Skip to main content
Fig. 1 | Journal of Translational Medicine

Fig. 1

From: Radiomics-based machine learning analysis and characterization of breast lesions with multiparametric diffusion-weighted MR

Fig. 1

Workflow of image processing. a MRI data of multi-b value sequences and quantitative maps from ME, BE, SE and DKI models. b 3D segmentations of lesions shown as surface shaded 3D renderings. c Extraction of radiomics features, i.e., First-order, Shape, GLCM, GLSZM and GLDM. d Radiomics analysis using four models (RF, SVM, PCA-LR, and L1R-LR), and e ROC curve analysis. ROC curves are used for the comparison of four methods, and diagnostic performance of radiomics features and mean diffusion metrics

Back to article page