- Book Chapter
- 10.3233/atde260195
Research on Multi-Object Tracking Algorithm Based on YOLO Combined DeepSORT
- Mar 13, 2026
- Haoyuan Yang + 1 more +1
We focused on researching multi-target tracking algorithms based on improved DeepSORT. The basic principles of multi-target tracking algorithms are introduced, followed by the basic architecture of DeepSORT. CIOU loss function, PP LCNet backbone, and SE attention mechanism are introduced to address the problems of previous algorithms, and an end-to-end tracking algorithm framework based on improved DeepSORT is constructed. This paper tells on the experimental index, criteria, and KITTI dataset. Through ablation experiments, the performance of individual modules is verified and improved. The obtained model is compared with baseline algorithms such as ByteTrack and BotSORT, and the improvement in accuracy, efficiency, and robustness of the improved model is very quantified.
Read more