Table 6.
Performance comparison for the proposed C-DenseNet against the classical DenseNet, and ResNet.
Network | Accuracy | Precision | Recall | F1-score |
---|---|---|---|---|
ResNet | 0.7343 | 0.7272 | 0.7276 | 0.7197 |
DenseNet | 0.9253 | 0.9254 | 0.9253 | 0.9248 |
C-DenseNet | 0.9790 | 0.9799 | 0.9799 | 0.9799 |
Each result is reported in the form of mean deviation. The ones in bold denote the best performance.