sparse angle CT image reconstruction method based on a weighted kernel norm minimization

Abstract

The invention discloses a sparse angle CT image reconstruction method based on a minimum weighted nuclear norm. The method comprises the steps of firstly, reconstructing a CT initial image by adoptinga SART algorithm; And correcting the CT initial image by using a weighted improved singular value truncation algorithm to obtain a reconstructed CT image. According to the invention, similarity and redundancy in a CT image domain are fully utilized; According to the method, the weighted nuclear norm of the CT image is used as a regular item and is introduced into the CT reconstruction problem, and the image can be clearly reconstructed under the condition that the sampling angle is sparse, so that the method can be suitable for sparse angle CT reconstruction work under the conditions of sparse angle and low dosage, and has important significance for reducing radiation of CT scanning to a human body.

Description

technical field [0001] The invention belongs to the technical field of CT image processing, and relates to a method for obtaining CT images based on projection data acquired by sparse angles, in particular to a method for reconstructing CT images of sparse angles based on a very small weighted kernel norm. Background technique [0002] In the medical field, X-ray computerized tomography (abbreviated as CT) is a widely used imaging method, and it is becoming more and more popular in medical aided diagnosis. Studies have shown that excessive radiation doses will increase the risk of cancer. Therefore, it is a subject of great significance to study how to obtain clinically acceptable medical images at low doses. [0003] Methods to reduce X-ray radiation dose are mainly divided into two categories. One is to reduce the sampling angle through hardware improvement, so as to achieve the effect of reducing radiation dose. The images obtained by such methods often contain noise an...

AI Technical Summary

Problems solved by technology

[0006] Although the above methods have improved the quality of sparse-angle CT reconstruction images to varying degrees, the above-mentioned methods still have the following defects: (1) The reconstructed image contains obvious artifacts when the sampling angle is sparse in the filtered back-projection method ; (2) The reconstruction quality of the algebraic reconstruction method is better than that of the filtered back-projection method, but it is still difficult to meet the requirements of medical diagnosis; (3) The method of using the total variation as a regular term, and other improvements to the total variation methods, since they are all based on the assumption of piecewise smoothness, block artifacts are inevitably introduced; (4) methods based on dictionary learning, non-local means, etc. do not make full use of the association between non-locally similar blocks in the image domain

Images