Skip to main content
. 2023 Aug 23;13:13774. doi: 10.1038/s41598-023-40599-3

Figure 2.

Figure 2

Deep learning framework. CT images were cropped to the region of interest, after which a 3D U-Net network was trained for automated tibia and fibula segmentation. After that, the segmentation masks were converted into 3D bony models. Created with Mimics® (version 21.0, Materialise, Leuven, Belgium, https://www.materialise.com/en/healthcare/mimics-innovation-suite/mimics).