Multiple human tracking with alternately updating trajectories and multi-frame action features

Abstract

In this paper, we propose a multiple human tracking method with alternately updating trajectories and multi-frame action features (MHT-MAF). Even though occlusion or motion blur occurs due to the sudden movement of the drone, ID switches are prevented by the stable MAF. In the experiments, we verified the effectiveness of the proposed method using the Okutama-Action dataset. Our code is available online (https://github.com/hitottiez/mht-paf).