Research And Implementation Of Compressive Sensing Magnetic Resonance Imaging Reconstruction

With the development of science and technology,Magnetic resonance imaging(MRI)is a revolutionary tool in medical imaging,which plays a signigicant role in clinical diagnosis.And it's noninvasion,non-invasive,pinoint accurancy and can be faulted from arbitrary direction.Recent years,Sparse representation also be led in analysis of MRI images.At the same time,Compressive sensing(CS)has shown great potential in significantly reducing the acquisition time of MRI scanning.Compareing with nature images,MRI images are featured with large area of smooth regions,sharp edges and rich textures,which means researchers need to combine previous knowledge about image reconstruction with characteristic of MRI images.Therefore,how to improve the reconstruction quality with limited k-space data is still a challenge.Motivated by the problems of the existing reconstruction methods,for example,quality of the reconstruction isn't accurate enough and the time is too long,we analyze the background knowledge of MRI image reconstruction at first.After learning more about sparse representation and compressive sensing,we propose two kinds of methods.The first is a nonlocal autoregressive model(NAM)for CS MRI reconstruction.Nonlocal similarity between image patches is exploited as a regularization term to constrain the nonlocal feature in MRI images,which is very helpful in preserving edge sharpness.While an autoregressive regularization term is employed to describe the linear correlation between neighboring pixels,which preserves more spatial details.Different from previous work,this method combines both patches and neighboring pixels.The second one is an optimization reconstruction algorithm based on dictionary learning.In view of the characteristics of MRI image block with geometric orientation,we select an adaptive multi-directional dictionary.The algorithm incorporates global auto-regression constraints based the classical reconstruction framework.The innovation point of the algorithm is that it takes the use of self-similarity and geometric orientation between image blocks.This provides a reasonable sparsity constraint for the sparse model.Extensive experimental results demonstrate that our methods outperform mainstream methods in MRI reconstruction in terms of both objective quality and subjective quality.