Acta Optica Sinica, Volume. 40, Issue 3, 0315001(2020)

Multi-Scale Kernel Correlation Filter Algorithm for Visual Tracking Based on the Fusion of Adaptive Features

Faling Chen1,2,3,4,5、*, Qinghai Ding1,6, Zheng Chang1,2,4,5, Hongyu Chen1,2,3,4,5, Haibo Luo1,2,4,5, Bin Hui1,2,4,5, and Yunpeng Liu1,2,4,5
Author Affiliations
  • 1Shenyang Institute of Automation, Chinese Academy of Sciences, Shenyang, Liaoning 110016, China
  • 2Institutes for Robotics and Intelligent Manufacturing, Chinese Academy of Sciences, Shenyang, Liaoning 110169, China
  • 3University of Chinese Academy of Sciences, Beijing 100049, China
  • 4Key Laboratory of Opto-Electronic Information Processing, Chinese Academy of Sciences, Shenyang, Liaoning 110016, China
  • 5Key Laboratory of Image Understanding and Computer Vision, Shenyang, Liaoning 110016, China
  • 6Space Star Technology Co., Ltd., Beijing 100086, China
  • show less
    Figures & Tables(14)
    Schematic of adaptive features fusion process
    Relationship between weight adjustment ρ and the target tracking performance
    Distance precision curves and overlap precision curves of three target tracking algorithms. (a) Distance precision; (b) overlap precision
    Comparisons of estimated scale by the proposed algorithm and actual scale on four sequences.(a) Blurcar2; (b) Dog1; (c) Doll; (d) Carscale
    Distance precision curves and overlap precision curves of different target tracking algorithms. (a) Distance precision; (b) overlap precision
    Comparison of tracking results among five algorithms on David sequence
    Comparison of tracking results among five algorithms on Basketball sequence
    Comparison of tracking results among five algorithms on Carscale sequence
    Comparison of tracking results among five algorithms on Jogging1 sequence
    Comparison of tracking results among five algorithms on Trellis sequence
    Comparison of tracking results among five algorithms on Trellis sequence
    • Table 1. Tracking results of the proposed algorithm on four scale variation sequences

      View table

      Table 1. Tracking results of the proposed algorithm on four scale variation sequences

      SequenceMean ECLPd /% (ECL=20)Po /% (So=0.5)
      Blurcar23.42100.0100.0
      Dog13.81100.0100.0
      Doll2.2699.399.6
      Carscale3.84100.0100.0
    • Table 2. Pd scores of the top ten algorithms on eleven attributes

      View table

      Table 2. Pd scores of the top ten algorithms on eleven attributes

      AlgorithmIVDEFSVOCCMBFMIPROPROVBCLR
      Proposed0.7800.7370.7390.7610.6530.5810.7040.7510.6650.7140.424
      DSST0.7300.6360.7380.6920.5440.5130.7680.7250.5110.6940.497
      KCF0.6570.6980.6480.6950.5710.5340.6910.6780.5900.6760.387
      Struck0.5580.5210.6390.5640.5510.6040.6170.5970.5390.5850.545
      SCM0.5940.5860.6720.6400.3390.3330.5970.6180.4290.5780.305
      CN0.5760.6070.5990.6210.5510.4820.6740.6450.4380.6290.408
      TLD0.5370.5120.6060.5630.5180.5510.5840.5960.5760.4280.349
      VTD0.5570.5010.5970.5450.3750.3520.5990.6200.4620.5710.168
      VTS0.5730.4870.5820.5340.3750.3530.5790.6040.4550.5780.187
      CXT0.5010.4220.5500.4910.5090.5150.6100.5740.5100.4430.371
    • Table 3. Po of the top ten algorithms on eleven attributes

      View table

      Table 3. Po of the top ten algorithms on eleven attributes

      AlgorithmIVDEFSVOCCMBFMIPROPROVBCLR
      Proposed0.7120.7330.7210.7380.5910.5320.6740.7020.6720.6480.419
      DSST0.6810.6100.6400.6320.5280.5030.6790.6320.5120.6270.437
      SCM0.5680.5650.6350.5990.3390.3350.5600.5750.4490.5500.308
      KCF0.5430.6280.4740.5800.5610.5230.6130.5790.6100.6300.355
      Struck0.4910.4730.4710.4930.5180.5670.5280.5060.5500.5450.410
      TLD0.4600.4560.4940.4680.4820.4730.4760.4970.5160.3880.327
      ALSA0.5030.4560.5440.4510.2810.2600.4880.4940.3590.4680.163
      CN0.4500.5110.4210.4790.4800.4370.5500.5010.4580.5310.399
      VTS0.5030.4410.4530.4650.3280.3250.4770.4960.5080.5160.183
      VTD0.4800.4430.4600.4680.3200.3190.5000.5100.4910.5150.170
    Tools

    Get Citation

    Copy Citation Text

    Faling Chen, Qinghai Ding, Zheng Chang, Hongyu Chen, Haibo Luo, Bin Hui, Yunpeng Liu. Multi-Scale Kernel Correlation Filter Algorithm for Visual Tracking Based on the Fusion of Adaptive Features[J]. Acta Optica Sinica, 2020, 40(3): 0315001

    Download Citation

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

    Category: Machine Vision

    Received: Jul. 25, 2019

    Accepted: Sep. 29, 2019

    Published Online: Feb. 17, 2020

    The Author Email: Faling Chen (chfling@sia.cn)

    DOI:10.3788/AOS202040.0315001

    Topics