自适应特征选择的相关滤波跟踪算法
|
摘要
针对相关滤波方法对快速运动与快速变形的目标跟踪稳定性较差的问题,提出一种自适应特征选择的相关滤波跟踪算法。利用位置滤波器和颜色概率模型提取候选区域中的基础特征,对基础特征以不同的权重分配方式进行融合,得到多个融合特征。对融合特征进行可信度判定,选择可信度较高的融合特征作为当前帧的跟踪特征,估计出目标的候选位置。若最高可信度低于可信度阈值,启动检测器重新检测目标位置,否则候选位置即为目标最终位置。与此同时,对目标模型进行更新,确保模型对目标描述的准确性。在标准数据集OTB50和OTB100上进行大量实验,测试结果表明,所提出的跟踪方法在运动模糊、光照变化、快速运动等条件下具有较高的跟踪准确率和较好的稳健性。
The conventional correlation filtering methods are known to demonstrate poor tracking stability for fast moving and fast deforming targets. Therefore, this paper proposes a correlation filter tracking algorithm for adaptive feature selection. First, the basic features are extracted in the candidate regions using a position filter and a color probability model and fused in different weight combinations to obtain multiple fusion features. Then, the credibility of the fusion features is determined and the features with relatively high credibility are selected as the tracking features of the current frame to estimate the candidate position of the target. Finally, if the maximum credibility is less than the credibility threshold, the detector is activated to redetect the target position; otherwise, the candidate position is just the final position. Meanwhile, the target model is updated to ensure the accuracy of target description. The experimental results on the standard OTB50 and OTB100 datasets show that the proposed tracking method has relatively high tracking accuracy and good robustness under the conditions of motion blurring, illumination variation, and fast motion.
The conventional correlation filtering methods are known to demonstrate poor tracking stability for fast moving and fast deforming targets. Therefore, this paper proposes a correlation filter tracking algorithm for adaptive feature selection. First, the basic features are extracted in the candidate regions using a position filter and a color probability model and fused in different weight combinations to obtain multiple fusion features. Then, the credibility of the fusion features is determined and the features with relatively high credibility are selected as the tracking features of the current frame to estimate the candidate position of the target. Finally, if the maximum credibility is less than the credibility threshold, the detector is activated to redetect the target position; otherwise, the candidate position is just the final position. Meanwhile, the target model is updated to ensure the accuracy of target description. The experimental results on the standard OTB50 and OTB100 datasets show that the proposed tracking method has relatively high tracking accuracy and good robustness under the conditions of motion blurring, illumination variation, and fast motion.
引文
[1] Zhang T Z,Xu C S,Yang M H.Multi-task correlation particle filter for robust object tracking[C]//2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR),July 21-26,2017,Honolulu,HI,USA.New York:IEEE,2017:4819-4827.
[2] Zhao G P,Shen Y P,Wang J Y.Adaptive feature fusion object tracking based on circulant structure with kernel[J].Acta Optica Sinica,2017,37(8):0815001.赵高鹏,沈玉鹏,王建宇.基于核循环结构的自适应特征融合目标跟踪[J].光学学报,2017,37(8):0815001.
[3] Zhang Z,Sun J,Yang L T.Tracking algorithm based on correlation filter fusing and keypoint matching[J].Acta Optica Sinica,2019,39(2):0215001.张哲,孙瑾,杨刘涛.融合相关滤波与关键点匹配的跟踪算法[J].光学学报,2019,39(2):0215001.
[4] Ge B Y,Zuo X Z,Hu Y J.Long-term object tracking based on feature fusion[J].Acta Optica Sinica,2018,38(11):1115002.葛宝义,左宪章,胡永江.基于特征融合的长时目标跟踪算法[J].光学学报,2018,38(11):1115002.
[5] Lu H C,Li P X,Wang D.Visual object tracking:a survey[J].Pattern Recognition and Artificial Intelligence,2018,31(1):61-76.卢湖川,李佩霞,王栋.目标跟踪算法综述[J].模式识别与人工智能,2018,31(1):61-76.
[6] Bolme D S,Beveridge J R,Draper B A,et al.Visual object tracking using adaptive correlation filters[C]//2010 IEEE Computer Society Conference on Computer Vision and Pattern Recognition,June 13-18,2010,San Francisco,CA,USA.New York:IEEE,2010:2544-2550.
[7] Henriques J F,Caseiro R,Martins P,et al.Exploiting the circulant structure of tracking by detection with kernels[M]//Fitzgibbon A,Lazebnik S,Perona P,et al.Computer Vision-ECCV 2012.Berlin,Heidelberg:Springer,2012,7575:702-715.
[8] 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,2015,37(3):583-596.
[9] Danelljan M,H?ger G,Khan F,et al.Accurate scale estimation for robust visual tracking[C]//British Machine Vision Conference 2014,September 1-5,2014,Nottingham.Durham,England,UK:BMVA Press,2014:65.
[10] 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 (CVPR),June 27-30,2016,Las Vegas,NV,USA.New York:IEEE,2016:1401-1409.
[11] Tang M,Feng J Y.Multi-kernel correlation filter for visual tracking[C]//2015 IEEE International Conference on Computer Vision (ICCV),December 7-13,2015,Santiago,Chile.New York:IEEE,2015:3038-3046.
[12] Zhang K H,Zhang L,Yang M H,et al.Fast tracking via spatio-temporal context learning[EB/OL].(2013-11-08)[2018-12-15].https://arxiv.org/pdf/1311.1939.pdf.
[13] Danelljan M,H?ger G,Khan F S,et al.Discriminative scale space tracking[J].IEEE Transactions on Pattern Analysis and Machine Intelligence,2017,39(8):1561-1575.
[14] Possegger H,Mauthner T,Bischof H.In defense of color-based model-free tracking[C]//2015 IEEE Conference on Computer Vision and Pattern Recognition (CVPR),June 7-12,2015,Boston,MA,USA.New York:IEEE,2015:2113-2120.
[15] Wang M M,Liu Y,Huang Z Y.Large margin object tracking with circulant feature maps[C]//2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR),July 21-26,2017,Honolulu,HI,USA.New York:IEEE,2017:4800-4808.
[16] Bian J W,Lin W Y,Matsushita Y,et al.GMS:grid-based motion statistics for fast,ultra-robust feature correspondence[C]//2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR),July 21-26,2017,Honolulu,HI,USA.New York:IEEE,2017:2828-2837.
[17] Wu Y,Lim J,Yang M H.Online object tracking:a benchmark[C]//2013 IEEE Conference on Computer Vision and Pattern Recognition,June 23-28,2013,Portland,OR,USA.New York:IEEE,2013:2411-2418.
[18] Wu Y,Lim J,Yang M H.Object tracking benchmark[J].IEEE Transactions on Pattern Analysis and Machine Intelligence,2015,37(9):1834-1848.
[19] Li Y,Zhu J K.A scale adaptive kernel correlation filter tracker with feature integration[M]//Agapito L,Bronstein M,Rother C.Computer Vision-ECCV 2014 Workshops.Cham:Springer,2015,8693:254-265.
[20] Danelljan M,H?ger G,Khan F S,et al.Learning spatially regularized correlation filters for visual tracking[C]//2015 IEEE International Conference on Computer Vision (ICCV),December 7-13,2015,Santiago,Chile.New York:IEEE,2015:4310-4318.
[21] Galoogahi H K,Fagg A,Lucey S.Learning background-aware correlation filters for visual tracking[C]//2017 IEEE International Conference on Computer Vision (ICCV),October 22-29,2017,Venice,Italy.New York:IEEE,2017:1144-1152.
[22] Li F,Tian C,Zuo W M,et al.Learning spatial-temporal regularized correlation filters for visual tracking[C]//2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition,June 18-23,2018,Salt Lake City,UT,USA.New York:IEEE,2018:4904-4913.
[23] Danelljan M,Bhat G,Khan F S,et al.ECO:efficient convolution operators for tracking[C]//2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR),July 21-26,2017,Honolulu,HI,USA.New York:IEEE,2017:6931-6939.
[2] Zhao G P,Shen Y P,Wang J Y.Adaptive feature fusion object tracking based on circulant structure with kernel[J].Acta Optica Sinica,2017,37(8):0815001.赵高鹏,沈玉鹏,王建宇.基于核循环结构的自适应特征融合目标跟踪[J].光学学报,2017,37(8):0815001.
[3] Zhang Z,Sun J,Yang L T.Tracking algorithm based on correlation filter fusing and keypoint matching[J].Acta Optica Sinica,2019,39(2):0215001.张哲,孙瑾,杨刘涛.融合相关滤波与关键点匹配的跟踪算法[J].光学学报,2019,39(2):0215001.
[4] Ge B Y,Zuo X Z,Hu Y J.Long-term object tracking based on feature fusion[J].Acta Optica Sinica,2018,38(11):1115002.葛宝义,左宪章,胡永江.基于特征融合的长时目标跟踪算法[J].光学学报,2018,38(11):1115002.
[5] Lu H C,Li P X,Wang D.Visual object tracking:a survey[J].Pattern Recognition and Artificial Intelligence,2018,31(1):61-76.卢湖川,李佩霞,王栋.目标跟踪算法综述[J].模式识别与人工智能,2018,31(1):61-76.
[6] Bolme D S,Beveridge J R,Draper B A,et al.Visual object tracking using adaptive correlation filters[C]//2010 IEEE Computer Society Conference on Computer Vision and Pattern Recognition,June 13-18,2010,San Francisco,CA,USA.New York:IEEE,2010:2544-2550.
[7] Henriques J F,Caseiro R,Martins P,et al.Exploiting the circulant structure of tracking by detection with kernels[M]//Fitzgibbon A,Lazebnik S,Perona P,et al.Computer Vision-ECCV 2012.Berlin,Heidelberg:Springer,2012,7575:702-715.
[8] 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,2015,37(3):583-596.
[9] Danelljan M,H?ger G,Khan F,et al.Accurate scale estimation for robust visual tracking[C]//British Machine Vision Conference 2014,September 1-5,2014,Nottingham.Durham,England,UK:BMVA Press,2014:65.
[10] 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 (CVPR),June 27-30,2016,Las Vegas,NV,USA.New York:IEEE,2016:1401-1409.
[11] Tang M,Feng J Y.Multi-kernel correlation filter for visual tracking[C]//2015 IEEE International Conference on Computer Vision (ICCV),December 7-13,2015,Santiago,Chile.New York:IEEE,2015:3038-3046.
[12] Zhang K H,Zhang L,Yang M H,et al.Fast tracking via spatio-temporal context learning[EB/OL].(2013-11-08)[2018-12-15].https://arxiv.org/pdf/1311.1939.pdf.
[13] Danelljan M,H?ger G,Khan F S,et al.Discriminative scale space tracking[J].IEEE Transactions on Pattern Analysis and Machine Intelligence,2017,39(8):1561-1575.
[14] Possegger H,Mauthner T,Bischof H.In defense of color-based model-free tracking[C]//2015 IEEE Conference on Computer Vision and Pattern Recognition (CVPR),June 7-12,2015,Boston,MA,USA.New York:IEEE,2015:2113-2120.
[15] Wang M M,Liu Y,Huang Z Y.Large margin object tracking with circulant feature maps[C]//2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR),July 21-26,2017,Honolulu,HI,USA.New York:IEEE,2017:4800-4808.
[16] Bian J W,Lin W Y,Matsushita Y,et al.GMS:grid-based motion statistics for fast,ultra-robust feature correspondence[C]//2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR),July 21-26,2017,Honolulu,HI,USA.New York:IEEE,2017:2828-2837.
[17] Wu Y,Lim J,Yang M H.Online object tracking:a benchmark[C]//2013 IEEE Conference on Computer Vision and Pattern Recognition,June 23-28,2013,Portland,OR,USA.New York:IEEE,2013:2411-2418.
[18] Wu Y,Lim J,Yang M H.Object tracking benchmark[J].IEEE Transactions on Pattern Analysis and Machine Intelligence,2015,37(9):1834-1848.
[19] Li Y,Zhu J K.A scale adaptive kernel correlation filter tracker with feature integration[M]//Agapito L,Bronstein M,Rother C.Computer Vision-ECCV 2014 Workshops.Cham:Springer,2015,8693:254-265.
[20] Danelljan M,H?ger G,Khan F S,et al.Learning spatially regularized correlation filters for visual tracking[C]//2015 IEEE International Conference on Computer Vision (ICCV),December 7-13,2015,Santiago,Chile.New York:IEEE,2015:4310-4318.
[21] Galoogahi H K,Fagg A,Lucey S.Learning background-aware correlation filters for visual tracking[C]//2017 IEEE International Conference on Computer Vision (ICCV),October 22-29,2017,Venice,Italy.New York:IEEE,2017:1144-1152.
[22] Li F,Tian C,Zuo W M,et al.Learning spatial-temporal regularized correlation filters for visual tracking[C]//2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition,June 18-23,2018,Salt Lake City,UT,USA.New York:IEEE,2018:4904-4913.
[23] Danelljan M,Bhat G,Khan F S,et al.ECO:efficient convolution operators for tracking[C]//2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR),July 21-26,2017,Honolulu,HI,USA.New York:IEEE,2017:6931-6939.