation, haar wavelets, edge orientation histograms.