An Automated Tensorial Classification Procedure for Left Ventricular Hypertrophic Cardiomyopathy

Cardiovascular diseases are the leading cause of death globally. Therefore, classi cation tools play a major role in prevention and treatment of these diseases. Statistical learning theory applied to magnetic resonance imaging has led to the diagnosis of a variety of cardiomyopathies states. We propose a two-stage classi cation scheme capable of distinguishing between heterogeneous groups of hypertrophic cardiomyopathies and healthy patients. A multimodal processing pipeline is employed to estimate robust tensorial descriptors of myocardial mechanical properties for both short-axis and long-axis magnetic resonance tagged images using the least absolute deviation method. A homomorphic ltering procedure is used to align the cine segmentations to the tagged sequence and provides 3D tensor information in meaningful areas. Results have shown that the proposed pipeline provides tensorial measurements on which classi ers for the study of hypertrophic cardiomyopathies can be built with acceptable performance even for reduced samples sets.