Review Article
A Review on Recent Developments for Detection of Diabetic Retinopathy
Table 3
CAD methods for diagnosis of DR.
| Algorithm | Image processing techniques | Database | Color space | Sensitivity | Specificity | Accuracy/AUC | Lesions detection |
| Aptel et al. [21] | Single-field nonmydriatic; single-field mydriatic; three-field nonmydriatic; three-field mydriatic | 79 patients (158 eyes) | Gray scale | 77%, 90%, 92%, 97% | 99%, 98%, 97%, 98% | 0.82, 0.90, 0.90, 0.95 | Both NPDR and PDR |
| Kahai et al. [22] | Decision support system (DSS) | 143 images Louisiana State University Eye Center | Gray scale | 100% | 67% | — | NPDR |
| Usher et al. [23] | Neural network | 1273 consecutive patients, St. Thomas Hospital | Gray scale | 94.8% | — | — | NPDR |
| Gardner et al. [24] | Back propagation neural network | 200 diabetic and 101 normal images of private hospital | Gray scale | 88.4% | 83.5% | — | Both NPDR and PDR |
| Reza and Eswaran [25] | Rule based classifier | STARE | Green channel | 97.2% | 100% | 97% | NPDR |
| Annie Grace Vimala and Kajamohideen [26] | Cost-effective computer-aided diagnostic system | Private eye Hospital | HSV | 91.6% | 90.5% | 91.2% | NPDR |
| Dupas et al. [27] | Automated fundus photograph analysis algorithms | Messidor | Gray scale | 83.9% | 72.7% | — | NPDR |
| Ashraf et al. [77] | Local Binary Pattern, SVM | DIARETB1 | Green channel | 87.48% | 85.99% | 86.15%/0.87 | NPDR |
|
|