Multiple Object Tracking Using Re-Identification Model with Attention Module

Multi-object tracking (MOT) has gained significant attention in computer vision due to its wide range of applications. Specifically, detection-based trackers have shown high performance MOT, but they tend fail occlusive scenarios such as the moment when objects overlap or separate. In this paper, we propose a triplet-based MOT network that integrates information and visual features object. Using image feature, can differentiate similar-looking objects, reducing number identity switches over long period. Furthermore, an attention-based re-identification model focuses on appearance was introduced extract feature vectors from images effectively associate objects. The extensive experimental results demonstrated proposed method outperforms existing methods ID switch metric improves detection system.