Table 1 Accuracy and AUC from the test fold for different types of preprocessing methods in both networks
From: Fully convolutional architecture vs sliding-window CNN for corneal endothelium cell segmentation
| Method | Accuracy | AUC |
|---|---|---|
| SW-net | ||
| Raw | 95.45 | 0.9932 |
| Normalize | 95.49 | 0.9933 |
| Standardize | 95.54 | 0.9937 |
| CLAHE | 95.82 | 0.9938 |
| CLAHE+Standardize | 95.88 | 0.9935 |
| U-net | ||
| Raw | 97.65 | 0.9958 |
| Normalize | 97.67 | 0.9954 |
| Standardize | 97.64 | 0.9956 |
| CLAHE | 97.63 | 0.9957 |
| CLAHE+Standardize | 97.65 | 0.9953 |