Real-Time Multiple Pedestrian Tracking Based on Object Identification
Dohun Kim, Heegwang Kim, Jungsup Shin, Yeongheon Mok, Joonki Paik · 2019
In this paper, we present a novel real-time multiple pedestrian tracking based on object identification. The proposed method decides whether an object detected in the current frame is the same one in the previous frame, and up-dates the coordinate of multiple objects and the corresponding histogram during the tracking process. Experimental results show that the proposed multiple object tracking method out-performs existing methods in challenging situation with partial occlusions.