Fast stereo-based pedestrian detection using hypotheses

Abstract

In this paper, we present a multiple hypotheses framework to detect pedestrians accurately and precisely. The multiple hypotheses framework consists of obstacle detection, pedestrian recognition and data association. Obstacle detection detects all obstacles on the road. Pedestrian recognition classifies the detected obstacles as persons or non-persons. The data association component assigns multiple results to the correct hypotheses with multiple similarity functions. The experimental results demonstrate that the proposed method enhances the accuracy and precision of the region of interest. © 2015 ACM.