作者：张文明¹, 侯建平¹, 朱向东², 李海滨¹, 高雅昆¹, 李雅倩¹

作者单位：1. 燕山大学 工业计算机控制工程河北省重点实验室, 河北 秦皇岛 066004;
2. 秦皇岛港股份有限公司, 河北 秦皇岛 066002

关键词：目标跟踪;时空上下文;尺度优化;滤波函数

摘要：

视觉目标跟踪在智能监控和人机交互等领域有着广泛的应用。该文针对时空上下文（STC）跟踪算法尺度适应性不强的问题进行研究，提出一种尺度优化的时空上下文目标跟踪算法。首先，提出一种新型加权滤波函数，滤除图像的高频信息，提升算法的精度；其次，定义两种判别标准，实现时空上下文模型的自适应更新；最后，通过相关性原理训练尺度滤波器，估计出目标的尺度大小，提高算法的尺度适应性。实验表明：提出的跟踪算法能有效改善STC跟踪算法的尺度更新缺陷问题，提高STC跟踪算法的跟踪精度，与近年来出现的基本跟踪算法相比，该算法有着良好的跟踪效果。该算法在AMD-A6处理器、2.7 GHz主频、4 GB内存的计算机硬件平台下，实现68 f/s的实时跟踪速度。

Scale-optimized spatio-temporal context target tracking
ZHANG Wenming¹, HOU Jianping¹, ZHU Xiangdong², LI Haibin¹, GAO Yakun¹, LI Yaqian¹
1. Key Lab of Industrial Computer Control Engineering of Hebei Province, Yanshan University, Qinhuangdao 066004, China;
2. Qinhuangdao Port Co., Ltd., Qinhuangdao 066002, China
Abstract: Visual object tracking has a wide range of applications in the field of intelligent monitoring and human-computer interaction. In this paper, a scale-optimized spatio-temporal context target tracking algorithm is proposed to solve the problem of scale adaptability of spatio-temporal context (STC) tracking algorithm. Firstly, a new weighted filter function is proposed to filter out the high frequency information of the image. Secondly, two kinds of discriminant criterias are defined and the updating of the spatio-temporal context model is optimized. Finally, the scale filter is trained by the correlation principle and the scale pyramid to estimate the size of the target scale.The experimental results demonstrate that the proposed algorithm can solve the problem of scale update defects of STC tracking algorithm and improve the tracking accuracy of STC tracking algorithm, and this algorithm has better tracking effect compared with some basic tracking algorithms which have appeared in recent years. The algorithm achieves real-time tracking speed of 68 f/s on AMD-A6 processor, 2.7 GHz frequency, 4 GB memory computer hardware platform.
Keywords: object tracking;spatio-temporal context;scale-optimized;filter function
2019, 45(4):1-8  收稿日期: 2018-03-10;收到修改稿日期: 2018-04-16
基金项目: 河北省自然科学基金（F2015203212）
作者简介: 张文明(1979-),男,吉林长春市人,副教授,博士,主要从事计算机视觉方面的研究
参考文献
[1] KALAL Z, MIKOLAJCZYK K, MATAS J. Tracking-Learning-Detection[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence (PAMI), 2012, 34(7):1409-1422
[2] ZHANG K, ZHANG L, YANG M H. Fast Compressive Tracking[J]. IEEE Transactions on Pattern Analysis & Machine Intelligence, 2014, 36(10):2002-2015
[3] LI Y, ZHU J K. A scale adaptive kernel correlation filter tracker with feature integration[C]//European Conference on Computer Vision, Workshop VOT2014(ECCVW), 2014:254-265.
[4] 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 (PAMI), 2015:583-596
[5] BOLME D S, BEVERIDGE J R, DRAPER B A, et al. Visual object tracking using adaptive correlation filters[C]//Computer Vision and Pattern Recognition. IEEE, 2010:2544-2550.
[6] RUI C, MARTINS P, BATISTA J. Exploiting the circulant structure of tracking-by-detection with kernels[C]//European Conference on Computer Vision. Springer-Verlag, 2012:702-715.
[7] BIBI A, GHANEM B. Multi-template Scale-Adaptive Kernelized Correlation Filters[C]//IEEE International Conference on Computer Vision Workshop. IEEE Computer Society, 2015:613-620.
[8] BIBI A, MUELLER M, GHANEM B. Target Response Adaptation for Correlation Filter Tracking[C]//European Conference on Computer Vision. Springer International Publishing, 2016:419-433.
[9] ZHANG K, ZHANG L, LIU Q, et al. Fast Visual Tracking via Dense Spatio-temporal Context Learning[C]//European Conference on Computer Vision. Springer, Cham, 2014:127-141.
[10] 刘威, 赵文杰, 李成. 时空上下文学习长时目标跟踪[J]. 光学学报, 2016, 36(1):179-186
[11] DAI M, LIN P, WU L, et al. Orderless and Blurred Visual Tracking via Spatio-temporal Context[C]//International Conference on Multimedia Modeling. Springer, Cham, 2015:25-36.
[12] 徐建强, 陆耀. 一种基于加权时空上下文的鲁棒视觉跟踪算法[J]. 自动化学报, 2015, 41(11):1901-1912
[13] JIANG H, LI J, WANG D, et al. Multi-feature tracking via adaptive weights[J]. Neurocomputing, 2016, 207(9):189-201
[14] DANELLJAN M, HÄGER G, KHAN F S. Accurate scale estimation for robust visual tracking[C]//British Machine Vision Conference, 2014(65):1-11.
[15] 刘曙, 狄红卫, 姚曼虹. 基于自适应尺度的TLD目标跟踪算法[J]. 光学技术, 2017, 43(6):542-546
[16] MONTERO A S, LANG J, LAGANIÈRE R. Scalable Kernel Correlation Filter with Sparse Feature Integration[C]//IEEE International Conference on Computer Vision Workshop. IEEE, 2016:587-594.
[17] LIU T, WANG G, YANG Q. Real-time part-based visual tracking via adaptive correlation filters[C]//IEEE Conference on Computer Vision and Pattern Recognition. IEEE Computer Society, 2015:4902-4912.
[18] WU Y, LIM J, YANG M H. Online Object Tracking:A Benchmark[C]//IEEE Conference on Computer Vision and Pattern Recognition. IEEE Computer Society, 2013:2411-2418.