| 399 | 12 | 236 |
| 下载次数 | 被引频次 | 阅读次数 |
精度和速度一直是跟踪领域的矛盾。相比而言,基于深度学习方法的模型精度更好,但基于相关滤波(可以用FFT加速)的模型速度快很多。算法KCF、DSST、Staple、SiamFC、ECO、CCOT等在VOT2016数据库上的跟踪性能对比结果显示,在跟踪准确率这方面,深度学习优于相关滤波跟踪算法,而在鲁棒性方面,相关滤波算法则占有优势,且其速度也一直领先。将深度卷积特征和相关滤波相结合,可以兼顾两者的优势,使相应算法表现出更好性能。未来应着重考虑发挥CNN在目标跟踪领域的作用,以其同时提高算法的实时性和训练的便捷性。跟踪和检测是分不开的,跟踪能够保证速度上的需要,而检测能够有效地修正跟踪的累计误差。不同的应用场合对跟踪的成功率、准确度和鲁棒性要求也不一样,达到实际的跟踪要求仍然需要更好的算法实现。
Abstract：Accuracy and speed have always been a contradiction in the tracking field.The model based on deep learning is more accurate,but the model based on correlation filtering(which can be accelerated by FFT)is much faster.The tracking performance comparison results of algorithms KCF,DSST,Staple,SiamFC,ECO and CCOT on VOT2016 database show that in terms of tracking accuracy,deep learning is superior to correlation filtering tracking algorithm,while in terms of robustness,correlation filtering algorithm has an advantage and its speed has always been ahead.By combining the features of deep convolution with correlation filtering,the advantages of both can be taken into account,so that the corresponding algorithm can perform better.In the future,we should focus on giving full play to CNN's role in the field of target tracking,so as to improve the real-time performance of the algorithm and the convenience of training.Tracking and detection are inseparable,and tracking can ensure the need for speed,while detection can effectively correct the accumulated tracking error.The success rate,accuracy and robustness of tracking are also different in different applications,and better algorithm is still needed to achieve the actual tracking requirements.
[1] LI X,HU W,SHEN C.A survey of appearance models in visual object tracking[J/OL].Acm Transactions on Intelligent Systems&Technology,2013,4(4):1-48[2017-12-06].http://dx.doi.org/10.1145/2508037.2508039.
[2] HARE S,SAFFARI A,TORR P H S.Struck:Structured output tracking with kernels[C/OL]//2011International Conference on Computer Vision.Barcelona:IEEE,2011:263-270[2017-12-13].http://dx.doi.org/10.1109/ICCV.2011.6126251.
[3] YIN Y J,XU D,X,WANG X G,et al.Online stateBased structured SVM combined with incremental PCA for robust visual Tracking[J/OL].IEEE transactions on Cybernetics,2015,45(9):1988-2000[2017-12-06].http://dx.doi.org/10.1109/TCYB.2014.2363078.
[4] GAO Y,SHAN X,HU Z,et al.Extended compressed tracking via random projection based on MSERs and online LS-SVM learning[J/OL].Pattern Recognition,2016,59(C):245-254[2017-12-13].http://dx.doi.org/10.1016/j.patcog.2016.02.012.
[5] NING J,YANG J,JIANG S,et al.Object Tracking via Dual Linear Structured SVM and Explicit Feature Map[C/OL]//2016IEEE Conference on Computer Vision and Pattern Recognition(CVPR).Las Vegas:IEEE,2016:4266-4274[2017-12-06].http://dx.doi.org/10.1109/CVPR.2016.462.
[6] BABENKO B,YANG M H,BELONGIE S.Robust object tracking with online multiple instance learning[J/OL].IEEE Transactions on Pattern Analysis and Machine Intelligence,2011,33(8):1619-1632[2017-12-13]. http://dx. doi. org/10.1109/TPAMI.2010.226.
[7] ZHANG K,SONG H.Real-time visual tracking via online weighted multiple instance learning[J/OL].Pattern Recognition,2013,46(1):397-411[2017-12-06].http://dx.doi.org/10.1016/j.patcog.2012.07.013.
[8]王丽佳,贾松敏,李秀智,等.基于改进在线多示例学习算法的机器人目标跟踪[J/OL].自动化学报,2014,40(12):2916-2925[2017-12-13].http://www.oalib.com/paper/4418212.DOI:10.3724/SP.J.1004.2014.02916.
[9] LI N.Object Tracking with Multiple Instance Learning and Gaussian Mixture Model[J/OL].Journal of Information&Computational Science,2015,12(11):4465-4477[2017-12-13]. http://dx.doi.org/10.12733/jics20106217.
[10]SHARMA V K,MAHAPATRA K K.MIL based visual object tracking with kernel and scale adaptation[J/OL].Signal Processing:Image Communication,2017,53:51-64[2017-12-06].http://dx.doi.org/10.1016/j.image.2017.01.007.
[11]BOLME D S,BEVERIDGE J R,DRAPER B A,et al.Visual object tracking using adaptive correlation filters[C/OL]//2010IEEE Computer Society Conference on Computer Vision and Pattern Recognition.San Francisco:IEEE,2010:2544-2550[2017-12-08].http://dx.doi.org/10.1109/CVPR.2010.5539960.
[12] HARE S,GOLODETZ S,SAFFARI A,et al.Struck:Structured output tracking with kernels[J/OL].IEEE Transactions on Pattern Analysis and Machine Intelligence,2016,38(10):2096-2109[2017-12-22]. http://dx. doi. org/10. 1109/TPAMI.2015.2509974.
[13]HENRIQUES J,O F,CASEIRO R MARTINS P,et al.Exploiting the Circulant Structure of Tracking-byDetection with Kernels[C/OL]//FITZGIBBON A,LAZEBNIK S,PERONA P,et al.Computer Vision–ECCV 2012:Lecture Notes in Computer Science(vol 7575).Berlin:Springer,2012:702-715[2017-12-22].http://dx.doi.org/10.1007/978-3-642-33765-9_50.
[14]KALAL Z,MIKOAJCZYK K,MATAS J.TrackingLearning-Detection[J/OL].IEEE Transactions on Pattern Analysis&Machine Intelligence,2012,34(7):1409-1422[2017-12-08].http://doi.ieeecomputersociety.org/10.1109/TPAMI.2011.239.
[15]DANELLJAN M,KHAN F S,FELSBERG M,et al.Adaptive Color Attributes for Real-Time Visual Tracking[C/OL]//2014IEEE Conference on Computer Vision and Pattern Recognition.Columbus:IEEE,2014:1090-1097[2017-12-08].http://dx.doi.org/10.1109/CVPR.2014.143.
[16]HENRIQUENS J F,CASERIO R,MARTINS P,et al.High-Speed Tracking with Kernelized Correlation Filters[J/OL].IEEE Transactions on Pattern Analysis&Machine Intelligence,2015,37(3):583-596[2017-12-22].http://dx.doi.org/10.1109/TPAMI.2014.2345390.
[17]LI Y,ZHU J.A Scale Adaptive Kernel Correlation Filter Tracker with Feature Integration[C/OL]//AGAPITO L,BRONSTEIN M,ROTHER C.Computer Vision-ECCV 2014Workshops.ECCV 2014:Lecture Notes in Computer Science(vol 8926).Cham:Springer,2014:254-265[2017-12-08].http://dx.doi.org/10.1007/978-3-319-16181-5_18.
[18]DANELLJAN M,HAGER G,KHAN F S,et al.Accurate Scale Estimation for Robust Visual Tracking[C/OL]//Proceedings of the British Machine Vision Conference 2014.[S.l.]:British Machine Vision Association,2014:65.1-65.11[2017-12-13].http://dx.doi.org/10.5244/C.28.65.
[19]DANELLJAN M,HAGER G,KHAN F S,et al.Discriminative Scale Space Tracking[J/OL].IEEE Transactions on Pattern Analysis&Machine ntelligence,2016,39(8):1561-1575[2017-12-06].http://doi. ieeecomputersociety. org/10. 1109/TPAMI.2016.2609928.
[20]DNAELLJAN M,HAGER G,KHAN F S,et al.Learning Spatially Regularized Correlation Filters for Visual Tracking[C/OL]//2015 IEEE International Conference on Computer Vision(ICCV).Santiago:IEEE,2015:4310-4318[2017-12-13].http://doi.ieeecomputersociety.org/10.1109/ICCV.2015.490.
[21]GALOOGAHI H K,SIM T,LUCEY S.Correlation filters with limited boundaries[C/OL]//2015IEEE Conference on Computer Vision and Pattern Recognition(CVPR).Boston:IEEE,2015:4630-4638[2017-12-06].http://doi.ieeecomputersociety.org/10.1109/CVPR.2015.7299094.
[22]DANELLJAN M,HAGER G,KHAN F S,et al.Convolutional Features for Correlation Filter Based Visual Tracking[C/OL]//2015 IEEE International Conference on Computer Vision Workshop(ICCVW).Santiago:IEEE,2015:621-629[2017-12-13].http://doi. ieeecomputersociety. org/10. 1109/ICCVW.2015.84.
[23]MA C,HUANG J B,YANG X,et al.Hierarchical Convolutional Features for Visual Tracking[C/OL]//2015IEEE International Conference on Computer Vision(ICCV).Santiago:IEEE,2015:3074-3082[2017-12-06].http://doi.ieeecomputersociety.org/10.1109/ICCV.2015.352.
[24]DANELLJAN M,HAGER G,KHAN F S,et al.Adaptive Decontamination of the Training Set:A Unified Formulation for Discriminative Visual Tracking[C/OL]//2016IEEE Conference on Computer Vision and Pattern Recognition(CVPR).Las Vegas:IEEE,2016:1430-1438[2017-12-13].http://dx.doi.org/10.1109/CVPR.2016.159.
[25]DANELLJAN M,ROBINSON A,KHAN F S,et al.Beyond Correlation Filters:Learning Continuous Convolution Operators for Visual Tracking[C/OL]LEIBE B,MATAS J,SEBE N,et al.Computer Vision–ECCV 2016.ECCV 2016:Lecture Notes in Computer Science(vol 9909).Cham:Springer,2016:472-488[2017-12-06].http://dx.doi.org/10.1007/978-3-319-46454-1_29.
[26] DANELLJAN M,BHAT G,KHAN F S,et al.ECO:Efficient Convolution Operators for Tracking[C/OL]//2017IEEE Conference on Computer Vision and Pattern Recognition(CVPR).Honolulu:IEEE,2017:6931-6939[2017-12-13].http://doi.ieeecomputersociety.org/10.1109/CVPR.2017.733.
[27]VALMADRE J,BERTINETTO L,HENRIQUES J,et al.End-to-End Representation Learning for Correlation Filter Based Tracking[C/OL]//2017IEEE Conference on Computer Vision and Pattern Recognition(CVPR).Honolulu:IEEE,2017:5000-5008[2017-12-06].http://doi.ieeecomputersociety.org/10.1109/CVPR.2017.531.
[28]BERTINETTO L,VALMADRE J,HENRIQUES J F,et al.Fully-Convolutional Siamese Networks for Object Tracking[C/OL]//HUA G,JEGOU H.Computer Vision–ECCV 2016Workshops.ECCV 2016:Lecture Notes in Computer Science(vol 9914):Cham:Springer,2016:850-865[2017-12-22].http://dx.doi.org/10.1007/978-3-319-48881-3_56.
[29]LANDOLA F N,HAN S,MOSKEWICZ M W,et al.SqueezeNet:AlexNet-level accuracy with 50xfewer parameters and <0.5 MB model size[J/OL].Computer Science:Computer Vision and Pattern Recognition,(2016-11-04)[2017-12-08].https://arxiv.org/abs/1602.07360v4.
[30]HOWARD A G,ZHU M L,CHEN B,et al.MobileNets:Efficient Convolutional Neural Networks for Mobile Vision Applications[J/OL].Computer Science:Computer Vision and Pattern Recognition,(2017-04-17)[2017-12-22].https://arxiv.org/abs/1704.04861v1.
[31]COURBARIAUX M,HUBARA I,SOUDRY D,et al.Binarized Neural Networks:Training Deep Neural Networks with Weights and Activations Constrained to+1 or-1[J/OL].Computer Science:Machine Learning,(2016-03-17)[2017-12-08].https://arxiv.org/abs/1602.02830v3.
[32]RASTEGARI M,ORDONEZ V,REDMON J,et al.XNOR-Net:ImageNet Classification Using Binary Convolutional Neural Networks[C/OL]//LEIBE B,MATAS J,SEBE N,et al.Computer Vision–ECCV 2016.ECCV 2016:Lecture Notes in Computer Science(vol 9908).Cham:Springer,2016:525-542[2017-12-22].http://dx.doi.org/10.1007/978-3-319-46493-0_32.
[33]YILMAZ A,JAVED O,SHAH M.Object tracking:a survey[J/OL].Acm Computing Surveys,2006,38(4):13[2017-12-08].http://dx.doi.org/10.1145/1177352.1177355.
[34]ZHANG K,ZHANG L,YANG M H.Real-Time Compressive Tracking[C/OL]//FITAGIBBON A,LAZEBNIK S,PERONA P,et al.Computer Vision–ECCV 2012.ECCV 2012:Lecture Notes in Computer Science(vol 7574).Berlin:Springer,2012:864-877[2017-12-22].http://dx.doi.org/10.1007/978-3-642-33712-3_62.
[35]WU Y,LIM J,YANG M H.Online Object Tracking:A Benchmark[C/OL]//2013 IEEE Conference on Computer Vision and Pattern Recognition.Portland:IEEE,2013:2411-2418[2017-12-08].http://doi.ieeecomputersociety.org/10.1109/CVPR.2013.312.
[36]WU Y,LIM J,YANG M H.Object Tracking Benchmark[J/OL].IEEE Transactions on Pattern Analysis&Machine Intelligence,2015,37(9):1834-1848[2017-12-22].http://doi.ieeecomputersociety.org/10.1109/TPAMI.2014.2388226.
[37]LIANG P,BLASCH E,LING H.Encoding Color Information for Visual Tracking:Algorithms and Benchmark[J/OL].IEEE Transactions on Image Processing,2015,24(12):5630-5644[2017-12-08].http://dx.doi.org/10.1109/TIP.2015.2482905.
[38]UELLER M,SMITH N,GHANEM B.A Benchmark and Simulator for UAV Tracking[C/OL]//LEIBE B,MATAS J,SEBE N,et al.Computer Vision–ECCV 2016.ECCV 2016:Lecture Notes in Computer Science(vol 9905).Cham:Springer,2016:445-461[2017-12-22].http://dx.doi.org/10.1007/978-3-319-46448-0_27.
[39]DENTON E,CHINTALA S,SZLAM A,et al.Deep generative image models using a Laplacian pyramid of adversarial networks[C/OL]//NIPS'15Proceedings of the 28th International Conference on Neural Information Processing Systems(Volume 1).Cambridge:MIT Press,2015:1486-1494[2017-12-20].http://dl.acm.org/citation.cfm?id=2969239.2969405.
[40]BERTINETTO L,VALMADRE J,GOLODETZ S,et al.Staple:Complementary Learners for Real-Time Tracking[C/OL]//2016IEEE Conference on Computer Vision and Pattern Recognition(CVPR).Las Vegas:IEEE,2016:1401-1409[2017-12-08].http://doi.ieeecomputersociety.org/10.1109/CVPR.2016.156.
[41]ZHU G,WANG J,WU Y,et al.Collaborative Correlation Tracking[C/OL]//XIE X H,JONES M W,TAM G K L.Proceedings of the British Machine Vision Conference(BMVC).[S.l.]:BMVA Press,2015:184.1-184.12[2017-12-24].http://dx.doi.org/10.5244/C.29.184.
[42]HUANG D F,LUO L,WEN M,et al.Enable Scale and Aspect Ratio Adaptability in Visual Tracking with Detection Proposals[C/OL]//XIE X H,JONES M W,TAM G K L.Proceedings of the British Machine Vision Conference(BMVC).[S.l.]:BMVA Press,2015:185.1-185.12.[2017-12-08].http://dx.doi.org/10.5244/C.29.185.
基本信息:
DOI：10.13682/j.issn.2095-6533.2018.04.010
中图分类号:TP391.41;TN713
引用信息:
[1]李娜,牒谨,刘颖.基于相关滤波的目标跟踪方法[J].西安邮电大学学报,2018,23(04):58-67.DOI:10.13682/j.issn.2095-6533.2018.04.010.
基金信息:
电子信息现场勘验应用技术公安部重点实验室开放课题(EISI2016006)
2018-07-10
2018-07-10