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. 59, Issue 6, 0617017(2022)

Automatic Detection of Retinal Diseases Based on Lightweight Convolutional Neural Network

Lingxiao Wang1, Jun Yang1, Wensai Wang1, and Ting Li1,2、*
Author Affiliations
  • 1Institute of Biomedical Engineering, Chinese Academy of Medical Sciences, Tianjin 300192, China
  • 2Chinese Institute for Brain Research, Beijing, Beijing 102206, China
    • show less
    AbstractGet PDF(in Chinese)
    Figures&Tables (8)
    Equations (3)
    References (23)
    Tools
    Paper Information
    Topics
    Figures & Tables(8)
    Sample image examples. (a) CNV; (b) DME; (c) DRUSEN; (d) NORMAL

    Fig. 1. Sample image examples. (a) CNV; (b) DME; (c) DRUSEN; (d) NORMAL

    Download full size
    Schematic diagrams of module structure . (a) Dilated SepConv ; (b) Residual Unit

    Fig. 2. Schematic diagrams of module structure . (a) Dilated SepConv ; (b) Residual Unit

    Download full size
    Schematic diagram of proposed model structure

    Fig. 3. Schematic diagram of proposed model structure

    Download full size
    ROC curves of proposed model. (a) ROC curve; (b) partial enlarged view

    Fig. 4. ROC curves of proposed model. (a) ROC curve; (b) partial enlarged view

    Download full size
    CAM. (a) CNV; (b) DME; (c) DRUSEN

    Fig. 5. CAM. (a) CNV; (b) DME; (c) DRUSEN

    Download full size

    • Table 1. Segmentation of retinal OCT image dataset

      View table

      Table 1. Segmentation of retinal OCT image dataset

      CategoryTraining datasetValidation datasetTesting datasetTotal
      CNV33485372025037205
      DME10214113425011348
      DRUSEN77558612508616
      NORMAL46026511425051140
      Total97480108291000109309

    • Table 2. Classification results of proposed model for various retinal images

      View table

      Table 2. Classification results of proposed model for various retinal images

      CategoryAccuracySensitivitySpecificity
      CNV0.9661.0000.947
      DME0.9860.9441.000
      DRUSEN0.9680.8721.000
      NORMAL1.0001.0001.000

    • Table 3. Overall evaluation results of different classification models

      View table

      Table 3. Overall evaluation results of different classification models

      ModelOv-AccOv-SeOv-SpLayersParameter Count /MBRatio
      Human Expert［12］0.9210.9940.943
      VGG 16［14］0.9860.97816138.38.6
      RongheNet［23］0.9720.970654.59.1
      Inception V3［12］0.9660.9780.9744223.60.6
      Proposed model0.9800.9540.987449.20.2
    Tools

    Get Citation

    Copy Citation Text

    Lingxiao Wang, Jun Yang, Wensai Wang, Ting Li. Automatic Detection of Retinal Diseases Based on Lightweight Convolutional Neural Network[J]. Laser & Optoelectronics Progress, 2022, 59(6): 0617017

    Download Citation

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

    Category: Medical Optics and Biotechnology

    Received: Jan. 13, 2022

    Accepted: Feb. 10, 2022

    Published Online: Mar. 8, 2022

    The Author Email: Ting Li (liting@bme.cams.cn)

    DOI:10.3788/LOP202259.0617017

    Topics

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