基于前序特征提示的单目标跟踪算法

郑颖; 马国梁; 郭健; 王梓屹

doi:10.12382/bgxb.2025.0829

您当前的位置：

首页 >

文章列表页 >

基于前序特征提示的单目标跟踪算法

更新时间：2026-04-23

- 基于前序特征提示的单目标跟踪算法
- Single Object Tracking Algorithm Based on Preceding-Feature Prompts
- 兵工学报 2026年
- 作者机构：
  
  1. 南京理工大学自动化学院,江苏,南京,210094
  2. 南京理工大学能源与动力工程学院,江苏,南京,210094
- 作者简介：
- 基金信息：
  
  江苏省科技重大专项项目(BG2024041)
- DOI：10.12382/bgxb.2025.0829
  中图分类号： TP391
- 收稿：2025-09-09，
  
  网络首发：2026-04-23，
- 稿件说明：
移动端阅览
郑颖，马国梁，郭健，等. 基于前序特征提示的单目标跟踪算法[J/OL]. 兵工学报, 2026(2026-04-23). https://doi.org/10.12382/bgxb.2025.0829.

ZHENG Y, MA G L, GUO J, et al. Single object tracking algorithm based on preceding-feature prompts[J/OL]. Acta Armamentarii, 2026(2026-04-23). https://doi.org/10.12382/bgxb.2025.0829. (in Chinese)
郑颖，马国梁，郭健，等. 基于前序特征提示的单目标跟踪算法[J/OL]. 兵工学报, 2026(2026-04-23). https://doi.org/10.12382/bgxb.2025.0829. DOI：

ZHENG Y, MA G L, GUO J, et al. Single object tracking algorithm based on preceding-feature prompts[J/OL]. Acta Armamentarii, 2026(2026-04-23). https://doi.org/10.12382/bgxb.2025.0829. (in Chinese) DOI：

摘要

现有单目标跟踪算法普遍依赖初始帧模板与当前帧搜索区域之间的相似性度量进行匹配，在复杂场景中易出现跟踪漂移和误匹配问题，为此提出一种基于前序特征提示的Transformer目标跟踪算法。通过广义关系建模模块对搜索区域内的目标与背景令牌进行自适应划分，仅允许目标令牌与模板令牌交互以有效抑制相似目标及环境噪声干扰；在Transformer目标跟踪算法框架中新增加了前序特征提示模块，结合令牌划分结果与当前帧预测结果生成精确的目标蒙版，并利用前序特征提示编码器对蒙版与搜索图像的拼接结果进行联合编码；设计的前序特征提示解码器根据当前帧搜索区域特征自适应聚合编码结果生成特定提示，从而提升了跟踪算法的适应能力。采用多个基准数据集进行测试。研究结果表明：所提算法在GOT-10K数据集上平均重叠率达到76.1%，在LaSOT数据集上成功率曲线下面积为72.3%、精准度为82.4%；所提算法跟踪性能较优，能够有效解决单目标跟踪时的遮挡、形变和相似干扰等问题。

Abstract

Most existing single-object tracking algorithms rely on similarity matching between the initial-frame template and the search region of the current frame. However

in complex scenarios

this often leads to tracking drift and mismatches. To address these issues

we propose a Transformer-based tracking algorithm enhanced with preceding-feature prompts. A generalized relation modeling module adaptively partitions target and background tokens within the search region

allowing only target tokens to interact with template tokens. This effectively suppresses interference from similar objects and environmental noise. Within the Transformer tracking framework

we introduce apreceding-featureprompt module that combines token partitioning results with the current prediction to generate precise target masks. These masks are jointly encoded with the search image by a prior feature prompt encoder. A corresponding decoder then adaptively aggregates the encoded information with search-region features to produce task-specific prompts

thereby improving adaptability. Evaluations on multiple benchmark datasets show that our method achieves an average overlap of 76.1% on GOT-10Kdataset

an area-under-curve of 72.3%

and a precision of 82.4% on LaSOTdataset. The results demonstrate superior tracking performance

particularly in handling occlusion

deformation

and interference from similar objects.

关键词

Keywords

references

ZHAO G C, MENG F Y, YANG C Z, et al. A review of object tracking based on deep learning[J]. Neurocomputing, 2025, 651: 130988.

丁奇帅,雷帮军,吴正平. 基于孪生网络的轻量型无人机单目标跟踪算法[J]. 航空学报, 2025, 46(4): 330925.

DING Q S, LEI B J, WU Z P. A lightweight single object tracking algorithm for UAVs based on Siamese network[J]. Acta Aeronautica et Astronautica Sinica, 2025, 46(4): 330925.(in Chinese)

才华,周鸿策,付强,等. 基于多层特征嵌入的单目标跟踪算法[J]. 兵工学报, 2025, 46(3): 240062-.

CAI H, ZHOU H C, FU Q, et al. Single object tracking algorithm based on multilayer feature embedding[J]. Acta Armamentarii, 2025, 46(3): 240062.(in Chinese)

杨绪祺,谭启凡,苏航,等. 面向无人机视觉制导的自适应目标跟踪方法[J]. 兵工学报, 2025, 46(2): 240284.

YANG X Q, TAN Q F, SU H, et al. Guidance-Tracker:an adaptive UAV siamese tracker for visual guidance[J]. Acta Armamentarii, 2025, 46(2): 240284.(in Chinese)

BOLME D S, BEVERIDGE J R, DRAPER B A, et al. Visual object tracking using adaptive correlation filters[C]// Proceedings of the 2010 IEEE Computer Society Conference on Computer Vision and Pattern Recognition. San Francisco, CA, US: IEEE, 2010: 2544-2550.

HENRIQUES J F, CASEIRO R, MARTINS P, BATISTA J. High-speed tracking with kernelized correlation filters[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2015, 37(3): 583-596.

BERTINETTO L, VALMADRE J, HENRIQUES J F, et al. Fully-convolutional Siamese networks for object tracking[C]//Proceedings of 2016 European Conference on Computer Vision Part II. Amsterdam, The Netherlands: Springer, 2016: 850-865.

LI B, WU W, WANG Q, et al. SiamRPN++: evolution of siamese visual tracking with very deep networks[C]// Proceedings of 2019 IEEE Conference on Computer Vision and Pattern Recognition. Long Beach, CA, US: IEEE, 2019: 4282-4291.

ZHANG Z P, PENG H W, FU J L, et al. Ocean: object-aware anchor-free tracking[C]//Proceedings of European Conference on Computer Vision. Glasgow, UK: Springer, 2020: 771-787.

GUO D Y, WANG J, CUI Y, et al. SiamCAR: siamese fully convolutional classification and regression for visual tracking[C]//Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. Piscataway, NJ, US: IEEE, 2020: 6268-6276.

CHEN X, YAN B, ZHU J W, et al. Transformer tracking[C]//Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. Nashville, TN, US: IEEE, 2021: 8126-8135.

YAN B, PENG H W, FU J L, et al. Learning spatio-temporal transformer for visual tracking[C]//Proceedings of the IEEE/CVF International Conference on Computer Vision. Montreal, QC, Canada: IEEE, 2021: 10428-10437.

DOSOVITSKIY A, BEYER L, KOLESNIKOV A, et al. An image is worth 16x16 words: Transformers for image recognition at scale[C]//Proceedings of International Conference on Learning Representations. Vienna, Austria: OpenReview.net, 2021.

LIU Z, LIN Y T, CAO Y, et al. Swin transformer: Hierarchical vision transformer using shifted windows[C]//Proceedings of the IEEE/CVF International Conference on Computer Vision. Montreal, Canada: IEEE, 2021: 10012-10022.

CUI Y T, JIANG C, WU L M, 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.

YE B T, CHANG H, MA B P, et al. Joint feature learning and relation modeling for tracking: a one-stream framework[C] // Computer Vision – ECCV 2022. Cham, Switzerland: Springer, 2022: 341-357.

CAI W R, LIU Q J, WANG Y H. Hiptrack: visual tracking with historical prompts[C]//Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. Seattle, WA, US: IEEE, 2024: 19258-19267.

GAO S Y, ZHOU C L, ZHANG J. Generalized relation modeling for transformer tracking[C]//Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. Vancouver, BC, Canada: IEEE, 2023: 18686-18695.

WU Q Q, YANG T Y, LIU Z Q, et al. DropMAE: masked autoencoders with spatial-attention dropout for tracking tasks[C]//Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. Vancouver, BC, Canada: IEEE, 2023: 14561-14571.

HE K, ZHANG C, XIE S, et al. Target-aware tracking with long-term context attention[C]//Proceedings of the AAAI Conference on Artificial Intelligence. Washington, DC, US：AAAI, 2023, 37(1): 773-780.

徐晗,董仕豪,张家伟,等.融合上下文感知注意力的Transformer目标跟踪方法[J].中国图象图形学报, 2025,30(1):212-224.

XU H, DONG S H, ZHANG J W, et al. Context-aware attention fused Transformer tracking[J]. Journal of Image and Graphics, 2025, 30(1):212-224.(in Chinese)

CHEN Z, LIU L J, ZHEN Y. Transformer tracking with auxiliary search token[J]. Expert Systems with Applications, 2025, 274: 126910.

ZHANG H L, FU W Q, QI R, et al. Target-background interaction modeling transformer for object tracking[J]. Knowledge-Based Systems, 2025, 315: 113230.

CAI Y, LIU J, TANG J, et al. Robust object modeling for visual tracking[C]//Proceedings of the IEEE/CVF International Conference on Computer Vision. Paris, France: IEEE, 2023: 9589-9600.

CHEN X, PENG H W, WANG D, et al. Seqtrack: sequence to sequence learning for visual object tracking[C]//Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. Vancouver, BC, Canada: IEEE, 2023: 14572-14581.

WEI X, BAI Y F, ZHENG Y C, et al. Autoregressive visual tracking[C]//Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. Vancouver, BC, Canada: IEEE, 2023: 9697-9706.

SONG Z K, LUO R, YU J Q, et al. Compact transformer tracker with correlative masked modeling[C]//Proceedings of the AAAI conference on artificial intelligence. Washington, DC, US:AAAI, 2023, 37(2): 2321-2329.

QIN H L, XU T F, LI T H, et al. MUST: the first dataset and unified framework for multispectral UAV single object tracking[C]//Proceedings of the Computer Vision and Pattern Recognition Conference. Nashville, TN, US: IEEE, 2025: 16882-16891.

浏览量

下载量

CNKI被引量

文章被引用时，请邮件提醒。

提交

工具集

关联资源

基于多层特征嵌入的单目标跟踪算法

面向无人机视觉制导的自适应目标跟踪方法

基于空间偏振混叠特性的红外偏振视频目标跟踪

基于红外相机和毫米波雷达融合的烟雾遮挡无人驾驶车辆目标检测与跟踪

事件域与RGB域跨模态融合的仿生目标跟踪算法