MRI Image Reconstruction Based On Undersampling Technology

Magnetic resonance imaging is one of the key means of medical diagnosis and research,the most important and difficult thing in present study is to reduce the imaging time without harming or even improve the quality of the whole or partial image.Most researchers adopt hardware improvement and software processing me thods to raise the speed of magnetic resonance imaging.Chirp Z transform can be seen as extensions of Fourier transform,it can also be regarded as normal Z transform's secondary sampling in the Z plane.So,can the Chirp Z transform's partial frequency refining ability be used into undersample magnetic resonance image and improve the image's local quality?To solve this problem,based on the study of compressed sensing-magnetic resonance imaging,this article explores the combination possibility of Chirp Z transform and compressed sensing-magnetic resonance imaging,puts forward a magnetic resonance image processing method which combined the Chirp Z transform with compressed sensing-magnetic resonance imaging's reconstruction algorithm.With the utilize of Chirp Z transform's advantage,frequency spectrum refine and multi-parameter control,this article proposes a new way to increase the ability of showing magnetic resonance image's local details,the main work includes:First of all,according to the principle and software algorithm study of Chirp Z transform and its inverse transformation,realize its application in image processing and reconstruction.Through a large number of simulation experiments,this article explores the influence of Chirp Z transform parameters on K space and reconstruction image,the parameters are initial sampling point radius A₀,initial sampling phase angle ?₀,spiral elongation W₀ and the phase angle between two sampling points ?₀.Secondly,based on the advantage of Chirp Z transform and compressed sensing-magnetic resonance imaging reconstruction algorithm,proposes a magnetic resonance image reconstruction method which combined Chirp Z transform with orthogonal matching pursuit algorithm.Experiments in the same simulation environment,with orthogonal matching pursuit reconstruction algorithm,compare the reconstruction result of Chirp Z transform,discrete Fourier transform,wavelet sparse transformatio n and discrete cosine transform,analyze different algorithms'advantages and disadvantages.Finally,puts forward a magnetic resonance image processing method which combined Chirp Z transform with wavelet tree sparsity algorithm,a new compressed sensing-magnetic resonance imaging reconstruction method,take the combination of Chirp Z,wavelet tree sparsity and total variation as magnetic resonance image reconstruction sparse constraint.The reconstruction process is based on alternating minimization ideas,incorporate with fast composite splitting algorithm.Experiments in the different sampling rate show that compared with other algorithms,the reconstruction results of Chirp Z transform and wavelet tree spars ity reconstruction algorithm's combination have a better edge display ability,the proposed algorithm fulfills image spectrum refinement.