Pedestrian Detection Based on Hybrid Features

In this paper, we propose a new approach for pedestrian detection in crowded scene from static images. The method is based on hybrid features, one type of middle-level features, which include haar-like features and gradient features, two low-level feature sets. The haar-like features focus on the local edges information of the image and the gradient features focus on the local regions information. We use two stages of Adaboost to train the final classifier. In the first stage, the whole image is divided into many small windows which all include numerous low-level features. Adaboost is used in each window to get one mid-level feature which composes of some best features including haar-like features and gradient features in this window. Secondly, from all mid-level features, Adaboost is used again to get the final classifier. Experiment results on common datasets and comparisons with some previous methods are given.