Table 2.
Summary of performance metrics for Fet-Net and prominent CNN architectures during the 5-fold cross-validation experiment.
| Architecture | Average Accuracy (%) | Average Loss | Number of Parameters |
|---|---|---|---|
| Fet-Net | 97.68 | 0.06828 | 10,556,420 |
| VGG16 | 96.72 | 0.12316 | 14,847,044 |
| VGG19 | 95.83 | 0.15412 | 20,156,740 |
| ResNet-50 | 88.37 | 0.40604 | 24,113,284 |
| ResNet-50V2 | 95.20 | 0.16328 | 24,090,372 |
| ResNet-101 | 82.12 | 0.47086 | 43,183,748 |
| ResNet-101V2 | 94.69 | 0.18866 | 43,152,132 |
| ResNet-152 | 84.12 | 0.41756 | 58,896,516 |
| ResNet-152V2 | 94.61 | 0.20502 | 58,857,220 |
| Inception-ResnetV2 | 94.20 | 0.21042 | 54,731,236 |
| InceptionV3 | 93.83 | 0.21720 | 22,328,356 |
| Xception | 96.08 | 0.13956 | 21,387,052 |