New NeuroImage paper on joint super-resolution and synthesis
May 26, 2021
Our new article "Joint super-resolution and synthesis of 1 mm isotropic MP-RAGE volumes from clinical MRI exams with scans of di erent orientation, resolution and contrast" has been accepted for publication in NeuroImage. In the paper, we present a method to turn one of more clinical scans from a subject into a 1mm isotropic MP-RAGE volume that we can use for analyses like registration, segmentation, etc. with standard packages like FreeSurfer. Crucially, the method relies on a neural network that is trained with synthetic data, such that one can train a network for any set of orientations / resolutions / contrasts without having to collect any training data. The code is publicly available on https://github.com/BBillot/SynthSR/tree/main/SynthSR, and the paper can be found [here].
Laboratory for Computational Neuroimaging Athinoula A. Martinos Center for Biomedical Imaging 149 Thirteenth Street Charlestown, Massachusetts 02129