Table 10 Impact of data augmentation and noise on model robustness for retinal fundus image classification using retinal fundus images dataset.
From: An enhanced diabetic retinopathy detection approach using optimized deep learning technique
Metric | Value |
|---|---|
Baseline Accuracy (%) | 94.6 |
Accuracy After Augmentation (%) | 92.1 |
Accuracy After Noise Injection (%) | 89.5 |
Robustness Factor (Augmentation) | 0.97 |
Robustness Factor (Noise) | 0.95 |
