Table 3 Performance of proposed Bayesian DenseNet-121 CNN models for DR Classification.
From: Uncertainty-aware diabetic retinopathy detection using deep learning enhanced by Bayesian approaches
Model | Train accuracy (%) | Train loss | Test accuracy (%) | Test loss | Precision (%) | Recall (%) | F1-Score (%) |
|---|---|---|---|---|---|---|---|
BCNN-MC Dropout | 98.03 | 0.018 | 97.68 | 0.08 | 0.97 | 0.97 | 0.97 |
BCNN-MFVI | 95.70 | 0.074 | 94.23 | 0.086 | 0.94 | 0.93 | 0.94 |
BCNN-Deterministic | 93.16 | 0.048 | 91.44 | 0.10 | 0.92 | 0.91 | 0.91 |
