Publicly Available CFD-enhanced Cerebrovascular 4D Flow MRI Dataset
Please login to view abstract download link
In this work we propose to generate publicly available paired sets of patient-specific realistic ground truth flow and cerebral 4D flow MRI data to investigate the impact of MR parameters including spatial resolution, undersampling, motion, and encoding strength on the quality of reconstructed flow fields. The results allowed us to better understand fundamental limitations of cerebral 4D flow MR and to accurately quantify errors in the reconstructed flow fields. Additionally, the augmentation of available in-vivo datasets with synthetic data paired with ground truth flow can be used to increase the confidence on validation of deep learning applications such as reconstruction, hemodynamic parameter inference, and super-resolution.
