Table 2 Test results of the CNN-based segmentation model’s performances combined with different preprocessing methods
| Â | Training set | Test set | ||||
|---|---|---|---|---|---|---|
Accuracy | F1 score | Mean squared error | Accuracy | F1 score | Mean squared error | |
Raw image | 0.9794 | 0.997 | 0.0019 | 0.9742 | 0.9935 | 0.0062 |
Stretching | 0.9801 | 0.998 | 0.00098 | 0.974 | 0.9932 | 0.007 |
Equalization | 0.982 | 0.998 | 0.00074 | 0.9756 | 0.9946 | 0.0049 |