Image Reconstruction Using Partial K-Space with Compressed Sensing-Sense and Sparse Blip
MetadataShow full item record
Magnetic Resonance Imaging is one of the most important imaging technologies these days. But one problem of MRI remains now — the low imaging speed. At present, there are many methods existing to deal with different kind of reconstructed problems. Compressed sensing (CS), parallel imaging and partial Fourier (PF) acquisition are all effective methods to reduce k-space sampling and therefore accelerate MR acquisition and reconstruction. In this thesis, I will introduce the method that combines these existing method to achieve higher reduction and faster reconstruction. We performed simulations on full k-space data to demonstrate the potential use of combined PF-CSSENSE or PF-SparseBLIP in MRI. CS-SENSE combined with PF sampling is able to provide better reconstructed images than CS-SENSE without PF for the same total acceleration. SparseBLIP combined with PF sampling also provides comparable result than original SparseBLIP method. Combining PF sampling with existing CS and SENSE offers us multiple choices to find higher reduction factors in MRI reconstruction or improve image quality with keeping reduction factor unchanged.