A Study On Parallel Magnetic Resonance Imaging Based On 3D Model

Magnetic resonance imaging(MRI)is one of the most important imaging methods in clinical diagnosis and medical research due to its advantages of non-ionizing radiation,high contrast and high resolution imaging.But long scanning time of the MRI machine limits its further application and promotion.Parallel magnetic resonance imaging(pMRI)uses multiple radio frequency receiving coils to receive signals and utilizes the redundancy of the coil data to sample parts of the k-space data.The pMRI technique greatly reduces the scanning time,but needs to study pMRI algorithm to reconstruct the target image.However,the pMRI problem is an undersampling system,and the noise caused by a MRI machine reduce the imaging quality.Thus it is meaningful to further research more effective pMRI techniques.Due to the redundancy of multiple coils,the correlation between the each k-space data is investigated.A 3D regularized model is proposed to solve the problems of aliasing artifacts and noise amplification from reconstruction.The main contributions are:1)each two-dimensional(2D)MRI coil images can be folded to three-dimensional(3D)data using the redundancy of the k-space data by multiple coils,and then the correlation of the3 D data can be used to reduce artifacts and noise by constraining the constructed image;2)2D coil images are decomposed by 2D wavelet transformer,and then one-dimensional Haar system is applied to 2D wavelet features to describe the relationship between the features of each coil;3)A 3D sparse optimization model based on 3D coil image model is proposed,and two nonlinear iterative 3DcGRAPPA and 3DrGRAPPA algorithms,under the framework of alternating direction multiplier method(ADMM),are proposed to estimate interpolating kernels and retrieve images for solving non-smoothness of the reconstructed model,respectively.Experiments shows that the proposed 3D sparse optimization model can accurately get calibration kernel weights,effectively suppress the artifacts and noise,and reconstruct high quality and high contrast medical image.