Research On Parallel Magnetic Resonance Imaging Algorithm Based On K-space

Magnetic Resonance Imaging (MRI) is an important biomedical diagnostic technology appeared in recent century. However, the scan time of magnetic resonance imaging was so long that the patient cannot understand and produce involuntary movement which had a serious impact of the reconstructed medical image quality. In order to reduce the imaging time in clinical application, the physical methods that through improve the intensity of main magnetic and gradient field have reached the end, so it must use certain mathematical methods to enhance reconstruction speed. Therefore, it is more urgent on how to get clear images through fast reconstruction. And the parallel magnetic resonance imaging (pMRI) has been proven to be a most effective reconstruction algorithm making the imaging time shorter, this paper mainly aims at the further study for parallel magnetic resonance imaging algorithm based on k-space.Firstly, based on the comparison of various parallel magnetic resonance imaging algorithms, this paper focus on the study of uncomplete fitting problem of k-space data while reconstruct the image using partial k-space data, and put forward a more effective k-space data fitting algorithm:an improved GRAPPA data fitting method. This method considers the changes in phase directions when take the data acquisition and make the fitting coefficient for nonlinear.The simulation results show that, this method can reduces k-space data acquisition and fit out the more accurate complete k-space data even in larger acceleration factors, thus it reduce the time of data acquisition as well as make the image clearer.Because the GRAPPA algorithm used undersampling partial k-space data to fit out the full k-space data, so the acquisition of k-space data exists valid data lost, and leads to the reconstruction image having Gibbs artifact. Therefore, based on GRAPPA data acquisition, this paper focus on the study of removing reconstruction artifact after the partial k-space data reconstruction and put forward a reasonable reconstruction algorithm which can remove artifact effectively: the weighted sum of squares algorithm of parallel magnetic resonance reconstruction based on regional mutual information.This new method divides the reconstruction image into small regions, and gets the image block gray information, then compute the image mutual information through information entropy, And then give different images different weight coefficient in the image reconstruction process. This method can avoid the deficiency of commonly used sum of squares reconstruction algorithm for the same weight, avoid image error for different coil data damage in the actual acquisition.The experiments provided evidence that this method can effectively resolve the problem of the imprecise pixel of original method greatly influence the reconstructed quality, it can effectively remove artifact, and improve the precision of the image reconstruction.