Research Article
A Comprehensive Convolutional Neural Network Survey to Detect Glaucoma Disease
Table 3
Review of glaucoma diagnosis using CNN.
| Year | Algorithm | Feature | Method | Dataset | Accuracy |
| 2016 | Deep CNN | Optic disc segmentation | CNN using 5 layers with VGG-19 network | DRIONS, RimOneR3 | F-score: 96 | 2017 | CNN | Optic cup and disc segmentation | Sampling methodology | Drishti-GS1 | Accuracy: 94.10% | 2017 | CNN | Disc segmentation | CNN with 7 layers | Drive | Accuracy: -Sensitivity: 87 | 2017 | Ant colony optimization | Optic cup segmentation | Ant colony optimization | RimoneR3 | Accuracy: -AUC: 0.7957 | 2017 | DCNN | Optic cup and disc segmentation | U-Net convolutional neural network | Drishti-GS | 2017 | 2018 | CNN | Optic disc segmentation | Two-branch CNN | Private | Accuracy: 81.69 | 2018 | Ensemble learning CNN | Optic disc segmentation | CNN using CDR ratio | Rimone | Accuracy: 81.69 | 2018 | DCNN | Fundus photography | Cross Entropy Inception V3 | Private (Kin’s Eye Hospital, South Korea | Accuracy: 87.9 | 2018 | CNN | ONH fundus images | GON in fundus photographs | Private (the ADAGES new study and Alabama, California) | Accuracy: -AUC: 0.91 | 2018 | DCNN | Optic cup and disc segmentation | Morphometric features for CNN multistage model | Drishti-GS1 Rimone | Accuracy: 88.9% | 2018 | CNN | Optic disc segmentation | Entropy sampling, Ensemble learning | Drishti-GS | Accuracy: 91.9% | 2019 | Joint RCNN | Optic cup and disc segmentation | Generative adversarial network GL-Net | Drishti-GS1 Origa | Accuracy: -AUC: 90.10 | 2019 | CNN | Proposed deep CNN | CNN with 18 layers | Private Kasturba medical college India | Accuracy: 98% | 2020 | CNN | Optic cup and disc segmentation | CNN OverFeat and VGG-S | Private | Accuracy: 90.50% | 2020 | DCNN | Optic Cup and Disc segmentation | Joint RCNN | Origa | Accuracy: 92.50% |
|
|