Laser & Optoelectronics Progress, Volume. 60, Issue 4, 0415007(2023)

Nonlocal Neural Network-Based Moving Target Tracking Method

Liguo Zhang, Zijian Ma*, Mei Jin, and Yihui Li
Author Affiliations
  • Institute of Electrical Engineering, Yanshan University, Qinhuangdao 066000, Hebei, China
  • show less
    Figures & Tables(13)
    Schematic diagram of tracking method
    Backbone network structure
    NL module structure
    Visualization diagrams of classified branch characteristics. (a) Original result; (b) improved result
    Classification score data distribution. (a) Offset; (b) overlap
    RL module structure
    Visualization of tracking results. (a) Original result; (b) improved result
    Analysis of tracking results. (a) Occlusion and deformation-success rate; (b) occlusion and deformation -precision;(c) motion blur and background clutters-success rate; (d) motion blur and background clutters-precision
    Visualization of tracking results. (a) Girl2; (b) Board
    • Table 1. Ablation analysis of tracking results

      View table

      Table 1. Ablation analysis of tracking results

      TestNLRLDPAUC
      Z1X1
      AZMZRZAXMXRXRZ
      10.8220.629
      20.8020.611
      30.8030.615
      40.8080.618
      50.8130.627
      60.8290.626
      70.8360.633
    • Table 2. Comparison results of different algorithms on OTB100 occlusion and deformation dataset

      View table

      Table 2. Comparison results of different algorithms on OTB100 occlusion and deformation dataset

      AlgorithmDPAUCDetection speed /(frame·s-1
      SiamFC40.7070.52085
      SiamRPN60.7170.535180
      SiamFC++70.8010.593228
      Proposed algorithm0.8460.626217
    • Table 3. Comparison results of different algorithms on OTB100 motion blur and background interference dataset

      View table

      Table 3. Comparison results of different algorithms on OTB100 motion blur and background interference dataset

      AlgorithmDPAUCDetection speed /(frame·s-1
      SiamFC40.7130.54582
      SiamRPN60.7310.548178
      SiamFC++70.7940.612227
      Proposed algorithm0.8200.623215
    • Table 4. Comparison results of different algorithms on UAV123 dataset

      View table

      Table 4. Comparison results of different algorithms on UAV123 dataset

      AlgorithmDPAUCDetection speed /(frame·s-1
      SiamFC40.6910.49778
      SiamRPN60.7030.525177
      SiamFC++70.7440.580205
      Proposed algorithm0.7750.595198
    Tools

    Get Citation

    Copy Citation Text

    Liguo Zhang, Zijian Ma, Mei Jin, Yihui Li. Nonlocal Neural Network-Based Moving Target Tracking Method[J]. Laser & Optoelectronics Progress, 2023, 60(4): 0415007

    Download Citation

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

    Category: Machine Vision

    Received: Nov. 12, 2021

    Accepted: Dec. 21, 2021

    Published Online: Feb. 13, 2023

    The Author Email: Ma Zijian (1239038456@qq.com)

    DOI:10.3788/LOP212946

    Topics