A Study on the Detection of Wheelchair Users Combined with Parts Tracking

In recent years, to support the self-support of wheelchair users, there has been an increasing demand for automatic detection of wheelchair users from a surveillance video. In a crowded scene where many pedestrians exist around, it is difficult to detect them because of occlusion by the pedestrians. In this report, we report on a detection method of wheelchair users robust to the occlusion. In case a detector cannot detect them, the proposed method estimates their location by combining it with parts tracking based on positional relation among parts and their trajectories. This improves their detection accuracy in a crowded scene. As a result of an experiment, the detection of wheelchair users with the proposed method achieved the highest accuracy in crowded scenes, compared with conventional methods.