CVPR2020

Temporal-Context Enhanced Detection of Heavily Occluded Pedestrians

Jialian Wu, Chunluan Zhou, Ming Yang, Qian Zhang, Yuan Li, Junsong Yuan

Abstract

As shown in Fig. 1 , we visualize some proposal tubes and detection results on the Caltech dataset [2] . Compared to the baseline detector and the baseline+FGFA [1], our proposed method achieves better detection of heavily occluded pedestrians.