TraSw: Tracklet-Switch Adversarial Attacks against Multi-Object Tracking

Multi-Object Tracking (MOT) has achieved aggressive progress and derives many excellent deep learning models. However, the robustness of the trackers is rarely studied, and it is challenging to attack the MOT system since its mature association algorithms are designed to be robust against errors during the tracking. In this work, we analyze the vulnerability of popular pedestrian MOT trackers and propose a novel adversarial attack method called Tracklet-Switch (TraSw) against the complete tracking pipeline of MOT. TraSw can fool the advanced deep trackers (i.e., FairMOT and ByteTrack) to fail to track the targets in the subsequent frames by attacking very few frames. Experiments on the MOT-Challenge datasets (i.e., 2DMOT15, MOT17, and MOT20) show that TraSw can achieve an extraordinarily high success rate of over 95% by attacking only four frames on average. To our knowledge, this is the first work on the adversarial attack against pedestrian MOT trackers. The code is available at https://github.com/DerryHub/FairMOT-attack .