Table 6.
AUC, accuracy, sensitivity and specificity related to the bright-lesion detection on DIARETDB1 database. Results were optimized to the best sensitivity-specificity trade-off.
Accuracy | Sensitivity | Specificity | ||
---|---|---|---|---|
Random Forests | 0.8037 ± 0.0319 | 0.8027 ± 0.0324 | 0.8038 ± 0.0319 | 0.9006 ± 0.0146 |
Linear-SVM | 0.8112 ± 0.0316 | 0.8107 ± 0.0315 | 0.8112 ± 0.0316 | 0.6207 ± 0.0440 |
RBF-SVM | 0.8166 ± 0.0322 | 0.8154 ± 0.0321 | 0.8155 ± 0.0322 | 0.6334 ± 0.0401 |
Gaussian Processes | 0.8184 ± 0.0324 | 0.8183 ± 0.0324 | 0.8184 ± 0.0324 | 0.6023 ± 0.0518 |