A cardiac disease prediction and clinical decision making framework based on automated segmentation of left and right ventricles in cardiac MRI images

In this research, a graphical computational tool for automatic segmentation of the left and right ventricles in cardiac MRI images was developed, tested, and analyzed. This research develops a tool that can aid cardiologists in the extraction of clinically relevant medical information such as ejection fraction, stroke volume, etc. from the patient's cardiac MRI images. The fully automatic tool was developed to allow the user to load any cardiac MRI image, and perform automatized segmentation upon the click of a button. This graphical GUI tool generates the segmented left and right ventricle extracted regions and their delineated boundaries upon clicks of buttons. The clinical motivation for ventricular segmentation in cardiac MRI images is to calculate the clinical parameters measured from the segmented left ventricle such as ED volume, ES Volume, SV (stroke volume) and EF (ejection fraction). These parameters are effective measures of the capability of the heart’s pumping mechanism, i.e. they provide clinical measures of how well the heart is able to pump blood out to the rest of the body. This clinical parameter estimation from cardiac MRI images is essential for cardiologists to assess cardiac function and predict cardiac diseases or disorders. . It was observed that the proposed methodology presented in this paper performed better in terms of accuracy and had the smallest mean square error (0.015) and had a high overall average correlation coefficient (0.9445). Keywords: Segmentation, Left ventricle, Right ventricle, MRI, GUI