Research On Compressed Sensing Algorithm And Its Application In MRI Image

In recent years,the compressed sensing technology proposed can realize the high probability reconstruction of the original signal at the receiving end only by a small number of measurements.In recent years,compressed sensing has received extensive attention at home and abroad.Signal reconstruction is an important step.Therefore,this paper makes some improvements on the shortcomings of the popular signal reconstruction algorithm and puts forward some new algorithms.And apply it to the field of MRI image.Firstly,this paper introduces MRI images and compressed sensing in detail,and studies CS MRI(compressed sensing nuclear magnetic image processing),which is currently popular,are carried out.The main algorithms of compressed sensing are analyzed and studied.According to different algorithms to understand its shortcomings and put forward the improvement scheme.First on the matching pursuit algorithm for the sparse degree of dependence of the K value and the selection of atoms procrastination,paper according to the rule of RIP a today's sparse degree K value estimation method,before the iteration algorithm,proposition is obtained by RIP criterion to anticipate,an approximate real value K value of the initial K,greatly improve the performance of the matching pursuit algorithm.Compared with some algorithms,it has been successfully applied to MRI images.Secondly,in view of the shortcomings of the L0 minimum norm method in signal reconstruction accuracy and anti-noise performance,the iterative threshold algorithm proposed many solutions but there are still many limitations.The lasso model based on logarithm and norm is studied in this paper,and based on this,an iterative and contraction operator for logarithm and norm is proposed according to ISTA algorithm and FISTA algorithm.According to the proposed operator,LISTA algorithm based on logarithm and norm model is implemented.The algorithm realizes the anti-noise mechanism of LISTA algorithm through the selection of regularization parameters and so on.And has been successfully applied to MRI images,which is improved compared with the current popular CS MRI algorithm.The good performance of the reconstruction algorithm proposed in this paper is verified,which lays a foundation for its application in other fields.