Compensation Tracker: Reprocessing for Lost Object

TL;DR: Wang et al. as mentioned in this paper proposed a compensation tracker based on motion compensation and objects selection, which can be embedded into other non-end-to-end tracking frameworks, and achieved 66% MOTA and 67% IDF1 in the 2020 datasets.

Abstract: At present, the main research direction of multi-object tracking framework is tracking by detection. Although the detection-based tracking framework can achieve good results, it is very dependent on the performance of the detector. The tracking results will be affected to a certain extent when the detector has the behaviors of omission and error detection. Therefore, in order to solve the problem of missing detection, we designs a compensation tracker based on motion compensation and objects selection. Besides the tracker can be embedded into other non-end-to-end tracking frameworks. Experiments show that after using the compensation tracker designed in this paper, evaluation indicators have improved in varying degrees on MOT Challenge datasets. With limit cost, the compensation tracker haves reached 66% MOTA and 67% IDF1 in the 2020 datasets of dense scenarios. This shows that the proposed method can effectively improve the tracking performance of the model.