Laser & Optoelectronics Progress, Volume. 58, Issue 20, 2010016(2021)

Weak-Light Image Segmentation Based on Intercept Histogram and Otsu Fusion

Yanyang Zeng, Gaosen Xie, and Jianchun Zhang*
Author Affiliations
  • Computer Science and Technical College, Henan Polytechnic University, Jiaozuo, Henan 454003, China
  • show less
    Figures & Tables(14)
    Regions in 2D histogram. (a) Histogram method; (b) line interpret histogram method
    Directions about a pixel. (a) Eight directions of center pixel; (b) NW direction of edge of SW-NE direction
    Algorithm segmentation results. (a) Original image; (b) result obtained by Otsu; (c) result obtained by directional fuzzy derivative; (d) fusion result
    Algorithm segmentation results. (a) Original image 1; (b) Otsu; (c) 2D maximum entropy; (d) algorithm in Ref. [18]; (e) algorithm in Ref. [23]; (f) proposed method
    Algorithm segmentation results. (a) Original image 2; (b) Otsu; (c) 2D maximum entropy; (d) algorithm in Ref. [18]; (e) algorithm in Ref. [23]; (f) proposed method
    Algorithm segmentation results. (a) Original image 3; (b)Otsu; (c) 2D maximum entropy; (d) algorithm in Ref. [18]; (e) algorithm in Ref. [23]; (f) proposed method
    Fuzzy derivative segmentation results. (a) Segmentation result of Fig. 4(a); (b) segmentation result of Fig. 5(a); (c) segmentation result of Fig. 6(a)
    Algorithm segmentation results of image 1. (a) Original image; (b) Otsu; (c) 2D maximum entropy; (d) algorithm in Ref. [18]; (e) algorithm in Ref. [23]; (f) proposed method
    Algorithm segmentation results of image 2. (a) Original image; (b) Otsu; (c) 2D maximum entropy; (d) algorithm in Ref. [18]; (e) algorithm in Ref. [23]; (f) proposed method
    Algorithm segmentation results of image 3. (a) Original image; (b) Otsu; (c) 2D maximum entropy; (d) algorithm in Ref. [18]; (e) algorithm in Ref. [23]; (f) proposed method
    Algorithm segmentation results of image 4. (a) Original image; (b) Otsu; (c) 2D maximum entropy; (d) algorithm in Ref. [18]; (e) algorithm in Ref. [23]; (f) proposed method
    Algorithm segmentation results of image 5. (a) Original image; (b) Otsu; (c) 2D maximum entropy; (d) algorithm in Ref. [18]; (e) algorithm in Ref. [23]; (f) proposed method
    • Table 1. Quantitative comparison of various segmentation methods

      View table

      Table 1. Quantitative comparison of various segmentation methods

      AlgorithmParameterFig. 8Fig. 9Fig. 10Fig. 11Fig. 12
      RMSE0.18520.18560.18780.18630.1898
      OtsuSSIM0.99640.99640.99640.99640.9964
      PSNR29.295729.254029.051029.190928.8709
      2D maximumentropyRMSE0.23560.15160.25860.27810.2970
      SSIM0.99400.99770.99270.99150.9902
      PSNR25.110532.770023.493322.231521.0909
      RMSE0.16860.18160.15200.21240.1954
      Algorithm in Ref. [18]SSIM0.99890.99860.99920.99790.9983
      PSNR30.921929.631832.731426.914828.3674
      RMSE0.18380.18420.18520.18430.1885
      Algorithm in Ref. [23]SSIM0.99640.99640.99640.99640.9963
      PSNR29.430129.387429.290829.374428.9850
      RMSE0.02920.02210.04350.04160.0416
      Proposed methodSSIM0.99980.99980.99980.99980.9998
      PSNR61.389566.261654.442455.241055.2410
    • Table 2. Quantitative comparison of various segmentation methods

      View table

      Table 2. Quantitative comparison of various segmentation methods

      AlgorithmParameterFig. 8Fig. 9Fig. 10Fig. 11Fig. 12
      RMSE0.22750.22850.30650.23920.2275
      OtsuSSIM0.99620.99610.99150.99590.9962
      PSNR25.724525.643720.540424.849425.7172
      2D maximumentropyRMSE0.26480.20070.28780.30470.3070
      SSIM0.99410.99740.99270.99160.9907
      PSNR23.080227.898621.635820.644820.5169
      RMSE0.21620.18410.17490.22860.2170
      Algorithm in Ref. [18]SSIM0.99800.99880.99900.99770.9980
      PSNR26.602129.400730.287025.640226.5443
      RMSE0.22700.22650.30450.23070.2261
      Algorithm in Ref. [23]SSIM0.99620.99620.99160.99610.9962
      PSNR25.762025.796220.658525.476125.8264
      RMSE0.14310.13950.16440.16730.1429
      Proposed methodSSIM0.99950.99950.99930.99930.9995
      PSNR33.779634.211831.365031.064333.7985
    Tools

    Get Citation

    Copy Citation Text

    Yanyang Zeng, Gaosen Xie, Jianchun Zhang. Weak-Light Image Segmentation Based on Intercept Histogram and Otsu Fusion[J]. Laser & Optoelectronics Progress, 2021, 58(20): 2010016

    Download Citation

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

    Category: Image Processing

    Received: Oct. 9, 2020

    Accepted: Jan. 13, 2021

    Published Online: Oct. 13, 2021

    The Author Email: Zhang Jianchun (1294082095@qq.com)

    DOI:10.3788/LOP202158.2010016

    Topics