Region-of-interest reduction using edge and depth images for pedestrian detection in urban areas

For real-time pedestrian detection, it is important to identify a relatively small set of region-of-interests (ROIs) accurately improve the computational efficiency. In this paper, we propose a ROI reduction method that exploits edges and depth information obtained from stereo images. In the proposed method, special features of urban areas along with depth and ground plane information are used to reduce the number of ROIs in stereo-based pedestrian detection. Experimental results show that the proposed method improves the speed of pedestrian detection while keeping the detection performance.