Research On Compressed Sensing Based On Image Prior Information

Magnetic resonance imaging is an important medical imaging technology with the advantages of accurate diagnosis and harmless to the human body,etc.Due to the principle of MR imaging,the long imaging time caused will cause physical discomfort for some patients,which has become a major bottleneck in the development of MRI.Therefore,on the premise of ensuring imaging,how to shorten the imaging time is a hot topic of recent research.Compressed Sensing(CS)theory developed in recent years is an innovative signal sampling theory.For sparse or compressible signals,the compressed sensing theory can also accurately recover the signal when the number of samples is much less than the traditional Shannon-Nyquist sampling theorem.Therefore,applying compressed sensing to MR imaging can greatly reduce the number of samples,thus providing a new idea for shortening MR imaging.Therefore,after a detailed study on the theory of compressed sensing and MRI,this article will study the MRI reconstruction technology based on compressed sensing.The main work and innovations of this article include the following three aspects:(1)Understand the common sparse basis of CS-MRI,study the sparse transform discrete shear wave and its properties used in this paper,and use the discrete shear wave as a sparse matrix.According to the experimental results,we can know the image reconstruction based on discrete shear wave Both the visual effect and the index data are superior to the image reconstruction based on wavelet transform.(2)On the basis of the traditional CS-MRI algorithm,using the region of interest of the MR brain image as the a priori information of the image,a new algorithm ROI-CS MRI model is proposed.First,the mask matrix W is designed by interactively selecting regions of interest,and added to the traditional CS-MRI mathematical framework.Numerical experiments show that the new algorithm helps to reconstruct images better,and it is more convenient for diagnosticians to study MR images.(3)On the basis of adding the region of interest,we proposed a new algorithm model----ROI-LACS MRI model based on the sparsity in the time domain,and based on the iterative soft threshold algorithm,Chose a more optimized algorithm,namely the fast soft threshold iteration method(SFISTA),and improved according to the objective function,and obtained a new reconstruction algorithm----RSFISTA.According to the numerical experiment,we can know the new model,get better imaging effect,and reduce the reconstruction time.