Inter-k-space motion based strategy for Super-Resolution in MRI

Magnetic Resonance Imaging (MRI) image reconstruction based on a frequency domain Super-Resolution (SR) algorithm, is presented in the paper. It is shown that the approach improves MRI spatial resolution in cases when Periodically Rotated Overlapping Parallel Lines with Enhanced Reconstruction (PROPELLER) sequences are used. The PROPELLER MRI method collects data in rectangular `blades' rotated around the origin of the k-space. Inter-blade patient motion is the premise for the use of SR technique. Images obtained from sets of irregularly located frequency domain samples are combined into the high resolution MRI image. The SR reconstruction replaces the usually applied direct averaging of low-resolution images. We extended iterative back projection approach to consider also tissue movement. Our approach has been based on simple premise, that LR blade images may be segmented into areas with coherent motion trajectory. Appropriately processed sub-images are combined into one Super-Resolution image by a modified IBP scheme. Our algorithm covers a cardiac and respiratory movements. The improvements resulted in lower approximation error and higher convergence speed.