Heuristic Maximum Clique Based Identity Switches Awareness for Tracking
Chao Li, Di Cui, Hao Sheng, Jiahui Chen, Yang Zhang · 2018
Multi-target tracking is a popular research topic in the field of computer vision.Most recent works based on tracking-by-detection paradigm have trouble when dealing with tracking errors (i.e.identity switches) that are caused by overlapping of different pedestrians.In this work, we propose a novel method which is capable to perceive the mismatched tracklets by using an Identity Switch Awareness Algorithm based on iterative maximum clique searching.Besides, we present an optimal method to accelerate the maximum clique searching process to alleviate the cost of computing.Finally, we introduce this algorithm into a network flow tracking method and test the validity of it on five challenging sequences.The results show obvious improvement on MOTA and IDs.