Research On Sparse Reconstruction Algorithm For MRI

Magnetic Resonance Imaging?MRI?is a medical imaging technology that can provide precise information about the structure and function of the human body.It has good contrast for soft tissue imaging and has no radiation damage.The main disadvantage of MRI is the slower imaging speed.In order to enhance the MRI speed and assure the imaging quality,this paper studies the use of sparse optimization theory methods to design an effective magnetic resonance imaging sparse reconstruction algorithm to make the performance of MRI reconstruction better.The main research contents of this article are as follows:?1?The group sparse transformation method based on the sparse structure of the wavelet tree overlap group can obviously increase the sparsity of the image,and the MRI reconstruction algorithm using the sparse model can effectively enhance the reconstruction quality of the imaging.But its reconstruction speed is relatively slow.For this reason,this paper suggests a fast magnetic resonance imaging sparse reconstruction method based on the wavelet tree overlapping group sparse model.The proposed algorithm uses the alternating direction multiplier method to solve the problem,which can clearly make the reconstruction performance of the algorithm better.The experimental simulation results show that the average reconstruction speed of the proposed algorithm is 3.3 times faster than the comparative algorithm while preserving the quality of the image.The algorithm proposed in this paper can distinctly enhance the imaging reconstruction speed.?2?Eigenvector-based SPIRi T?ESPIRi T?is a parallel MRI reconstruction model that can estimate coil sensitivity.Due to the effects of dynamic corrosion,chemical shift,ghosting,and small field of view?FOV?,the reconstruction of the ESPIRi T model requires multiple sets of sensitivity maps.In addition,the ₁ regularization term sparse constraint method of wavelet transform cannot obviously increase the sparseness of the image,resulting in a small amount of overlapping artifacts in the reconstructed image.To this end,based on the ESPIRi T model,this paper proposes a parallel MRI reconstruction algorithm based on ESPIRi T's multiple sets of sensitivity maps with Total Variation?TV?regularization term constraints.The simulation test results show that the proposed TV sparse reconstruction algorithm based on ESPIRi T can eliminate overlapping artifacts and make the quality of the reconstructed image better.?3?Although the reconstruction algorithm based on ESPIRi T's TV regularization term constraint proposed in?2?can effectively remove overlapping artifacts,there are still some slight artifacts in the reconstructed image.This is because although the TV regularization term constraint method can evidently retain the edge-preserving of the image,it is also produced a small amount of staircase effects in the reconstruction.In order to further enhance the reconstruction quality of images,a sparse parallel MRI reconstruction algorithm based on ESPIRi T's _p pseudo-norm joint TV regularization term constraint is proposed.The experimental results show that the proposed new algorithm can further improve the quality of MRI,and under high acceleration,the proposed new algorithm has better imaging effect.