Research And Application Of Rapid Magnetic Resonance Imaging Technology Based On ROI-CS

Magnetic resonance imaging(MRI)is widely used in clinical medical diagnosis by virtue of its advantages of no radiation,arbitrary tomography scanning and excellent soft tissue contrast.However,the slow data acquisition speed has been restricting the development of magnetic resonance imaging technology.The new Compressed Sensing(CS)sampling technique can be applied to MRI by randomly collecting a small amount of k space data to improve the speed of data acquisition.A high quality image can be constructed using a nonlinear iterative reconstruction algorithm.The conventional CS based MRI fast imaging is usually based on a nonlinear reconstruction for the global image,but in the diagnosis of some diseases such as tumors,more attention is often pay on region of interest area.This paper focuses on the research of compressed sensing technology based on Region of Interest(ROI),a new algorithm about continuity constraint based on ROI in image domain is proposed for CS reconstruction.Proposed ROI-CS technology is applied in fast vascular imaging and real-time dynamic MRI.The main work and innovation of this paper includes:(1)Study the fast magnetic resonance imaging based on compressed sensing.Study the basic theory of compressed sensing and its application in fast MRI.The compressed sensing theory mainly includes three parts: sparse representation of signal,design of observation matrix,and nonlinear reconstruction.Under different under-sampling ratios,wavelet transform and TV transform were used as sparse basis,and guarantee the fidelity of image data,high quality two-dimensional magnetic resonance image was reconstructed by solving the minimum L1 norm for the signal in sparse transform domain based on the guarantee of data consistency constrain.(2)Study the fast magnetic resonance angiography based on ROI-CS.Study the ROI-CS based compressed sensing model and its reconstruction method by combining ROI with CS.Study on the method of zero filling reconstruction of random sampling data under different under-sampling ratio in k-space.Study the Maximum Density Projection(MIP)for multi-layer artery images.The ROI region is segmented interactively in the MIP for zero filling reconstruction images,which is added to the CS reconstruction model as a data continuity constraint in image domain.Preliminary simulation experiments show that ROI-CS can be applied to fast magnetic resonance angiography,and under same under-sampling ratio,the quality of ROI-CS reconstruction is improved compared with the conventional CS reconstruction.(3)Study real-time on-line dynamic magnetic resonance imaging technology based on ROI-CS.In dynamic magnetic resonance imaging,the difference between frames is mainly caused by the movement of the tissue(such as the contraction of the heart,etc.),since there is redundant information between frames,a real-time on-line dynamic magnetic resonance imaging based on ROI-CS method is proposed.Study on the prediction method based on Motion Estimation/Motion Compensation(ME/MC)and fast reconstruction for residual signals based on an iterative soft threshold algorithm.A new motion estimation/motion compensation method based on ROI is proposed,the Adaptive Cross Search Algorithm is used only for the ROI region of the image,and the block artifacts is suppressed by using overlapping blocks in MC.Simulation experiment shows that the real-time on-line dynamic MRI based on ROI-CS can effectively reduce the complexity of the algorithm and improve the imaging speed.Under the same under-sampling ratio,compared with the existing real-time online imaging methods,the image quality of the proposed method is improved and the reconstruction time is shortened.