Table 11.
The training execution time (sec) of the three CNN used in MB-AI-His CADx and after the fusion stage of MB-AI-His CADx for both binary and multi-class classification levels.
Multi-Class Classification Level | |||||||
---|---|---|---|---|---|---|---|
DL Method | Time | Feature Reduction Method | Linear-SVM Time | Cubic-SVM Time | k-NN Time | LDA Time | ESD Time |
ResNet-50 | 249 | DCT | 2.15 | 2.03 | 1.98 | 2 | 3.39 |
MobileNet | 125 | PCA | 2.74 | 3.17 | 4.75 | 2.79 | 6.23 |
DenseNet-201 | 850 | ||||||
Binary-Class Classification Level | |||||||
DL Method | Time | Feature Reduction Method | Linear-SVM Time | Cubic-SVM Time | k-NN Time | LDA Time | ESD Time |
ResNet-50 | 125 | DCT | 0.873 | 0.886 | 0.858 | 0.888 | 2.54 |
MobileNet | 132 | PCA | 2.05 | 2 | 2.07 | 1.996 | 3.314 |
DenseNet-201 | 540 |