Using 3D realistic blood vessel structures machine learning for MR vascular Fingerprinting

International audience; MR vascular ngerprinting aims at mapping cerebral vascular properties such as blood volume and oxygenation. We to improve the technique by generating dictionaries based 3D vascular networks segmented from whole brain high-resolution (3 µm isotropic) microscopy In order to compensate for the limited number of available data and long times, we used a machine-learning reconstruction process generalize our results and tested our approach in healthy, stroke and tumor animal models. results show high quality maps with expected contrast and baseline values in healthy animals as well expected trends in pathological tissues.