Figure 7

Model segmentation accuracy and misclassification rate for liver parenchyma segmentation (left), portal veins segmentation (middle), and hepatic veins segmentation (right). The results are shown for the central region of the liver (upper row) and the peripheral region of the liver (bottom row). In-phase (In), water (W), and opposed-phase (Opp) constituted single-modal inputs. In-phase, water (In-W); in-phase, opposed-phase, water (In-Opp-W); in-phase, opposed-phase (In-Opp); and in-phase, opposed-phase, fat, water (In-Opp-F-W) constituted multi-modal inputs. Liver parenchyma segmentation is shown in red, portal veins segmentation in purple, hepatic veins segmentation in blue, and background segmentation in black. The colors are identical to Figs. 4 and 5. Best viewed in screen.