Cardiac Cohort Classification based on Morphologic and Hemodynamic Parameters extracted from 4D PC-MRI Data.

摘要: An accurate assessment of the cardiovascular system and prediction diseases (CVDs) are crucial. Measured cardiac blood flow data provide insights about patient-specific hemodynamics, where many specialized techniques have been developed for visual exploration such sets to better understand influence morphological hemodynamic conditions on CVDs. However, there is a lack machine learning approaches that allow feature-based classification heart-healthy people patients with In this work, we investigate potential characteristics, extracted from measured in aorta, volunteers bicuspid aortic valve (BAV). Furthermore, research if characteristic features classify male female as well older BAV patients. We propose analysis pipeline status, encompassing feature selection, model training hyperparameter tuning. our experiments, use several selection methods algorithms train separate models healthy subgroups report performance predictive power regard defined groups. Finally, identify key best models.