Table 7.
The classification performance of the proposed model with CLAHE enhanced images compared to other CNN models
| CNN Models | Precision | Recall | F1-score | Trainable Parameters |
|---|---|---|---|---|
| VGG16 | 40% | 53% | 48% | 14,716,227 |
| VGG19 | 42% | 53% | 47% | 20,025,923 |
| InceptionV3 | 76% | 75% | 75% | 21,808,931 |
| Xception | 79% | 63% | 70% | 20,867,627 |
| ResNet50 | 69% | 68% | 68% | 23,593,859 |
| ResNet152 | 55% | 52% | 53% | 58,377,091 |
| ResNet50V2 | 64% | 65% | 64% | 23,570,947 |
| ResNet152V2 | 58% | 55% | 56% | 58,337,795 |
| MobileNetV2 | 61% | 62% | 61% | 2,261,827 |
| DenseNet121 | 76% | 73% | 74% | 7,040,579 |
| DenseNet169 | 79% | 69% | 74% | 12,647,875 |
| DenseNet201 | 89% | 87% | 88% | 18,327,747 |
| Proposed Model with raw images | 95% | 96% | 95% | 18,561,541 |
| Proposed Model with CLAHE enhanced images | 97% | 96% | 96% | 18,561,541 |