Chinese Journal of Liquid Crystals and Displays, Volume. 40, Issue 8, 1202(2025)
Review of single object tracking algorithm based on deep learning
Figures & Tables(9)
Table 1. Comparison of representative algorithms based on traditional sequence models
View tableView in ArticleTable 1. Comparison of representative algorithms based on traditional sequence models
跟踪算法 AUC P MDNet 0.590 0.814 RTT 0.592 0.827 SANet 0.616 0.837 SiamFC 0.582 0.771 SiamRPN 0.568 0.748 SiamRPN++ 0.696 0.914 DaSiamRPN 0.658 0.880 SiamFC++ 0.683 0.896
Table 2. Comparison of representative algorithms based on the CNN-Transformer
View tableView in ArticleTable 2. Comparison of representative algorithms based on the CNN-Transformer
跟踪算法 AUC P TrDiMP 0.708 0.925 TransT 0.695 0.899 BANDT 0.706 0.911 TrTr 0.712 0.931 STARK 0.680 0.884 CSWinTT 0.671 0.872 AiATrack 0.696 0.917
Table 3. Comparison of representative algorithms based on the Transformer
View tableView in ArticleTable 3. Comparison of representative algorithms based on the Transformer
跟踪算法 AUC P SwinTrack 0.691 0.902 SparseTT 0.698 0.905 MixFormer 0.704 0.922 SimTrack 0.661 0.857 OSTrack 0.681 0.887 GRM 0.689 0.900 DropMAE 0.696 0.911 MixFormerV2 0.708 0.925 HIPTrack 0.710 0.933
Table 4. Comparison of the performance of different algorithms among three types on the LaSOT and GOT-10k datasets
View tableView in ArticleTable 4. Comparison of the performance of different algorithms among three types on the LaSOT and GOT-10k datasets
分类 跟踪算法 LaSOT GOT-10k FPS AUC P Pnorm AO SR0.5 SR0.75 基于传统序列的目标跟踪算法 SiamFC 0.336 0.339 0.420 0.348 0.353 0.098 86 MDNet 0.397 0.373 0.460 0.299 0.303 0.099 1 SiamRPN++ 0.496 0.569 0.491 0.517 0.616 0.325 5.17 SiamFC++ 0.560 0.566 0.651 0.595 0.695 0.473 45.27 基于CNN-Transformer的目标跟踪算法 TrDiMP 0.639 0.662 0.730 0.671 0.777 0.583 15.75 TransT 0.649 0.690 0.738 0.671 0.768 0.609 21.15 CSWinTT 0.662 0.709 0.752 0.694 0.789 0.542 8.76 STARK 0.671 0.722 0.769 0.688 0.781 0.641 17.90 AiATrack 0.690 0.738 0.794 0.696 0.800 0.632 31.22 基于Transformer的目标跟踪算法 MixFormer 0.700 0.763 0.799 0.707 0.800 0.678 8.02 OSTrack 0.687 0.746 0.781 0.736 0.830 0.717 - ARTrack 0.704 0.766 0.795 0.735 0.822 0.709 26 DropTrack 0.715 0.779 0.815 0.759 0.868 0.720 20.64 SeqTrack 0.725 0.792 0.815 0.748 0.819 0.722 5.81 AQATrack 0.719 0.802 0.826 0.760 0.852 0.749 - HIPTrack 0.727 0.774 0.829 0.774 0.880 0.745 -
Tools
Get Citation
Copy Citation Text
Shiyan GAO, Jie LIU, Wenyi CHEN, Zemin HE, Haiyan YANG, Zongcheng MIAO. Review of single object tracking algorithm based on deep learning[J]. Chinese Journal of Liquid Crystals and Displays, 2025, 40(8): 1202
Paper Information
Category:
Received: Apr. 10, 2025
Accepted: --
Published Online: Sep. 25, 2025
The Author Email: Zongcheng MIAO (miaozongcheng@nwpu.edu.cn)
© Copyright 2018-2021 | Chinese Laser Press.
All Rights Reserved 沪ICP备15018463号-20