高效的多特征自适应相关滤波跟踪器

doi:10.16182/j.issn1004731x.joss.21-0274

系统仿真学报 ›› 2022, Vol. 34 ›› Issue (8): 1864-1873.doi: 10.16182/j.issn1004731x.joss.21-0274

• 物理效应/模拟器仿真技术 • 上一篇下一篇

高效的多特征自适应相关滤波跟踪器

张思贤¹^,²(), 杨艺¹^,²(), 张猛¹^,², 米鹏博¹^,²

^1.西安交通大学机械结构强度与振动国家重点实验室，陕西西安 710049
^2.西安交通大学航天航空学院，陕西西安 710049

收稿日期:2021-03-31 修回日期:2021-08-12 出版日期:2022-08-30 发布日期:2022-08-15
通讯作者: 杨艺 E-mail:touchmeteor@stu.xjtu.edu.cn;jiafeiyy@mail.xjtu.edu.cn
作者简介:张思贤(1999-)，男，博士生，研究方向为多元信息融合等。E-mail：touchmeteor@stu.xjtu.edu.cn
基金资助:
国家自然科学基金(61671370);中国博士后科学基金(2016M592790);中央高校基本科研业务费专项资金(xjj2016066);陕西省博士后基金(2016BSHEDZZ46)

An Efficient Tracker via Multi-feature Adaptive Correlation Filter

Sixian Zhang¹^,²(), Yi Yang¹^,²(), Meng Zhang¹^,², Pengbo Mi¹^,²

^1.State Key Laboratory for Strength and Vibration of Mechanical Structure, Xi'an Jiaotong University, Xi'an 710049, China
^2.School of Aerospace, Xi'an Jiaotong University, Xi'an 710049, China

Received:2021-03-31 Revised:2021-08-12 Online:2022-08-30 Published:2022-08-15
Contact: Yi Yang E-mail:touchmeteor@stu.xjtu.edu.cn;jiafeiyy@mail.xjtu.edu.cn

摘要/Abstract

摘要：

针对基于手工特征的相关滤波跟踪器在快速变形、背景杂乱等挑战性跟踪场景效果不佳的问题，在Staple跟踪器的基础上提出了一种新型的相关滤波跟踪器。结合方向梯度直方图特征与颜色命名特征构建了目标外观模型，增强其对快速变形以及背景杂乱等场景的鲁棒性；设计了自适应评分函数对2种特征进行融合，得到更具有鉴别性的特征；针对不同的特征分别提出了在线更新策略以减小训练过拟合与模型漂移。实验结果表明：该跟踪器在跟踪的准确性与实时性上均有着优良表现。

Abstract:

Aiming at the low tracking effect of the correlation filters tracker based on manual features in challenging scenes of rapid deformation and background clutter, a new correlation filter tracker based on Staple tracker is proposed. An appearance model based on HOG features and color-naming features is built to enhance the robustness to the challenging scenes of rapid deformation and background clutter. A self-adjust evaluation function is designed to merge the two kinds of feature information and a more discriminative feature is obtained. The novel online update strategies to reduce the training over-fitting and model drift for different features are proposed. The tracker shows excellent performance in accuracy and real-time capability on OTB2015 benchmark.

Key words: object tracking, correlation filter, self-adjust evaluation function, color-naming feature, online update

中图分类号:

TP391.4

张思贤, 杨艺, 张猛, 米鹏博. 高效的多特征自适应相关滤波跟踪器[J]. 系统仿真学报, 2022, 34(8): 1864-1873.

Sixian Zhang, Yi Yang, Meng Zhang, Pengbo Mi. An Efficient Tracker via Multi-feature Adaptive Correlation Filter[J]. Journal of System Simulation, 2022, 34(8): 1864-1873.

图/表 7

图1

表1

实验参数

主要参数	量值	主要参数	量值
学习率 $η$	0.01	融合参数 $α 2$	0.6
颜色特征	Color Naming	检测区域	150×150
融合参数 $α 1$	0.3	Hog单元尺寸	4×4

表1

图2

表2

图3

图4

表3

参考文献 26

1	Bolme D S, Ross Beveridge J, Draper Bruce A, et al. Visual Object Tracking Using Adaptive Correlation Filters[C]//2010 IEEE Computer Society Conference on Computer Vision and Pattern Recognition. San Francisco, CA: IEEE, 2010: 2544-2550.
2	Zhang T, Liu S, Xu C, et al. Correlation Particle Filter for Visual Tracking[J]. IEEE Transactions on Image Processing(S1057-7149), 2018, 27(6): 2676-2687.
3	Zhang T, Xu C, Yang M. Learning Multi-Task Correlation Particle Filters for Visual Tracking[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence(S0162-8828), 2019, 41(2): 365-378.
4	Li S, Zhao S, Cheng B, et al. Robust Visual Tracking Via Hierarchical Particle Filter and Ensemble Deep Features[J]. IEEE Transactions on Circuits and Systems for Video Technology(S1051-8215), 2020, 30(1): 179-191.
5	Henriques J, Caseiro R, Martins P, et al. Exploiting the Circulant Structure of Tracking-by-Detection with Kernels[C]//12th European Conference on Computer Vision. Berlin, Heidelberg: Springer, 2012: 702-715.
6	Danelljan M, Häger G, Khan F S, et al. Learning Spatially Regularized Correlation Filters for Visual Tracking[C]//2015 IEEE International Conference on Computer Vision. Santiago: IEEE, 2015: 4310-4318.
7	Song Z, Sun J, Duan B. Collaborative Correlation Filter Tracking with Online Re-detection[C]//2019 IEEE 3rd Information Technology, Networking, Electronic and Automation Control Conference (ITNEC). Chengdu: IEEE, 2019: 1303-1313.
8	Zhang S, Lu W, Xing W, et al. Learning Scale-Adaptive Tight Correlation Filter for Object Tracking[J]. IEEE Transactions on Cybernetics(S2168-2267), 2020, 50(1): 270-283.
9	Henriques J F, Caseiro R, Martins P, et al. High-Speed Tracking with Kernelized Correlation Filters[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence (S0162-8828), 2015, 37(3): 583-596.
10	Lukežič A, Čehovin Zajc L, Kristan M. Deformable Parts Correlation Filters for Robust Visual Tracking[J]. IEEE Transactions on Cybernetics(S2168-2267), 2018, 48(6): 1849-1861.
11	Martin Danelljan, Gustav Häger, Khan Shahbaz Fahad, et al. Discriminative Scale Space Tracking[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence (S0162-8828), 2017, 39(8): 1561-1575.
12	Danelljan M, Khan F S, Felsberg M, et al. Adaptive Color Attributes for Real-Time Visual Tracking[C]//2014 IEEE Conference on Computer Vision and Pattern Recogni-tion. Columbus, OH: IEEE, 2014: 1090-1097.
13	Gundogdu E, Alatan A A. Good Features to Correlate for Visual Tracking[J]. IEEE Transactions on Image Processing(S1057-7149), 2018, 27(5): 2526-2540.
14	Bertinetto L, Valmadre J, Golodetz S, et al. Staple: Complementary Learners for Real-Time Tracking[C]//2016 IEEE Conference on Computer Vision and Pattern Recognition. Las Vegas, NV: IEEE, 2016: 1401-1409.
15	Valmadre J, Bertinetto L, Henriques J, et al. End-to-End Representation Learning for Correlation Filter Based Tracking[C]//2017 IEEE Conference on Computer Vision and Pattern Recognition. Honolulu, HI: IEEE, 2017: 5000-5008.
16	Feng J, Pu S, Zhao K, et al. Enhanced Initialization with Multi-Stage Learning for Robust Visual Tracking[C]//2019 IEEE Visual Communications and Image Processing. Sydney, NSW: IEEE, 2019: 1-4.
17	Bhat G, Danelljan M, Van Gool L, et al. Learning Discriminative Model Prediction for Tracking[C]//2019 IEEE/CVF International Conference on Computer Vision (ICCV). Seoul, Korea(South): IEEE, 2019: 6181-6190.
18	Danelljan M, Bhat G, Khan F S, et al. ATOM: Accurate Tracking by Overlap Maximization[C]//2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition. Long Beach, CA: IEEE, 2019: 4655-4664.
19	Guo D, Wang J, Cui Y, et al. SiamCAR: Siamese Fully Convolutional Classification and Regression for Visual Tracking[C]//2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition. Seattle, WA: IEEE, 2020: 6268-6276.
20	Voigtlaender P, Luiten J, Torr P H S, et al. SiamR-CNN: Visual Tracking by Re-Detection[C]// 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition. Seattle, WA: IEEE, 2020: 6577-6587.
21	Li B, Wu W, Wang Q, et al. SiamRPN++: Evolution of Siamese Visual Tracking with Very Deep Networks[C]//2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition. Long Beach, CA: IEEE, 2019: 4277-4286.
22	Hu J, Shen L, Albanie S, et al. Squeeze-and-Excitation Networks[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence(S0162-8828), 2020, 42(8): 2011-2023.
23	Danelljan M, Van Gool L, Timofte R. Probabilistic Regression for Visual Tracking[C]// 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition. Seattle, WA: IEEE, 2020: 7181-7190.
24	Martin Danelljan, Goutam Bhat, Khan Fahad Shahbaz, et al. ECO: Efﬁcient Convolution Operators for Tracking[C]//2017 IEEE Conference on Computer Vision and Pattern Recognition. Honolulu, HI: IEEE, 2017: 6931-6939.
25	Wang M, Liu Y, Huang Z. Large Margin Object Tracking with Circulant Feature Maps [C]//2017 IEEE Conference on Computer Vision and Pattern Recognition. Honolulu, HI: IEEE, 2017: 4021-4029.
26	Wu Y, Lim J, Yang M H. Object Tracking Benchmark[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence (S0162-8828), 2015, 37(9): 1834-1848.

跟踪器	帧数	跟踪器	帧数
CSK	275.0	ECO_HC	37.5
KCF	120.5	DSST	33.7
Staple	55.1	CN	13.5
STACL	50.3	SRDCF	9.8

challenges	STACL	Staple
变形	0.640	0.617
背景杂波	0.682	0.623
光照变化	0.631	0.596
面外旋转	0.594	0.561
面内旋转	0.589	0.549
尺度变化	0.613	0.552
遮挡	0.593	0.621
抖动	0.631	0.625
快速移动	0.657	0.651
低分辨率	0.570	0.569

[1]	周维, 刘宇翔, 廖广平, 马鑫. 结合交并比损失的孪生网络目标跟踪算法研究[J]. 系统仿真学报, 2022, 34(9): 1956-1967.
[2]	李琪, 墨瀚林, 王向东, 李华. 短道速滑场景下的多目标跟踪与运动仿真[J]. 系统仿真学报, 2021, 33(5): 1039-1050.
[3]	柳有权, 裴雪, 李婉, 刘正雄. 基于邻近目标置信度评估的视觉目标跟踪与定位[J]. 系统仿真学报, 2020, 32(7): 1294-1300.
[4]	张良, 郝凯锋. 基于RSSI变化序列的MAC地址与目标匹配方法[J]. 系统仿真学报, 2020, 32(1): 113-121.
[5]	孙永侃, 张萍萍. 基于横向地球坐标系的极区超视距目标跟踪[J]. 系统仿真学报, 2018, 30(9): 3502-3513.
[6]	李菊, 曹明伟, 余烨, 夏瑜, 周立凡. 抗遮挡自适应的粒子滤波算法[J]. 系统仿真学报, 2018, 30(9): 3552-3557.
[7]	姜明新, 潘志庚, 王兰芳, 胡铸鑫. 基于深度去噪自编码器的RGB-D视频目标跟踪[J]. 系统仿真学报, 2018, 30(11): 4276-4283.
[8]	张志利, 王超, 龙勇, 王韶迪. 基于核相关滤波器无标识AR的目标跟踪仿真[J]. 系统仿真学报, 2018, 30(11): 4462-4468.
[9]	牛为华, 梁贵书, 赵鹏. 基于稀疏表示和M-ELM的断路器故障诊断模型[J]. 系统仿真学报, 2017, 29(11): 2828-2839.
[10]	陈金广, 江梦茜, 马丽丽. 存在目标交叉情形的扩展目标跟踪算法[J]. 系统仿真学报, 2017, 29(1): 7-13.
[11]	周彦, 胡岚, 王冬丽. 基于IMM-UKF-CS的混合环境目标跟踪[J]. 系统仿真学报, 2017, 29(1): 43-48.
[12]	马丽丽, 秦晓姗, 陈金广. 一种状态与杂波相关条件下的GM-CPHD算法[J]. 系统仿真学报, 2016, 28(7): 1637-1643.
[13]	夏候凯顺, 陈善星, 邬依林. 基于双目云台相机的目标跟踪系统建模与仿真[J]. 系统仿真学报, 2015, 27(2): 362-368.
[14]	吴静静, 尤丽华, 王瑶, 宋淑娟. 基于自适应新生目标强度的概率假设密度滤波[J]. 系统仿真学报, 2015, 27(11): 2741-2747.

高效的多特征自适应相关滤波跟踪器

An Efficient Tracker via Multi-feature Adaptive Correlation Filter

RichHTML

PDF (PC)

可视化

摘要/Abstract

引用本文

使用本文

图/表 7

参考文献 26

相关文章 14

编辑推荐

Metrics

本文评价