Super-resolution of Magnetic Resonance Images (MRI) acquired under clinical protocols

Collaborating institution: School of Informatics, The University of Edinburgh (UK)

About the project

Vast amounts of Magnetic Resonance Images (MRI) are routinely acquired in clinical practice but, to speed up acquisition, these scans are typical of a quality that is sufficient for clinical diagnosis but sub-optimal for precision medicine and large-scale scientific research. This project aims to develop a state-of-the-art machine learning scheme that using a dataset of paired low- and high-resolution brain MRI scans (i.e., acquired in clinical and research settings respectively) for training, effectively up-samples clinical (low-resolution) scans. Preliminary evaluations from an academic project show that the up-sampled MRI are of high quality and qualitatively faithful to the ground-truth high-quality research scans. We also aim to highlight further directions and make the code freely available.

 

Awarded grants

Ongoing collaboration in form of academic projects due to synergies in research interests.

 

Publications

Manuscript in preparation

Dr Maria Valdes Hernandez

 

Key contact

Please, get in touch with Dr Maria Valdes-Hernandez for more information about this project and further collaboration.

 M.Valdes-Hernan@ed.ac.uk

Dr Maria Valdes-Hernandez research profile

This article was published on