Fig. 2
From: Classification prediction of pancreatic cystic neoplasms based on radiomics deep learning models
A: Violin plots of the feature range of the 3 most important deep learning features evaluated in construction of SCA differential diagnostic model; B: ROC curves of five-fold cross-validation with the deep learning radiomics model for the SCA differential diagnosis; C: ROC curves of the radiomics model, deep learning radiomics model and fused model for the SCA differential diagnosis; D: Predictions of the radiomics model, deep learning radiomics model and fused model for each patient shown as a 3D pattern; E: Nomogram of the fused model for the SCA differential diagnosis