Fig. 6From: 3D convolutional neural networks applied to CT angiography in the detection of acute ischemic strokeASPECTS points of the manual segmentation and three CNN outputs for the group A patients (test data). All the scores are calculated from the CTA-SI; if a single positive voxel belongs to the specified region, it is marked as stroke positive. The number of positive regions is subtracted from 10 to get the ASPECTS points and the minimum (worst case) between the hemispheres is the reported valueBack to article page