Table 7.
Comparison of classification performance changes between three methods.
| Method | Original segmentation | Optimally mapped segmentation | ||||||
|---|---|---|---|---|---|---|---|---|
| DES images | LE images | DES images | LE images | |||||
| AUC | Accuracy | AUC | Accuracy | AUC | Accuracy | AUC | Accuracy | |
| Proposed MLP | 0.76±0.05 | 68.3% | 0.75±0.05 | 72.1% | 0.74±0.05 | 68.5% | 0.85±0.04 | 78.4% |
| Remove SMOTE | 0.63±0.06 | 70.19% | 0.70±0.05 | 68.46% | 0.75±0.05 | 62.16% | 0.69±0.05 | 65.76% |
| Remove CFS and SMOTE | 0.56±0.07 | 65.38% | 0.71±0.05 | 66.67% | 0.61±0.06 | 64.86% | 0.59±0.06 | 63.06% |