Research On CT Image Reconstruction Based On Total Variation And Dictionary Learning

Nowadays,Computed Tomography(CT)is widely used in modern medicine and provides important evidence for clinical diagnosis.However,studies have shown that excessive X-ray doses or long exposure times can cause radiation damage to the human body.At the same time,the quality of computer tomographic images is closely related to the dose of X-rays.The higher the dose of X-rays,the higher the resolution of the image.In order to reduce the dose of X-rays and reduce the damage to the human body while ensuring that the CT image quality is sufficient for clinical diagnosis,Low-dose CT imaging has become a research hotspot in the field of medical imaging.Iterative reconstruction algorithm is an effective method to solve the problem of low-dose CT reconstruction.It can optimize the reconstructed image by establishing an objective function to iterate.This method does not require high completeness of the projection data and can pass prior information and appropriate constraints.Improve image quality.For the problem of low-dose CT image reconstruction,this article will focus on iterative reconstruction algorithms to improve the quality of CT images by combining regularization information.The main research work of this paper is as follows:(1)Based on the theory of Compress Sensing(CS),this paper improves and implements a statistical iterative reconstruction algorithm based on Total Variation(TV)regularization and dictionary learning(DL).This algorithm uses the Penalized Weighted Least-Square(PWLS)as the reconstruction algorithm.At the same time,it introduces a priori information,combines full variational regularization and dictionary learning,integrates it into an objective function,and solves it in a circular iterative manner.The objective function continuously updates the image to be reconstructed during the iteration process,thereby achieving the purpose of effectively suppressing artifacts and noise.(2)Aiming at the problems of iterative reconstruction algorithm combined with dictionary learning,which are computationally complex and time-consuming,this paper proposes an algebraic iterative reconstruction algorithm based on improved dictionary learning.This algorithm uses the Algebraic Reconstruction Technique(ART)algorithm as the reconstruction algorithm.It adds orthogonal constraints to the dictionary learning process and uses incomplete projection data to perform orthogonal dictionary learning to obtain an adaptive orthogonal dictionary that matches the image.The singular value solution method is used to reduce the computational complexity,shorten the reconstruction time,and ensure the quality of CT images.In this paper,two iterative reconstruction algorithms are proposed and implemented for low-dose CT image reconstruction.The feasibility and effectiveness of the algorithm are verified in the experimental part.Compared with conventional CT image reconstruction algorithms,the method proposed in this paper can effectively suppress artifacts and noise in CT images and significantly improve the quality of reconstructed images.