Laser & Optoelectronics Progress, Volume. 56, Issue 22, 221502(2019)

Tracking Algorithm of Correlation Filter with Multiple Features Based on Temporal Consistency and Spatial Pruning

Yixuan Wang, Xiaojun Wu*, and Tianyang Xu
Author Affiliations
  • International Joint Laboratory of Pattern Recognition and Computational Intelligence, School of Internet of Things Engineering, Jiangnan University, Wuxi, Jiangsu 214122, China
  • show less
    Figures & Tables(10)
    Flow chart of correlation filter tracking algorithm for multiple features based on temporal consistency and spatial pruning
    Schematic of binary matrix mask
    Comparison of generated masks by proposed algorithm and CSRDCF algorithm. (a) Mask generated by proposed algorithm based on by single frame image in Matrix video; (b) mask generated by CSRDCF algorithm based on by single frame image in Matrix video; (c) single frame image in Matrix video; (d) mask generated by proposed algorithm based on by single frame image in Basketball video; (e) mask generated by CSRDCF algorithm based on by single frame image in Basketball video; (f) single frame image in
    Average AUC and average DP curves of 7 algorithms on OTB-100 dataset. (a) Average AUC; (b) average DP
    Comparison of 9 algorithms on VOT2016 dataset. (a) EAO; (b) AUC
    Comparison of tracking results of 6 algorithms on Bird1 and Lemming. (a) Bird1; (b) Lemming
    Comparisonof tracking results of 6 algorithms on Iron Man and Matrix. (a) Iron Man; (b) Matrix
    • Table 1. Average AUC results of 6 algorithms in 6 challenges on OTB-100 dataset

      View table

      Table 1. Average AUC results of 6 algorithms in 6 challenges on OTB-100 dataset

      AlgorithmOCCBCDEFMBIPRSV
      Ours(D+HC)0.6690.6940.6510.6990.6580.670
      ECO-HC0.6140.6400.6120.6290.6060.608
      CCOT0.6200.6130.5850.6630.5820.601
      BACF0.5560.6050.5720.5700.5750.575
      SRDCF0.5540.5830.5400.5900.5420.565
      CSRDCF0.5300.5440.5310.5830.5240.528
    • Table 2. Average DP results of 6 algorithms in 6 challenges on OTB-100 dataset

      View table

      Table 2. Average DP results of 6 algorithms in 6 challenges on OTB-100 dataset

      AlgorithmOCCBCDEFMBIPRSV
      Ours(D+HC)0.8690.9140.8700.8880.8820.882
      ECO-HC0.8100.8430.8250.7900.7780.803
      CCOT0.8510.8320.8280.8440.8330.822
      BACF0.7310.8010.7660.7330.7900.773
      SRDCF0.7280.7750.7310.7600.7380.747
      CSRDCF0.7320.7560.7500.7470.7480.750
    • Table 3. Tracking results of 9 algorithms on VOT2016 dataset

      View table

      Table 3. Tracking results of 9 algorithms on VOT2016 dataset

      PerformanceOurs (D+HC)CCOTTCNNSSATSTAPLEDDCEBTSTAPLEpDeepSRDCF
      EAO0.400.330.320.320.300.290.290.290.28
      Failures8.9216.5817.9419.2723.9020.9815.1924.3220.35
      AO0.530.470.490.520.390.390.370.390.43
    Tools

    Get Citation

    Copy Citation Text

    Yixuan Wang, Xiaojun Wu, Tianyang Xu. Tracking Algorithm of Correlation Filter with Multiple Features Based on Temporal Consistency and Spatial Pruning[J]. Laser & Optoelectronics Progress, 2019, 56(22): 221502

    Download Citation

    EndNote(RIS)BibTexPlain Text
    Save article for my favorites
    Paper Information

    Category: Machine Vision

    Received: Apr. 11, 2019

    Accepted: May. 9, 2019

    Published Online: Nov. 2, 2019

    The Author Email: Xiaojun Wu (xiaojun_wu_jnu@163.com)

    DOI:10.3788/LOP56.221502

    Topics