搜索

x

留言板

尊敬的读者、作者、审稿人, 关于本刊的投稿、审稿、编辑和出版的任何问题, 您可以本页添加留言。我们将尽快给您答复。谢谢您的支持!

姓名
邮箱
手机号码
标题
留言内容
验证码

基于残差矩阵估计的稀疏表示目标跟踪算法

陈典兵 朱明 高文 王慧利 杨航

引用本文:
Citation:

基于残差矩阵估计的稀疏表示目标跟踪算法

陈典兵, 朱明, 高文, 王慧利, 杨航

Visual tracking based on the estimation of representation residual matrix

Chen Dian-Bing, Zhu Ming, Gao Wen, Wang Hui-Li, Yang Hang
PDF
导出引用
  • 基于稀疏表示的目标跟踪算法多数利用稀疏系数计算目标位置信息,而忽略了稀疏表示过程中的残差所包含的信息.因此,本文设计了一种基于残差矩阵估计的跟踪模型.该模型在粒子滤波的框架下利用L1范数分别约束稀疏表示系数与残差矩阵,并且利用L2范数建立残差矩阵与观测模型之间的联系.本文给出了相应求解模型的表示系数与残差矩阵的迭代算法,并利用残差矩阵更新模板字典.相比应用稀疏系数的跟踪算法,本文算法考虑了残差矩阵对跟踪结果的影响,使得算法对于候选目标的评估更加精确,同时在模板更新部分引入残差矩阵,使得字典能够更好地描述目标的变化.实验数据表明,本文算法优于现今主流算法.
    In recent years,sparse representation theory has acquired considerable progress and has extensively been used in visual tracking.Most trackers used the sparse coefficients to merely calculate the position of the target according to the reconstruction error relative to sparse coefficients,and often neglected the information contained by representation residual matrix in representing step.Consequently,we present a novel sparse representation based tracker which takes representation residual matrix into consideration.First of all,at initialization of a new frame,we reconstruct the frame by singular value decomposition (SVD) to eliminate noise and useless information,which contributes a friendly frame for the following representation step.To obtain the compact representation of the target,we build L2-norm regularization according to the distance between the candidates wrapped in particle framework and the reconstruction calculated by dictionary templates and residual matrix.Additionally,we use the L1-norm constraint to restrict the sparse coefficients and the residual matrix of each candidate.Secondly,as the built optimization problem does not have closed-form solution,we design a method to compute the coefficients and the residual matrix iteratively.During each iteration,the coefficients are obtained by solving classical least absolute shrinkage and selectionator operator (LASSO) model,and the residual matrix is achieved by shrinkage operation.After solving the optimization problem,we compute the score of each candidate for evaluating the truth target with considering coefficients and residual matrix.The score is formulated as weighted reconstruction error which consists of dictionary templates,candidates,coefficients and residual matrix. The weight is the exponential value of absolute value of elements in residual matrix.Finally,for capturing the varying appearance of target in series,we update the dictionary template with assembled template,which is composed of residual matrix,selected candidate and dictionary template.In this paper,the template to be replaced is determined according to the score which is inversely proportional to the distance between the selected candidate and each dictionary template. Then we update the dictionary frame by frame during tracking process.Contributions of this work are threefold:1) the representation model captures holistic and local features of target and makes the tracker robust to varying illumination, shape transformation,and background clutter,profiting from preprocessing of SVD reconstruction,the model exhibits a more compact representation of target without disturbance of noisy variance;2) we employ a weight matrix to adjust reconstruction error in candidate evaluation step,as described above,the weight matrix strengthens the effect of error in residual matrix for evaluating candidates from which target is selected,it is noted that weights are all greater than one,which leads to reconstruction error expanding according to the error value of residual matrix,and keeps pixels where there is small error value believable for evaluation;and 3) we adopt an assembled template to update dictionary template and reconstruction of coefficients of selected candidate,which alleviates dictionary degradation and tracking drift problems and provides an accurate description of new appearance of target.In order to illustrate the performance of the proposed tracker,we enforce the algorithm on several challenging sequences and compare the proposed algorithm with five state-of-art methods,whose codes are all supplied by the authors.For complete illustration,both qualitative evaluation and quantitative evaluation are presented in experiment section.Through the experimental results,we could conclude that the proposed algorithm has a more favorable and robust performance than other state-of-art algorithms when dealing with kinds of situations during tracking.
      通信作者: 朱明, zhu_mingca@163.com
    • 基金项目: 国家自然科学基金(批准号:61401425)和吉林省科技发展计划青年科研基金(批准号:20150520057JH)资助的课题.
      Corresponding author: Zhu Ming, zhu_mingca@163.com
    • Funds: Project supported by the National Natural Science Foundation of China (Grant No. 61401425), and the Science and Technology Development Plan for Youth Foundation of Jilin Province, China (Grant No. 20150520057JH).
    [1]

    Gao W, Tang Y, Zhu M 2015 Acta Phys. Sin. 64 014205 (in Chinese) [高文, 汤洋, 朱明2015物理学报64 014205]

    [2]

    Xu Y, Zhang B, Zhong Z F 2015 Pattern Recogn. Lett. 68 9

    [3]

    Fan Q, Qi C 2016 Neurocomputing 175 81

    [4]

    Kim M, Han D K, Ko H 2016 Information Fusion 27 198

    [5]

    Mei X, Ling H B 2009 Proceedings of IEEE International Conference on Computer Vision Kyoto, Japan, September 27-October 4, 2009 p1436

    [6]

    Liu B Y, Huang J Z, Yang L, Kulikowsk C 2011 Proceedings of IEEE Computer Vision, Pattern Recognition Colorado, Springs, June 21-252011 p1313

    [7]

    Jia X, Lu H C, Yang M H 2012 Proceedings of IEEE Computer Vision, Pattern Recognition Providence, Rhode Island, June 16-21, 2012 p1822

    [8]

    Liu H P, Sun F C 2010 Proceedings of International Conference on Pattern Recognition Istanbul, Turkey, August 23-26, 2010 p1702

    [9]

    Wang B X, Zhao B J, Tang L B, Wang S G, Wu J H 2014 Acta Phys. Sin. 63 234201 (in Chinese) [王保宪, 赵保军, 唐林波, 王水根, 吴京辉2014物理学报63 234201]

    [10]

    Liu B Y, Yang L, Huang J Z, Meer P, Gong L G, Kulikowski C 2010 Proceedings of the 11th European Conference on Computer Vision Crete, Greece, September 5-11, 2010 p624

    [11]

    Wang Q, Chen F, Xu W L, Yang M H 2012 Proceedings of I EE E Workshop on Applications of Computer Vision Breckenridge, C O, January 9-11, 2012 p425

    [12]

    Bao C L, Wu Y, Ling H B, Ji H 2012 Proceedings of IEEE Computer Vision, Pattern Recognition Providence, Rhode Island, June 16-21, 2012 p1830

    [13]

    Pérez P, Hue C, Vermaak J, Gangnet M 2002 European Conference on Computer Vision Copenhagen, Denmark, May 28-31, 2002 p661

    [14]

    Zhang T Z, Ghanem B, Liu S, Ahuja N 2013 Int. J. Comput. Vision 101 367

    [15]

    Zhuang B H, Lu H C, Xiao Z Y, Wang D 2014 IEEE Trans. Image Proces. 23 1872

    [16]

    Zhong W, Lu H C, Yang M H 2012 Proceedings of IEEE Computer Vision, Pattern Recognition Providence, Rhode Island, June 16-21, 2012 p1838

    [17]

    Donoho D L 2006 IEEE Trans. Inform. Theory 52 1289

    [18]

    Donoho D L, T SA IG Y 2006 Signal Proces. 86 533

    [19]

    Rao S R, Tron R, Vidal R, Ma Y 2009 IEEE Trans. PAMI. 32 1832

    [20]

    Wang D, Lu H C 2012 IEEE Signal Proces. Lett. 19 711

    [21]

    Yan H, Yang J 2016 Neurocomputing 173 1936

    [22]

    Efron B, Hastie T, Johnstone I, Tibshirani R 2004 Ann. Statist. 32 407

    [23]

    Hale E T, Yin W, Zhang Y 2008 SIAM J. Opt. 19 1107

    [24]

    Wu Y, Lim J, Yang M H 2013 Proceedings of IEEE Computer Vision, Pattern Recognition Portland, Oregon, June 23-28, 2013 p2411

    [25]

    Ross D, Lim J, Lin R, Yang M H 2008 Int. J. Comput. Vision 77 125

    [26]

    Kalal Z, Mikolajczyk K, Matas J 2012 IEEE Trans. on PAMI 34 1409

    [27]

    Everingham M, Gool L V, Williams C K I, Winn J M, Zisserman A 2010 Int. J. Comput. Vision 88 303

  • [1]

    Gao W, Tang Y, Zhu M 2015 Acta Phys. Sin. 64 014205 (in Chinese) [高文, 汤洋, 朱明2015物理学报64 014205]

    [2]

    Xu Y, Zhang B, Zhong Z F 2015 Pattern Recogn. Lett. 68 9

    [3]

    Fan Q, Qi C 2016 Neurocomputing 175 81

    [4]

    Kim M, Han D K, Ko H 2016 Information Fusion 27 198

    [5]

    Mei X, Ling H B 2009 Proceedings of IEEE International Conference on Computer Vision Kyoto, Japan, September 27-October 4, 2009 p1436

    [6]

    Liu B Y, Huang J Z, Yang L, Kulikowsk C 2011 Proceedings of IEEE Computer Vision, Pattern Recognition Colorado, Springs, June 21-252011 p1313

    [7]

    Jia X, Lu H C, Yang M H 2012 Proceedings of IEEE Computer Vision, Pattern Recognition Providence, Rhode Island, June 16-21, 2012 p1822

    [8]

    Liu H P, Sun F C 2010 Proceedings of International Conference on Pattern Recognition Istanbul, Turkey, August 23-26, 2010 p1702

    [9]

    Wang B X, Zhao B J, Tang L B, Wang S G, Wu J H 2014 Acta Phys. Sin. 63 234201 (in Chinese) [王保宪, 赵保军, 唐林波, 王水根, 吴京辉2014物理学报63 234201]

    [10]

    Liu B Y, Yang L, Huang J Z, Meer P, Gong L G, Kulikowski C 2010 Proceedings of the 11th European Conference on Computer Vision Crete, Greece, September 5-11, 2010 p624

    [11]

    Wang Q, Chen F, Xu W L, Yang M H 2012 Proceedings of I EE E Workshop on Applications of Computer Vision Breckenridge, C O, January 9-11, 2012 p425

    [12]

    Bao C L, Wu Y, Ling H B, Ji H 2012 Proceedings of IEEE Computer Vision, Pattern Recognition Providence, Rhode Island, June 16-21, 2012 p1830

    [13]

    Pérez P, Hue C, Vermaak J, Gangnet M 2002 European Conference on Computer Vision Copenhagen, Denmark, May 28-31, 2002 p661

    [14]

    Zhang T Z, Ghanem B, Liu S, Ahuja N 2013 Int. J. Comput. Vision 101 367

    [15]

    Zhuang B H, Lu H C, Xiao Z Y, Wang D 2014 IEEE Trans. Image Proces. 23 1872

    [16]

    Zhong W, Lu H C, Yang M H 2012 Proceedings of IEEE Computer Vision, Pattern Recognition Providence, Rhode Island, June 16-21, 2012 p1838

    [17]

    Donoho D L 2006 IEEE Trans. Inform. Theory 52 1289

    [18]

    Donoho D L, T SA IG Y 2006 Signal Proces. 86 533

    [19]

    Rao S R, Tron R, Vidal R, Ma Y 2009 IEEE Trans. PAMI. 32 1832

    [20]

    Wang D, Lu H C 2012 IEEE Signal Proces. Lett. 19 711

    [21]

    Yan H, Yang J 2016 Neurocomputing 173 1936

    [22]

    Efron B, Hastie T, Johnstone I, Tibshirani R 2004 Ann. Statist. 32 407

    [23]

    Hale E T, Yin W, Zhang Y 2008 SIAM J. Opt. 19 1107

    [24]

    Wu Y, Lim J, Yang M H 2013 Proceedings of IEEE Computer Vision, Pattern Recognition Portland, Oregon, June 23-28, 2013 p2411

    [25]

    Ross D, Lim J, Lin R, Yang M H 2008 Int. J. Comput. Vision 77 125

    [26]

    Kalal Z, Mikolajczyk K, Matas J 2012 IEEE Trans. on PAMI 34 1409

    [27]

    Everingham M, Gool L V, Williams C K I, Winn J M, Zisserman A 2010 Int. J. Comput. Vision 88 303

  • [1] 刘杰, 张建勋, 代煜. 基于多引导滤波的图像增强算法. 物理学报, 2018, 67(23): 238701. doi: 10.7498/aps.67.20181425
    [2] 张文杰, 王世元, 冯亚丽, 冯久超. 基于Huber的高阶容积卡尔曼跟踪算法. 物理学报, 2016, 65(8): 088401. doi: 10.7498/aps.65.088401
    [3] 段晓亮, 王一博, 杨慧珠. 基于逆散射理论的地震波速度正则化反演. 物理学报, 2015, 64(7): 078901. doi: 10.7498/aps.64.078901
    [4] 高文, 汤洋, 朱明. 目标跟踪中目标模型更新问题的半监督学习算法研究. 物理学报, 2015, 64(1): 014205. doi: 10.7498/aps.64.014205
    [5] 逯志宇, 王大鸣, 王建辉, 王跃. 基于时频差的正交容积卡尔曼滤波跟踪算法. 物理学报, 2015, 64(15): 150502. doi: 10.7498/aps.64.150502
    [6] 苏勇, 范东明, 游为. 利用GOCE卫星数据确定全球重力场模型. 物理学报, 2014, 63(9): 099101. doi: 10.7498/aps.63.099101
    [7] 刘广东, 张开银. 二维电磁逆散射问题的时域高斯-牛顿反演算法. 物理学报, 2014, 63(3): 034102. doi: 10.7498/aps.63.034102
    [8] 高文, 汤洋, 朱明. 复杂背景下目标检测的级联分类器算法研究. 物理学报, 2014, 63(9): 094204. doi: 10.7498/aps.63.094204
    [9] 王保宪, 赵保军, 唐林波, 王水根, 吴京辉. 基于双向稀疏表示的鲁棒目标跟踪算法. 物理学报, 2014, 63(23): 234201. doi: 10.7498/aps.63.234201
    [10] 王新迎, 韩敏, 王亚楠. 含噪混沌时间序列预测误差分析. 物理学报, 2013, 62(5): 050504. doi: 10.7498/aps.62.050504
    [11] 赵延来, 黄思训, 杜华栋. 基于变分方法的有限区域风场分解与重构I: 理论框架和仿真实验. 物理学报, 2013, 62(3): 039204. doi: 10.7498/aps.62.039204
    [12] 周树波, 袁艳, 苏丽娟. 基于双阈值Huber范数估计的图像正则化超分辨率算法. 物理学报, 2013, 62(20): 200701. doi: 10.7498/aps.62.200701
    [13] 何然, 黄思训, 周晨腾, 姜祝辉. 遗传算法结合正则化方法反演海洋大气波导. 物理学报, 2012, 61(4): 049201. doi: 10.7498/aps.61.049201
    [14] 龙智勇, 石汉青, 黄思训. 利用卫星云图反演云导风的新思路. 物理学报, 2011, 60(5): 059202. doi: 10.7498/aps.60.059202
    [15] 赵小峰, 黄思训. 垂直天线阵观测信息反演大气折射率廓线. 物理学报, 2011, 60(11): 119203. doi: 10.7498/aps.60.119203
    [16] 赵延来, 黄思训, 杜华栋, 仲跻芹. 正则化方法同化多普勒天气雷达资料及对降雨预报的影响. 物理学报, 2011, 60(7): 079202. doi: 10.7498/aps.60.079202
    [17] 姜祝辉, 黄思训, 何然, 周晨腾. 合成孔径雷达资料反演海面风场的正则化方法研究. 物理学报, 2011, 60(6): 068401. doi: 10.7498/aps.60.068401
    [18] 刘广东, 张业荣. 二维有耗色散介质的时域逆散射方法. 物理学报, 2010, 59(10): 6969-6979. doi: 10.7498/aps.59.6969
    [19] 盛峥, 黄思训. 变分伴随正则化方法从雷达回波反演海洋波导(Ⅱ):实际反演试验. 物理学报, 2010, 59(6): 3912-3916. doi: 10.7498/aps.59.3912
    [20] 盛峥, 黄思训. 变分伴随正则化方法从雷达回波反演海洋波导(Ⅰ):理论推导部分. 物理学报, 2010, 59(3): 1734-1739. doi: 10.7498/aps.59.1734
计量
  • 文章访问数:  4933
  • PDF下载量:  465
  • 被引次数: 0
出版历程
  • 收稿日期:  2016-04-05
  • 修回日期:  2016-07-14
  • 刊出日期:  2016-10-05

/

返回文章
返回