Advanced PDE-based methods for automatic quantification of cardiac function and scar from magnetic resonance imaging

This chapter focuses on the application of geometric deformable models based on partial differential equations (PDEs) for cardiac magnetic resonance imaging data processing. In particular we present two possible segmentation approaches: the basic level set model and a “region-based” level set model, we applied for left and right ventricular chamber segmentation and extraction of volumes and derived functional parameters. The application of these shape independent models directly in the three dimensional domain to data acquired with a 3D imaging system could potentially achieve the clinical need for correct and complete interpretation of ventricular morphology and pathology and for fast quantification of cardiac chamber volumes, ventricular function and myocardial scar in various situations. The two models have been detailed described in the chapter and the results obtained applying them to cardiac magnetic resonance data are also presented.