Journals Articles Conferences News About CLP
Submit a paper Email Alert
Search
Search
Articles
Journals
News
Advanced Search
Top Searches
2D MaterialsTransformation opticsQuantum PhotonicsSmall lasersSemiconductor UV PhotonicsSupersymmetric microring laser arrays

Laser & Optoelectronics Progress, Volume. 61, Issue 10, 1037007(2024)

Low-Light Image Enhancement Algorithm Based on Multiscale Depth Curve Estimation

Hongda Guo1、*, Xiucheng Dong1, Yongkang Zheng2, Yaling Ju1, and Dangcheng Zhang1
Author Affiliations
  • 1School of Electrical Engineering and Electronic Information, Xihua University, Chengdu 610039, Sichuan , China
  • 2Sichuan Electric Power Research Institute, State Grid, Chengdu 610041, Sichuan , China
    • show less
    AbstractGet PDF(in Chinese)
    Figures&Tables (13)
    Equations (7)
    References (26)
    Cited By (1)
    Tools
    Paper Information
    Topics
    Figures & Tables(13)
    Overall network framework

    Fig. 1. Overall network framework

    Download full size
    Network framework of multi-scale parameter estimation module

    Fig. 2. Network framework of multi-scale parameter estimation module

    Download full size
    Network framework of single-scale parameter estimation module

    Fig. 3. Network framework of single-scale parameter estimation module

    Download full size
    Comparison results of the first set of images. (a) Low-light image; (b) Zero-DCE; (c) Zero-DCE++; (d) KinD++; (e) LIME; (f) URetinex-Net; (g) TFR; (h) proposed method

    Fig. 4. Comparison results of the first set of images. (a) Low-light image; (b) Zero-DCE; (c) Zero-DCE++; (d) KinD++; (e) LIME; (f) URetinex-Net; (g) TFR; (h) proposed method

    Download full size
    Comparison results of the second set of images. (a) Low-light image; (b) Zero-DCE; (c) Zero-DCE++; (d) KinD++; (e) LIME; (f) URetinex-Net; (g) TFR; (h) proposed method

    Fig. 5. Comparison results of the second set of images. (a) Low-light image; (b) Zero-DCE; (c) Zero-DCE++; (d) KinD++; (e) LIME; (f) URetinex-Net; (g) TFR; (h) proposed method

    Download full size
    Comparison results of the third set of images. (a) Low-light image; (b) Zero-DCE; (c) Zero-DCE++; (d) KinD++; (e) LIME; (f) URetinex-Net; (g) TFR; (h) proposed method

    Fig. 6. Comparison results of the third set of images. (a) Low-light image; (b) Zero-DCE; (c) Zero-DCE++; (d) KinD++; (e) LIME; (f) URetinex-Net; (g) TFR; (h) proposed method

    Download full size
    Loss function ablation experiment. (a) Low-light image; (b) proposed method; (c) model A; (d) model B; (e) model C; (f) model D

    Fig. 7. Loss function ablation experiment. (a) Low-light image; (b) proposed method; (c) model A; (d) model B; (e) model C; (f) model D

    Download full size
    Ablation experiment of multi-scale parameter estimation module. (a) Single scale; (b) two scales; (c) three scales

    Fig. 8. Ablation experiment of multi-scale parameter estimation module. (a) Single scale; (b) two scales; (c) three scales

    Download full size
    Ablation experiment of noise refinement module. (a)‒(c) Results without using noise refinement; (d)‒(f) results of selecting noise refinement

    Fig. 9. Ablation experiment of noise refinement module. (a)‒(c) Results without using noise refinement; (d)‒(f) results of selecting noise refinement

    Download full size

    • Table 1. Full reference image quality evaluation results of different methods

      View table

      Table 1. Full reference image quality evaluation results of different methods

      MethodSSIMPSNR
      Zero-DCE0.638814.6304
      Zero-DCE++0.653214.7195
      KinD++0.713315.3417
      LIME0.699214.1325
      URetinex-Net0.729715.6845
      TFR0.721315.5998
      Proposed method0.729515.7043

    • Table 2. NIQE results of different methods in two test sets

      View table

      Table 2. NIQE results of different methods in two test sets

      MethodDICM datasetLIME dataset
      Zero-DCE3.71254.7037
      Zero-DCE++3.80374.5185
      KinD++3.51034.4169
      LIME3.94515.0280
      URetinex-Net3.11574.2377
      TFR3.14554.5350
      Proposed method3.20254.2157

    • Table 3. Comparison results of bright channel loss function and exposure control loss function

      View table

      Table 3. Comparison results of bright channel loss function and exposure control loss function

      MethodSSIMPSNR
      Zero-DCE0.638814.6304
      Zero-DCE-bri0.644214.6413

    • Table 4. Loss function of each model

      View table

      Table 4. Loss function of each model

      LossModel AModel BModel CModel D
      Lbri×√√√
      Lspa√×√√
      LtvA√√×√
      Lcol√√√×
    Tools

    Get Citation

    Copy Citation Text

    Hongda Guo, Xiucheng Dong, Yongkang Zheng, Yaling Ju, Dangcheng Zhang. Low-Light Image Enhancement Algorithm Based on Multiscale Depth Curve Estimation[J]. Laser & Optoelectronics Progress, 2024, 61(10): 1037007

    Download Citation

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

    Category: Digital Image Processing

    Received: Aug. 28, 2023

    Accepted: Oct. 9, 2023

    Published Online: Apr. 29, 2024

    The Author Email: Hongda Guo (476642413@qq.com)

    DOI:10.3788/LOP231997

    CSTR:32186.14.LOP231997

    Topics

    laser devices and laser physics
    Lasers and Laser Optics
    Laser physics
    laser manufacturing
    Instrumentation, Measurement and Metrology