Table 5 Precision, Recall, and F1 scores of subtype multiclass classification task using CNN and Pyradiomics. Models were trained with cGAN MRIs, and a combined dataset of cGAN and real patient MRIs. 20 percent testing sets were used for evaluation
Model (Dataset) | Precision | Recall | F1 |
|---|---|---|---|
CNN (cGAN MRI) | 0.7988 | 0.7917 | 0.7695 |
CNN (cGAN cGAN + real MRI) | 0.8444 | 0.8435 | 0.8336 |
PyRadiomics (cGAN MRI) PyRadiomics (cGAN + real MRI) | 0.4879 0.5001 | 0.5507 0.5732 | 0.4386 0.5344 |