Variational multi-task MRI reconstruction: Joint reconstruction, registration and super-resolution

Motion degradation is a central problem in Magnetic Resonance Imaging (MRI). This work addresses the of how to obtain higher quality, super-resolved motion-free reconstructions from highly undersampled MRI data. In this work, we present for first time variational multi-task framework that allows joining three relevant tasks MRI: reconstruction, registration and super-resolution. Our takes set multiple MR acquisitions corrupted by motion into novel optimisation model, which composed an L2 fidelity term sharing representation between tasks, super-resolution foundations hyperelastic deformations model biological tissue behaviors. We demonstrate combination yields significant improvements over sequential models other bi-task methods. results exhibit fine details compensate producing sharp textured images compared state art methods while keeping low CPU time. are appraised on both clinical assessment statistical analysis.