Fig. 2From: Fully automated convolutional neural network-based affine algorithm improves liver registration and lesion co-localization on hepatobiliary phase T1-weighted MR imagesScheme of the two-step fully automated affine registration algorithm using intensity masks. First, an independently developed two-dimensional liver segmentation algorithm was used to extract liver masks populated with intensities. Intensity masks were registered using an affine transformation network to geometrically align the moving series (follow-up) to the static series (baseline). Optimal affine transformation parameters were determined by maximizing the similarity between baseline and registered follow-up. Affine transformation parameters were then used to map the entire moving series to the static image series spaceBack to article page