[1]李欣桐,刘航,黄德天.采用剪枝策略的目标跟踪方法[J].华侨大学学报(自然科学版),2025,46(5):581-588.[doi:10.11830/ISSN.1000-5013.202509002]
 LI Xintong,LIU Hang,HUANG Detian.Object Tracking Method Using Pruning Strategy[J].Journal of Huaqiao University(Natural Science),2025,46(5):581-588.[doi:10.11830/ISSN.1000-5013.202509002]
点击复制

采用剪枝策略的目标跟踪方法()
分享到:

《华侨大学学报(自然科学版)》[ISSN:1000-5013/CN:35-1079/N]

卷:
第46卷
期数:
2025年第5期
页码:
581-588
栏目:
出版日期:
2025-09-20

文章信息/Info

Title:
Object Tracking Method Using Pruning Strategy
文章编号:
1000-5013(2025)05-0581-08
作者:
李欣桐 刘航 黄德天
华侨大学 工学院, 福建 泉州 362021
Author(s):
LI Xintong LIU Hang HUANG Detian
College of Engineering, Huaqiao University, Quanzhou 362021, China
关键词:
目标跟踪 Transformer 剪枝策略 计算效率 注意力机制
Keywords:
target tracking Transformer pruning strategy computational efficiency attention mechanism
分类号:
TP181
DOI:
10.11830/ISSN.1000-5013.202509002
文献标志码:
A
摘要:
为了解决现有的基于Transformer的目标跟踪方法重精度轻效率而导致模型的计算开销过大的问题,提出一种基于剪枝策略的目标跟踪方法(PSTrack)。首先,引入激活模块,对Transformer的注意力层进行自适应调整,以避免不必要的计算,提升计算效率。然后,采用混合注意力与交叉注意力机制,根据不同阶段的需求优化计算流程。最后,在多个标准数据集上进行实验验证。结果表明:相较于传统的目标跟踪方法,PSTrack在精度和效率上均具备明显优势。
Abstract:
To address the issue of excessive computational overhead in existing based Transformer object tracking methods that prioritize accuracy over efficiency, a pruning strategy-based object tracking method(PSTrack)is proposed. First, an activation module is introduced to adaptively adjust the attention layers of the Transformer, thereby avoiding unnecessary computations and improving efficiency. Then, a hybrid attention and cross-attention mechanism is employed to optimize the computational process according to the requirements of different stages. Finally, experiments are conducted on multiple standard datasets. The results show that, compared with traditional object tracking methods, PSTrack achieves significant advantages in both accuracy and efficiency.

参考文献/References:

[1] DANELLJAN M,BHAT G,KHAN F S,et al.ECO: Efficient convolution operators for tracking[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition.Las Vegas:IEEE Press,2017:6638-6646.DOI:10.1109/CVPR.2017.733.
[2] DANELLJAN M,BHAT G,KHAN F S,et al.ATOM: Accurate tracking by overlap maximization[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition.Long Beach:IEEE Press,2019:4660-4669.DOI:10.1109/CVPR.2019.00479.
[3] YE Botao,CHANG Hong,MA Bingpeng,et al.Joint feature learning and relation modeling for tracking: A one-stream framework[C]//Proceedings of the European Conference on Computer Vision.Cham:Springer,2022:341-357.DOI:10.1007/978-3-031-20047-2_20.
[4] CHEN Boyu,LI Peixia,BAI Lei,et al.Backbone is all you need: A simplified architecture for visual object tracking[C]//Proceedings of the European Conference on Computer Vision.Cham:Springer,2022:375-392.DOI:10.1007/978-3-031-20047-2_22.
[5] CUI Yutao,JIANG Cheng,WU Gangshan,et al.MixFormer: End-to-end tracking with iterative mixed attention[J].IEEE Transactions on Pattern Analysis and Machine Intelligence,2024,46(6):4129-4146.DOI:10.1109/TPAMI.2024.3349519.
[6] WU You,WANG Xucheng,ZENG Dan,et al.Learning motion blur robust vision transformers with dynamic early exit for real-time UAV tracking[EB/OL].(2024-07-07)[2025-09-01] .https://arxiv.org/abs/2407.05383.
[7] YANG Xiangyang,ZENG Dan,WANG Xucheng,et al.Adaptively bypassing vision transformer blocks for efficient visual tracking[J].Pattern Recognition,2025,161:111278.DOI:10.1016/j.patcog.2024.111278.
[8] XUE Chaocan,ZHONG Bineng,LIANG Qihua,et al.Similarity-guided layer-adaptive vision transformer for UAV tracking[EB/OL].(2025-03-09)[2025-09-01] .https://arxiv.org/abs/2503.06625.
[9] HUANG Lianghua,ZHAO Xin,HUANG Kaiqi.GOT-10k: A large high-diversity benchmark for generic object tracking in the wild[J].IEEE Transactions on Pattern Analysis and Machine Intelligence,2019,43(5):1562-1577.DOI:10.1109/TPAMI.2019.2957464.
[10] FAN Heng,LIN Liting,YANG Fan,et al.LaSOT: A high-quality benchmark for large-scale single object tracking[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition.Long Beach:IEEE Press,2019:5374-5383.DOI:10.1109/CVPR.2019.00552.
[11] MUELLER M,BIBI A,GIANCOLA S,et al.TrackingNet: A large-scale dataset and benchmark for object tracking in the wild[C]//Proceedings of the European Conference on Computer Vision.Cham:Springer,2018:300-317.
[12] LIN T Y,MAIRE M,BELONGIE S,et al.Microsoft COCO: Common objects in context[C]//Proceedings of the European Conference on Computer Vision.Cham:Springer,2014:740-755.DOI:10.1007/978-3-319-10602-1_48.
[13] LOSHCHILOV I,HUTTER F.Decoupled weight decay regularization[EB/OL].(2017-11-14)[2025-09-01] .https://arxiv.org/abs/1711.05101.
[14] KIANI G H,FAGG A,HUANG C,et al.Need for speed: A benchmark for higher frame rate object tracking[C]//Proceedings of the IEEE International Conference on Computer Vision.Venice:IEEE Press,2017:1125-1134.DOI:10.1109/ICCV.2017.128.
[15] MUELLER M,SMITH N,GHANEM B.A benchmark and simulator for UAV tracking[C]//Proceedings of the European Conference on Computer Vision.Cham:Springer,2016:445-461.DOI:10.1007/978-3-319-46448-0_27.
[16] YAN Bin,PENG Houwen,WU Kaiwen,et al.LightTrack: Finding lightweight neural networks for object tracking via one-shot architecture search[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition.Los Alamitos:IEEE Press,2021:15180-15189.DOI:10.1109/CVPR46437.2021.01493.
[17] BORSUK V,VEI R,KUPYN O,et al.FEAR: Fast, efficient, accurate and robust visual tracker[C]//Proceedings of the European Conference on Computer Vision.Cham:Springer,2022:644-663.
[18] CHEN Xin,KANG Ben,WANG Dong,et al.Efficient visual tracking via hierarchical cross-attention transformer[C]//Proceedings of the European Conference on Computer Vision.Cham:Springer,2022:461-477.DOI:10.1007/978-3-031-25085-9_26.
[19] BLATTER P,KANAKIS M,DANELLJAN M,et al.Efficient visual tracking with exemplar transformers[C]//Proceedings of the IEEE Winter Conference on Applications of Computer Vision.Los Alamitos:IEEE Press,2023:1571-1581.DOI:10.1109/WACV56688.2023.00162.
[20] KANG Ben,CHEN Xin,WANG Dong,et al.Exploring lightweight hierarchical vision transformers for efficient visual tracking[C]//Proceedings of the IEEE International Conference on Computer Vision.Paris:IEEE Press,2023:9612-9621.DOI:10.1109/ICCV51070.2023.00881.
[21] CHEN Xin,YAN Bin,ZHU Jiawen,et al.Transformer tracking[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition.Los Alamitos:IEEE Press,2021:8126-8135.DOI:10.1109/CVPR46437.2021.00803.
[22] YAN Bin,PENG Houwen,FU Jianlong,et al.Learning spatio-temporal transformer for visual tracking[C]//Proceedings of the IEEE International Conference on Computer Vision.Montreal:IEEE Press,2021:10428-10437.DOI:10.1109/ICCV48922.2021.01028.
[23] LIN Liting,FAN Heng,ZHANG Zhipeng,et al.SwinTrack: A simple and strong baseline for Transformer tracking[J].Advances in Neural Information Processing Systems,2022,35:16743-16754.
[24] WEI Xing,BAI Yifan,ZHENG Yongchao,et al.Autoregressive visual tracking[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition.Vancouver:IEEE Press,2023:9697-9706.DOI:10.1109/CVPR52729.2023.00935.
[25] KUGARAJEEVAN J,KOKUL T,RAMANAN A,et al.Optimized information flow for Transformer tracking[EB/OL].(2024-02-13)[2025-09-01] .https://arxiv.org/abs/2402.08195.

相似文献/References:

[1]钟必能,潘胜男.选择性搜索和多深度学习模型融合的目标跟踪[J].华侨大学学报(自然科学版),2016,37(2):207.[doi:10.11830/ISSN.1000-5013.2016.02.0207]
 ZHONG Bineng,PAN Shengnan.Multi-Clue Fusion Target Tracking Algorithm Based on Selective Search and Deep Learning[J].Journal of Huaqiao University(Natural Science),2016,37(5):207.[doi:10.11830/ISSN.1000-5013.2016.02.0207]
[2]刘韶涛,姚灿荣.结合PSO的改进压缩跟踪方法[J].华侨大学学报(自然科学版),2017,38(1):121.[doi:10.11830/ISSN.1000-5013.201701024]
 LIU Shaotao,YAO Canrong.Improved Compress Tracking Algorithm Based on PSO[J].Journal of Huaqiao University(Natural Science),2017,38(5):121.[doi:10.11830/ISSN.1000-5013.201701024]
[3]黄德天,顾培婷,柳培忠,等.改进的自适应核相关滤波目标跟踪[J].华侨大学学报(自然科学版),2017,38(5):693.[doi:10.11830/ISSN.1000-5013.201606062]
 HUANG Detian,GU Peiting,LIU Peizhong,et al.Improved Adaptive Target Tracking Based on Kernelized Correlation Filters[J].Journal of Huaqiao University(Natural Science),2017,38(5):693.[doi:10.11830/ISSN.1000-5013.201606062]
[4]顾培婷,黄德天,黄炜钦,等.抗遮挡的相关滤波目标跟踪算法[J].华侨大学学报(自然科学版),2018,39(4):611.[doi:10.11830/ISSN.1000-5013.201608031]
 GU Peiting,HUANG Detian,HUANG Weiqin,et al.Anti-Occlusion Object Tracking Algorithm Based on Kernelized Correlation Filters[J].Journal of Huaqiao University(Natural Science),2018,39(5):611.[doi:10.11830/ISSN.1000-5013.201608031]

备注/Memo

备注/Memo:
收稿日期: 2025-09-01
通信作者: 黄德天(1985-),男,副教授,博士,主要从事计算机视觉、深度学习的研究。E-mail:huangdetian@hqu.edu.cn。
基金项目: 国家自然科学基金资助项目(61901183); 国家重点研发计划(2021YFE0205400); 福建省自然科学基金资助项目(2023J01140); 福建省泉州市科技计划重点项目(2023C007R)
更新日期/Last Update: 2025-09-20