Compressive Sensing MR Imaging Based On Adaptive Tight Frame And Reference Image

Magnetic resonance(MR)imaging is a medical imaging technique that is widely used in medical diagnosis.It has superior performance such as non-ionizing radiation,high-quality display of anatomical structures,and soft tissue changes.However,in the MR imaging process,it is usually necessary to collect and process a large amount of data,resulting in a slower imaging speed,making the study of rapid MR imaging become one of the current hot spots.In order to speed up the imaging speed of MR imaging,it is a feasible way to reduce the amount of data collected,but limited by the traditional principle of sampling,it is difficult to reconstruct high quality MRI images by a small amount of data.The compressive sensing theory that has emerged in recent years as a new technology that can achieve high-precision signal reconstruction through few data provides a new direction for rapid MR imaging research.However,in the current MR image reconstruction method based on the compressed sensing,a predefined operator is usually used as a tool for sparse representation,and most of the search for prior information only focuses on the image itself to be reconstructed.Based on the deficiencies of the above two aspects,this paper discusses a method of MR image reconstruction based on adaptive tight frame and reference image.The main contributions are following three aspects:1)The adaptive tight frame is introduced as a sparse representation operator in the reconstruction model.Since MR images contains abundant structural information,the existing pre-defined sparse representation usually cannot fully extract sparse information of the image,while adaptive tight frame proposed in this paper can effectively obtain the sparse information of MR images because of the adaptability of the reconstructed image and the perfect reconstruction of the tight frame.2)In the reconstruction model,we explored the similarity prior information of MR images.In general,there is similarity information between MR images with the same imaging target.In this paper,we use the prior information of similarity to construct a corresponding MR images reconstruction model.Meanwhile,the similar prior information between images is effectively utilized by using an adaptive weighting method,thus improving the reconstruction quality of MR images.3)Based on the smoothing-based fast iterative shrinkage-threshold algorithm(SFISTA),the corresponding model’s solution algorithm is designed,which effectively solved the reconstruction optimization model.The simulation experiments verify the superiority of the proposed method compared with other contrast methods.