TABLE V.
Classification Accuracy of Deep Learning Frameworks
| Datasets | Type | Acc, % | Sen, % | Spe, % |
|---|---|---|---|---|
| (1) EyePACS-u | Single | 92.1 | 86.5 | 96.3 |
| (2) Messidor | Single | 99.1 | 98.3 | 100 |
| (3) EyeP_Mess | Cross | 69.7 | 34.5 | 97.4 |
| (4) Mess_EyeP | Cross | 81.5 | 57.7 | 99.3 |
| (5) Mess_EyeP | Merged | 94.6 | 88.5 | 98.3 |
| (6) rDR | Merged | 91.2 | 92.2 | 91.2 |
| (7) vtDR | Merged | 98.6 | 98.2 | 99.1 |