Lumped-parameter models of the cardiovascular system can improve the understanding of cardiovascular function and assist treatment planning. The clinical applicability of these models improves when they are subject-specific. This work proposes an approach to personalizing a model of the heart and the systemic circulation using exclusively non-invasive measurements from routine cardiovascular MRI and 4D Flow MRI. Personalized models were constructed for eight healthy volunteers. The model-based pressures and flows agreed well with the in-vivo measurements for each subject. The proposed approach can be used to synthesize medical data into clinically relevant information and estimate parameters that cannot be measured clinically.