自适应上下文感知相关滤波目标跟踪

何冉; 陈自力; 刘建军; 高喜俊

doi:doi:10.3969/j.issn.1671-637x.2019.05.011

电光与控制, 2019, 26 (5): 59, 网络出版: 2021-02-01

自适应上下文感知相关滤波目标跟踪

Correlation Filter Target Tracking with Adaptive Context Sensing

论文大纲

何冉陈自力刘建军高喜俊

作者单位

陆军工程大学石家庄校区，石家庄050003

摘要

针对传统相关滤波目标跟踪算法在目标快速运动、遮挡、复杂背景等情况下跟踪精度低的问题, 提出了一种自适应上下文感知的相关滤波目标跟踪算法。在相关滤波算法框架的基础上, 重点针对循环移位带来的边界效应与固定学习率进行改进: 首先,在分类器训练阶段提出一种基于响应图极值的自适应采样策略加入上下文信息;然后, 采用了一种分段学习率调整策略使算法更好地适应目标变化;最后,在标准数据集上验证了算法的性能。实验结果表明,提出的算法提高了DCF与SAMF算法的跟踪精度, 不仅在目标快速运动、遮挡、复杂背景等情况下鲁棒性较好, 而且还能作为一种框架集成到大部分相关滤波类算法中。

Abstract

The traditional correlation filter tracking algorithms have low precision when the target is in fast motion, occluded, or under complex background. To solve the problem, a correlation filter target tracking algorithm with adaptive context sensing was proposed. Based on the framework of the correlation filter algorithm, the boundary effects brought by the cyclic shifts and the fixed learning rate were mainly improved. Firstly, an adaptive sampling strategy based on response map was used to sample the context information in the classifier training stage. Then, a segmented learning rate adjustment strategy was used to make the algorithm more adaptive to the changes of target. Finally, the performance of the proposed algorithm was verified on the standard data set. The experimental results showed that the proposed algorithm can improve the tracking precision of DCF and SAMF algorithms. It not only has better robustness in the case of fast motion, occlusion, and complex background, but also can be used as a framework integrated into most of the correlation filter trackers.

参考文献

[1] 张微, 康宝生.相关滤波目标跟踪进展综述［J］.中国图象图形学报, 2017, 22(8): 1017-1033.

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

[3] LI Y, ZHU J K.A scale adaptive kernel correlation filter tracker with feature integration［C］//Proceedings of European Conference on Computer Vision, 2014:254-265.

[4] DANELLJAN M, H?GER G, KHAN F, et al.Learning spatially regularized correlation filters for visual tracking［C］//International Conference on Computer Vision(ICCV), IEEE, 2015:4310-4318.

[5] GALOOGAHI H K, SIM T, LUCEY S.Correlation filters with limited boundaries［C］//Conference on Computer Vision and Pattern Recognition(CVPR), 2015:4630-4638.

[6] GALOOGAHI H K, FAGG A, LUCEY S.Learning background-aware correlation filters for visual tracking［C］//International Conference on Computer Vision(ICCV), IEEE, 2017:1144-1152.

[7] MUELLER M, SMITH N, GHANEM B.Context-aware correlation filter tracking［C］//Conference on Computer Vision and Pattern Recognition(CVPR), 2017:1387-1395.

[8] 熊昌镇, 赵璐璐, 郭芬红.自适应特征融合的核相关滤波跟踪算法［J］.计算机辅助设计与图形学学报, 2017, 29(6): 1068-1074.

[9] WU Y, LIM J, YANG M H.Online object tracking:a benchmark［C］//Conference on Computer Vision and Pattern Recognition(CVPR), 2013:2411-2418.

[10] 王艳川, 黄海, 李邵梅, 等.基于在线检测和尺度自适应的相关滤波目标跟踪［J］.光学学报, 2018,38(2):0215002.

[11] 忽晓伟, 陈娟.融合颜色特征的核相关滤波器目标跟踪［J］.电光与控制, 2017,24(6):43-46.

[12] 陈倩茹, 刘日升, 樊鑫, 等.多相关滤波自适应融合的鲁棒目标跟踪［J］.中国图象图形学报, 2018,23(2):269-276.

何冉, 陈自力, 刘建军, 高喜俊. 自适应上下文感知相关滤波目标跟踪[J]. 电光与控制, 2019, 26(5): 59. HE Ran, CHEN Zi-li, LIU Jian-jun, GAO Xi-jun. Correlation Filter Target Tracking with Adaptive Context Sensing[J]. Electronics Optics & Control, 2019, 26(5): 59.

自适应上下文感知相关滤波目标跟踪

关于本站 Cookie 的使用提示

全站搜索

自适应上下文感知相关滤波目标跟踪

相关论文

相关资讯

关于本站 Cookie 的使用提示

全站搜索