Online multiple object tracking using joint detection and embedding network. (October 2022)

• Record Type: Journal Article
• Authors: Chan, Sixian; Jia, Yangwei; Zhou, Xiaolong; Bai, Cong; Chen, Shengyong; Zhang, Xiaoqin
• Is Part Of: Pattern recognition. Volume 130(2022)
• Publication Date: 2022-10
• Subjects: One-shot MOT -- Joint detection and tracking -- YOLO tracker; Pattern perception -- Periodicals; Perception des structures -- Périodiques; Patroonherkenning; 006.4
• DOI: 10.1016/j.patcog.2022.108793
• Languages: English
• ISSNs: 0031-3203

Abstract

Highlights: A novel architecture was proposed called YOLOTracker that performs online MOT by exploiting joint detection and embedding network, which shows the powerful and efficient performance in MOT benchmarks. The Path Aggregation Network was harnessed to combine the lowresolution and the high-resolution information for intergrating texture features and semantic information to mitigate alignment of the Re-ID features. The BNNeck is designed to avail the network for learning the best embeddings. The effectiveness of the two-stage progressive learning approach (BNNeck) is discussed and analyzed for ID-discriminative embedding of joinly trained detection and tracking. It effectively reduces the number of ID switches in MOT. The proposed tracker outperforms other state-of-the-art MOT trackers in terms of accuracy and efficiency on the three publicly challenging datasets of MOT15, MOT16 and MOT17. Abstract: Multiple object tracking (MOT) generally employs the paradigm of tracking-by-detection, where object detection and object tracking are executed conventionally using separate systems. Current progress in MOT has focused on detecting and tracking objects by harnessing the representational power of deep learning. Since existing methods always combine two submodules in the same network, it is particularly important that they must be trained effectively together. Therefore, the development of a suitable network architecture for the end-to-end joint training of detection and tracking submodules remains a challenging issue. The present work addresses this issue by proposing a novel architecture denoted as YOLOTracker that performs online MOT by exploiting a joint detection and embedding network. First, an efficient and powerful joint detection and tracking model is constructed to accomplish instance-level embedded training, which can ensure that the proposed tracker achieves highly accurate MOT results with high efficiency. Then, the Path Aggregation Network is employed to combine low-resolution and high-resolution features for integrating textural features and semantic information and mitigating the misalignment of the re-identification features. Experiments are conducted on three challenging and publicly available benchmark datasets and results demonstrate the proposed tracker outperforms other state-of-the-art MOT trackers in terms of accuracy and efficiency.