Research On Medical Image Reconstruction Based On Elastic Network Regularization

Medical image reconstruction technology is an important application and research hotspot in the field of biomedical engineering,which involves multidisciplinary knowledge of mathematics,digital image processing and biomedical engineering.It has been successfully applied to medical diagnosis,surgical planning,simulation,radiation therapy,virtual endoscopy etc.In this paper,the existing algorithms are improved by using the flexible elastic network regularity strategy and the stable alternating direction method of multipliers(ADMM).The parallel imaging and iterative reconstruction algorithms are studied in this paper,and applied to medical images,including magnetic resonance imaging(MRI)and computer tomography(CT)reconstruction.The proposed algorithms achieve good experimental results.The main work of this paper is as follows:(1)A new algorithm for parallel reconstruction of MRI images based on elastic network regularity is proposed.Based on the basic algorithm of parallel magnetic resonance imaging,the SENSE algorithm based on image domain reconstruction and GRAPPA algorithm and SPIRi T algorithm based on k space domain reconstruction are proposed.A new model based on elastic network regularity is proposed,ADMM for variable separation solution,and achieved good reconstruction effect.The SENSE algorithm uses the coil sensitivity to reconstruct the spatial coding function and is the optimal reconstruction method at the known coil sensitivity.The GRAPPA and SPIRi T algorithms first calculate the weighting factor with the full sampling check data,and then use the weighting factor to calculate the acquired data to reconstruct the image.The experimental results show that the improved parallel algorithm can improve the image quality.(2)A new algorithm for iterative reconstruction of CT images based on elastic network regularity is proposed.Due to the noise statistic characteristics of CT measurement data,the statistical iterative reconstruction method has a great advantage compared with the traditional analytic reconstruction algorithm in noise removal and spatial resolution.However,due to the large amount of statistical iteration calculations and the slow convergence rate,it is difficult to apply to clinical examination.Based on the variable separation technique,this paper presents an iterative model based on the elastic network regularity and applies the ADMM algorithm to CT statistical iterative reconstruction.The experimental results show that the new iterative model can improve the quality of the reconstructed image compared with the classical filtered backprojection,and has faster convergence speed and smaller computational time complexity.