Research On Parallel MRI SENSE Algorithm And Motion Artifact Elimination

The Magnetic Resonance Imaging(MRI)technology obtains the images about various organizations of human beings based on the characteristic of spin motion of the atomic nucleus.MRI has become one of the most widely used tools in clinical diagnosis as its high resolution,more imaging parameters and no ionizing radiation damage.However,owing to the longer imaging time,the emergence of patients' independent and non-independent movements will result in the motion artifacts and NSR reduction of reconstructed images.Parallel Magnetic Resonance Imaging(pMRI)breaks through the limitations of hardware condition and magnetic properties which caused too long imaging time,it mainly uses multiple phased-array coils to receive the sensor signals simultaneously,then encodes the space signal based on different spatial sensitivity of these coils,and the code frequency in the direction of phase encoding gradient is gradually decreased,so that the imaging speed is improved by shortening the scanning time.But in the application of MRI SENSE algorithm,the reconstructed results are not robust.In order to improve the accuracy of reconstructed algorithm,this thesis makes a deeper study of the SENSE algorithm.At the same time,the translational motion artifacts of patients are extremely difficult to completely overcome.The superimposed artifacts will make an influence on determining the imaging results,and the post-processing algorithm of translational motion artifacts can effectively improve the artifacts.Then this thesis also makes a deeper research on the Fourier projection algorithm.The parallel MRI SENSE reconstruction algorithm is considered to be the most optimal algorithm based on the image domain,and it can make full use of the space sensitivity information of each phased-array coil to encode.However,the optimizing section of SENSE algorithm uses the Least Squares(LS)algorithm,which leads to the non-robustness of reconstruction results under the influence of some residual data.Therefore,an improved SENSE reconstruction algorithm combined with robust estimation is proposed.This algorithm firstly combines with Deterministic Annealing technology to improve the M estimation to obtain the optimal objective function,and then the application of AM estimation to SENSE algorithm can obtain robust reconstruction results.Compared with the SENSE algorithm,simulation results show that the improved algorithm can improve the accuracy of the reconstructed image and the SNR.Nowadays,through a variety of research,the traditional Fourier projection algorithm has become one of the most widely used translational motion artifacts elimination methods.But this algorithm does not take the conditions of large offsets and subpixel offsets into account.To solve these problems,this thesis proposes an improved translational motion artifact correction method based on the Fourier projection algorithm and Genetic Algorithm(GA).This algorithm makes a frequency domain expansion of Fourier projection,then the large offsets in the X direction will be corrected.Then according to the GA,the subpixel offsets in the X direction and offsets in the Y direction will be corrected.When compared with the conventional Fourier projection algorithm,the simulation results prove that the proposed method can correct translational motion artifacts effectively and improve the quality of the reconstructed image.