Figure 3:

Receiver Operating Characteristic (ROC) curves generated from the classifications of the Convolutional Neural Network (CNN). (a) depicts the ROC curves obtained for each sub-group when making a 2-class classification, (b) depicts the ROC curves for each sub-group for the mTICI 0,1,2a class when making a 3-class classification, (c) depicts the ROC curves for each sub-group for the mTICI 2b class when making a 3-class classification, and (d) depicts the ROC curves for each sub-group for the mTICI 2c,3 class when making a 3-class classification. The shaded region around the ROC curve depicts the standard deviations at each point. High AUROC values and thin spread of standard deviations in (a) indicates the best performance is achieved when making a 2-class classification. In all 4 sub-plots, while the ROC curves and the standard deviations overlap between the three sub-groups, the McNemar’s p-test values indicate significant improvement in performance when ensembled networks over AP and lateral view networks.