Variational methods and partial differential equations in cardiac image analysis

Cardio-vascular diseases is a major cause of deaths worldwide. Early diagnosis is quite often associated with more efficient treatment leading to a significant decrease of the mortality rate. In this paper, we present tools developed to process different cardiac modalities over the past years with aim to facilitate/automate computer aided diagnosis. In particular we present algorithms and advanced techniques for the segmentation of the left ventricle for magnetic resonance and echocardiographic imaging. Furthermore, to account for user interaction, we extent our method towards user-aided boundary delineation where the user has the ability to correct local discrepancies of automated results.