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. 57, Issue 14, 141021(2020)

Low-Light Image Enhancement Based on Multi-Branch All Convolutional Neural Network

Ruoyou Wu, Dexing Wang*, Hongchun Yuan**, Peng Gong, Guanqi Chen, and Dan Wang
Author Affiliations
  • School of Information, Shanghai Ocean University, Shanghai 201306, China
    • show less
    AbstractGet PDF(in Chinese)
    Figures&Tables (10)
    Equations (0)
    References (38)
    Cited By (1)
    Tools
    Paper Information
    Topics
    Figures & Tables(10)
    Structure of convolutional block attention module

    Fig. 1. Structure of convolutional block attention module

    Download full size
    Architecture of multi-branch all convolutional neural network

    Fig. 2. Architecture of multi-branch all convolutional neural network

    Download full size
    Noise extraction module

    Fig. 3. Noise extraction module

    Download full size
    Subjective visual comparison of synthetic low-light images without noise. (a) Image of parrots; (b) image of building2; (c) image of buildings; (d) image of monarch

    Fig. 4. Subjective visual comparison of synthetic low-light images without noise. (a) Image of parrots; (b) image of building2; (c) image of buildings; (d) image of monarch

    Download full size
    Subjective visual comparison of synthetic low-light images with noise. (a) Image of parrots; (b) image of building2; (c) image of buildings; (d) image of monarch

    Fig. 5. Subjective visual comparison of synthetic low-light images with noise. (a) Image of parrots; (b) image of building2; (c) image of buildings; (d) image of monarch

    Download full size
    Subjective visual comparison of real low-light images. (a)(b) Images from LIME dataset; (c)(d) images from DICM dataset; (e) image from MEF dataset

    Fig. 6. Subjective visual comparison of real low-light images. (a)(b) Images from LIME dataset; (c)(d) images from DICM dataset; (e) image from MEF dataset

    Download full size

    • Table 1. Comparison of objective evaluation indicators for synthetic low-light images without noise

      View table

      Table 1. Comparison of objective evaluation indicators for synthetic low-light images without noise

      ImagePSNR/dBMSEMAEMS-SSIMQVIF
      Originalimage10.036/9.4606447.9000/7440.800074.8505/76.81000.3925/0.34000.2407/0.24000.4889/0.5100
      Image ofCLAHE[3]15.596/15.8501792.4000/1817.000039.3139/36.46000.7988/0.78150.6357/0.69000.7582/0.7500
      Image ofSSR[6]22.389/20.560375.1150/658.800014.3500/19.17000.8247/0.78000.6489/0.68000.8437/0.7400
      Image ofMSRCR[8]11.135/11.4105006.9000/4775.500067.8155/62.91000.7602/0.69000.6030/0.59000.6348/0.5800
      Image of methodin Ref. [12]19.040/17.430811.0260/1220.000022.8847/28.73000.8001/0.74000.6243/0.63000.6229/0.4200
      Image of methodin Ref. [30]19.959/18.250656.4077/1028.000018.3382/22.66000.8439/0.78000.6522/0.65000.6801/0.6700
      Image of methodin Ref. [16]22.679/21.860350.8300/434.670016.5200/17.30000.9130/0.86500.7360/0.76400.8750/0.8510
      Image ofMBACNN23.869/22.550266.7760/384.200013.8820/15.50000.9229/0.87000.7735/0.77000.8834/0.8630

    • Table 2. Comparison of objective evaluation indicators for synthetic low-light images with noise

      View table

      Table 2. Comparison of objective evaluation indicators for synthetic low-light images with noise

      ImagePSNR/dBMSEMAEMS-SSIMQVIF
      Originalimage9.8380/8.70006749.6000/8878.000074.8583/84.90000.3050/0.22000.1364/0.08000.3249/0.3400
      Image ofCLAHE[3]13.4239/11.96002955.9000/4215.000049.0863/56.70000.5606/0.41000.3990/0.27000.5170/0.5100
      Image ofSSR[6]18.9663/18.3000825.0010/979.000021.3554/23.60000.6690/0.51000.5018/0.40000.6058/0.5800
      Image ofMSRCR[8]10.4440/17.04005870.6000/1344.000072.3333/28.90000.5592/0.50000.3809/0.37500.5436/0.6280
      Image of methodin Ref. [12]19.5938/16.2000714.0061/1600.000019.2488/33.70000.6630/0.48000.4639/0.41000.5216/0.4200
      Image of methodin Ref. [30]17.8858/16.50001058.0000/1535.700022.6885/30.30000.6145/0.54000.3896/0.32000.4593/0.4700
      Image of methodin Ref. [16]18.3000/19.3300961.3000/776.550025.1700/21.67000.6900/0.62000.5300/0.48000.7370/0.7190
      Image ofMBACNN21.1500/19.8200499.0299/697.300016.3177/20.24000.7970/0.68000.6251/0.52000.6903/0.7400

    • Table 3. Comparison of objective evaluation indicators for real low-light images

      View table

      Table 3. Comparison of objective evaluation indicators for real low-light images

      ImageNRSSEntropy of informationNIQE
      Image of CLAHE[3]0.8862/0.95007.6180/7.30003.9221/5.2000
      Image of SSR[6]0.9257/0.94007.6845/7.36004.2682/5.0100
      Image of MSRCR[8]0.8946/0.93007.3285/7.53003.9577/4.2600
      Image of method in Ref. [12]0.9159/0.91907.7841/7.46004.1566/4.6100
      Image of method in Ref. [30]0.8785/0.92207.2858/7.41804.4679/4.9800
      Image of method in Ref. [16]0.9240/0.96306.6100/6.22505.1130/5.2530
      Image of MBACNN0.9359/0.97907.5844/7.49005.3669/4.6770

    • Table 4. Algorithm processing time comparison

      View table

      Table 4. Algorithm processing time comparison

      MethodTraining time /hTest time /s
      CLAHE[3]—0.77
      SSR[6]—2.65
      MSRCR[8]—1.02
      Method in Ref. [12]—0.72
      Method in Ref. [30]—1.66
      Method in Ref. [16]8.965.88
      MBACNN5.996.78
    Tools

    Get Citation

    Copy Citation Text

    Ruoyou Wu, Dexing Wang, Hongchun Yuan, Peng Gong, Guanqi Chen, Dan Wang. Low-Light Image Enhancement Based on Multi-Branch All Convolutional Neural Network[J]. Laser & Optoelectronics Progress, 2020, 57(14): 141021

    Download Citation

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

    Category: Image Processing

    Received: Oct. 15, 2019

    Accepted: Dec. 17, 2019

    Published Online: Jul. 28, 2020

    The Author Email: Dexing Wang (dxwang@shou.edu.cn), Hongchun Yuan (hcyuan@shou.edu.cn)

    DOI:10.3788/LOP57.141021

    Topics

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