The detection and rectification for identity-switch based on unfalsified control

TL;DR

This paper introduces a novel approach using unfalsified control to detect and correct ID-switch errors in multi-object tracking, enhancing robustness against occlusions and rapid movements.

Contribution

It proposes a new detection and rectification module based on appearance information variations and a strategy for ambiguous data association in MOT.

Findings

Effective ID-switch detection and correction demonstrated on public datasets

Improved robustness against occlusions and rapid movements

Enhanced tracking accuracy and stability

Abstract

The purpose of multi-object tracking (MOT) is to continuously track and identify objects detected in videos. Currently, most methods for multi-object tracking model the motion information and combine it with appearance information to determine and track objects. In this paper, unfalsified control is employed to address the ID-switch problem in multi-object tracking. We establish sequences of appearance information variations for the trajectories during the tracking process and design a detection and rectification module specifically for ID-switch detection and recovery. We also propose a simple and effective strategy to address the issue of ambiguous matching of appearance information during the data association process. Experimental results on publicly available MOT datasets demonstrate that the tracker exhibits excellent effectiveness and robustness in handling tracking errors caused by occlusions and rapid movements.