Abstract
This paper proposes a method to discriminate between a bicycle rider and a pusher from a video sequence. It is important to discriminate a bicycle rider from a pusher in a monitoring camera system. However, in previous works using shape feature, it is difficult to discriminate between similar shape objects such as a bicycle rider and a pusher. The proposed method uses not only shape feature but also movement feature. The leg movements of a bicycle rider and a pusher are different. We apply Cubic Higher-order Local Auto-Correlation (CHLAC) feature to these leg movements to calculate the difference of these movements. We conduct an experiment in a video sequence to verify the effectiveness of this proposed method. As a result, we confirmed the effectiveness of the proposed method.
