Fig. 4From: Fully automated convolutional neural network-based affine algorithm improves liver registration and lesion co-localization on hepatobiliary phase T1-weighted MR imagesBaseline static (a) and follow-up moving images (b, c). In b, the follow-up image registered to baseline manually using the bifurcation of the right portal vein as anatomical reference: differences in liver morphology are pronounced in the posterior aspect of the liver (circle) and in the left lobe (circle). In c, the follow-up image registered to baseline using the automated affine algorithm shows better correspondence to baseline imageBack to article page