Detection of pedestrian crossing road using action classification model

Abstract

This paper presents a detection of pedestrian crossing road method using action classification model. The model incorporates the pedestrian pose recognition, lateral speed and spatial layout of the environment. The spatial body language ratio is used for recognize the pedestrian pose. The center of mass of the body relative to its width and height is used to define the pedestrian pose. Motion trajectory is obtained by using point tracking on the centroid of detected human region. And then estimated velocity is determined. Spatial layout is determined by the distance of the pedestrian to the road lane boundary. These models will be then hierarchically separated according to their action (walking, starting, bending and stopping). In order to classify the pedestrian crossing road, a walking human model is performed. A walking human is defined by ratio of the centroid location from the ground plane divided by the height of bounding box. It should satisfy a constraint. The proposed algorithms are evaluated using publicly (Caltech and ETH) datasets and our pedestrian dataset. The performance results shown the correct pedestrian crossing road classification is 98.10%.