• Infrared and Laser Engineering
  • Vol. 51, Issue 6, 20210532 (2022)
Xikui Miao, Yanxiu Zhang, Hengwei Zhang, Xiaohu Liu, and Qianjin Zou
Author Affiliations
  • Key Laboratory of Electro-Optical Countermeasures Test & Evaluation Technology, Luoyang 471003, China
  • show less
    DOI: 10.3788/IRLA20210532 Cite this Article
    Xikui Miao, Yanxiu Zhang, Hengwei Zhang, Xiaohu Liu, Qianjin Zou. Image background clutter modeling method based on directional selectivity mechanism[J]. Infrared and Laser Engineering, 2022, 51(6): 20210532 Copy Citation Text show less
    Illustration of the excitation and inhibition response of neurons in the local receptive field
    Fig. 1. Illustration of the excitation and inhibition response of neurons in the local receptive field
    Visual pattern based on the directional selectivity mechanism in 8-neighbors
    Fig. 2. Visual pattern based on the directional selectivity mechanism in 8-neighbors
    Directional selectivity vision pattern of target and background clutter
    Fig. 3. Directional selectivity vision pattern of target and background clutter
    Directional selectivity vision pattern (before/after dimension reduction)
    Fig. 4. Directional selectivity vision pattern (before/after dimension reduction)
    The difference between visual pattern histograms of tank image and background clutter
    Fig. 5. The difference between visual pattern histograms of tank image and background clutter
    Image gradient filter
    Fig. 6. Image gradient filter
    Clutter metricC50EPLCCSRCCRMSE
    SV3.87394.06790.41960.64320.1513
    TSSIM0.2529−3.99870.57990.71000.1357
    POE121.32470.77720.44460.63980.1387
    VSD0.08793.02480.7510.7520.0718
    ESSIM0.620418.95650.8700.8060.0569
    SRrms0.973214.03220.8790.8130.0541
    SRavg0.856916.87650.8830.8150.0549
    Table 1. Correlation between detection probability and clutter models
    Clutter metricC50EPLCCSRCCRMSE
    SV5.91792.67910.59240.53660.0762
    TSSIM0.700215.70010.69100.54500.0830
    POE4.51270.78560.61800.55410.0821
    VSD0.08152.77510.8090.7740.0511
    ESSIM0.600314.22590.8450.8320.0459
    SRrms0.87618.87540.8420.8320.0451
    SRavg1.54226.67670.8450.8340.0462
    Table 2. Correlation between false probability and clutter models
    Clutter metricxyPLCCSRCCRMSE
    SV8.3006−2.06840.46350.65833.9075
    TSSIM0.5255−0.72450.42300.75623.2510
    POE995.5658.40360.61560.75053.8601
    VSD0.779−0.06330.7560.52803.1501
    ESSIM0.5792−0.60740.8970.59502.1287
    SRrms0.7654−0.76210.8810.60452.1143
    SRavg0.8012−0.81040.8750.60512.1138
    Table 3. Correlation between search time and clutter models
    Threshold TPDFDST
    30.07540.07722.2656
    40.06720.06832.2078
    50.05940.05072.1592
    60.05410.04512.1143
    70.05890.04962.1672
    80.06220.06452.2134
    90.07810.07852.2704
    Table 4. RMSE between PD, FD, ST and subjective test results under different directional similarity thresholds
    Xikui Miao, Yanxiu Zhang, Hengwei Zhang, Xiaohu Liu, Qianjin Zou. Image background clutter modeling method based on directional selectivity mechanism[J]. Infrared and Laser Engineering, 2022, 51(6): 20210532
    Download Citation