Table 3 Performance metrics evaluated from the pre-trained models used for classification of Plus form on Validation set.
From: Automated diagnosis of plus form and early stages of ROP using deep learning models
Model | Accuracy | F1 score | Sensitivity | Specificity |
|---|---|---|---|---|
ResNext50 | 0.97 | 0.86 | 0.79 | 0.9942 |
ResNet152 | 0.98 | 0.89 | 0.90 | 0.986 |
RegNetX3 | 0.95 | 0.68 | 0.52 | 0.9977 |
VGG16 | 0.95 | 0.71 | 0.59 | 0.9907 |
GoogleNet | 0.97 | 0.80 | 0.68 | 0.9977 |
MobileNetV2 | 0.98 | 0.88 | 0.89 | 0.9849 |
EfficientNetB4 | 0.996 | 0.98 | 0.99 | 0.997 |
