Skip to main content
Figure 1 | Journal of Cardiovascular Magnetic Resonance

Figure 1

From: Steps towards automated image segmentation as part of a 3D printing pipeline in congenital heart disease

Figure 1

In Stage 1 of the pipeline, the favourable contrast properties of the gadolinium enhanced MRA are used to generate an initial estimate (red geometry) of the cardiovascular anatomy and guide an automated segmentation of the intracardiac structures (green geometry) in the morphology data. A hybrid level set procedure, reliant on the parameters alpha, beta, mu and grad is used. In Stage 2, Relevant extracardiac vasculature (orange and light blue geometries) are automatically generated. In Stage 3 the automated segmentation geometries (intracardiac, full and reduced extracardiac masks) are then manually combined using using Boolean and region growing operations. Finally, the slice editing and region growing tools provided by Mimics 18.0 (Materialise NV, Leuven, Belgium, 1992 - 2015) are used to make fianl adjustments to the final segmentation (dark blue geometry).

Back to article page