Table 3.
SVM Classifiers Classification Accuracies achieved with incorporated and Reduced features of three CNNs learned with Radiomics images against CT and X-ray images.
Model | LDA |
Linear-SVM |
Quadratic-SVM |
ESD |
||||
---|---|---|---|---|---|---|---|---|
Dataset I | ||||||||
CT | Fused | CT | Fused | CT | Fused | CT | Fused | |
Dark-53 | 96.5 | 97.4 | 94.8 | 95.3 | 96.9 | 97.8 | 95.7 | 97.1 |
MobileNet | 96.9 | 98.5 | 96.8 | 98.3 | 98.1 | 99.4 | 96.6 | 97.9 |
DenseNet-201 |
96.9 |
98.5 |
96.1 |
98.6 |
97.9 |
99.4 |
95.8 |
98.2 |
Model |
Dataset II | |||||||
X-Ray |
Fused |
X-Ray |
Fused |
X-Ray |
Fused |
X-Ray |
Fused |
|
Dark-53 | 96.7 | 97.3 | 96.4 | 96.7 | 96.6 | 97.2 | 96.6 | 97.5 |
MobileNet | 96.8 | 97.5 | 96.5 | 97.7 | 96.8 | 98.0 | 96.6 | 97.6 |
DenseNet-201 | 98.2 | 98.7 | 98.1 | 99.0 | 98.3 | 99.0 | 98.2 | 98.9 |