The Research Of The Magnetic Of Resonance Imaging Reconstruction Based On Compressed Sensing

Magnetic resonance imaging is an important imaging method in current medical imaging,and its advantages such as its safety and accuracy have been widely used.However,the slow imaging method is the bottleneck restricting its further development.Compressive sensing theory is a new type of data acquisition method generated under this background.Compressed sensing theory proposed to achieve a smaller amount of sample data to accurately reconstruct the original signal.Greatly speed up the imaging speed of the image.The research was carried out from the previous research of the Magnetic Resonance Imaging Reconstruction based on compressed sensing.The existing nonlinear reconstruction algorithms were improved,and the image reconstruction accuracy has been effectively improved by adopting the variable density sampling methods.For the traditional under sampling method,it is easy to cause aliasing artifacts,Through the study of several common sampling methods,it was found that for an image signal,the energy was mainly concentrated in the low frequency band.The high frequency part of the image contained less energy distribution,so a variable density sampling method was proposed.Compared with the traditional uniform sampling method,the simulation results showed that the sampling method used could get more accurate reconstruction results Reduction of aliasing artifacts during image reconstruction..The sparse representation of signals,the design of Observation Matrix and the selection of optimal reconstruction algorithm were the three most important modules in compressed perceptual image reconstruction.The selection of the reconstruction algorithm would directly affect the visual effect of the reconstructed image.On the basis of the split Bregman iterative Algorithm,the sparsity of different features in different transform domain was found,and the combination of total variation and wavelet transform was used as a regular term by making full use of the prior information of the image.An improved split Bregman iterative image reconstruction algorithm was proposed.Finally,three different images were selected for reconstruction in the simulation experiment.The simulation results showed that,compared with the total variation regular constraint algorithm,the wavelet transformed constraint algorithm and the traditional norm ₁l model reconstruction algorithm,the improved reconstruction algorithm could effectively reduce the image reconstruction error and improve the image reconstruction accuracy.Figure 25;Table 3;Reference 58...