CT图像重建算法的改进  

Improvement of CT Image Reconstruction Algorithm Abstract

作　　者：马海英[1] 向顺灵 

机构地区：[1]广西民族大学信息科学与工程学院,广西南宁530006

出　　处：《河南科技》2016年第3期44-45,共2页Henan Science and Technology

摘　　要：CT图像重建算法是医学影像学的主要研究领域,然而CT图像重建时不可避免地存在噪声、伪影,因此,需要研究一种新的算法对CT图像进行去噪处理,进而重建出低噪声、高分辨率的CT图像。基于此,本文提出基于过完备字典稀疏的去噪模型,并运用于CT图像去噪,同时将基于低秩矩阵分解应用于CT重建,核心思想是利用CT图像投影矩阵的稀疏特性,提出了一种新的CT图像重建法,其重建过程分成2个步骤:一个是过完备字典稀疏法(K-SVD)进行图像去噪预处理,一个是低秩矩阵分解(LRMD)更新CT图像。结果表明,提出的方法能更好地保持图像细节,低秩矩阵的特性使得算法的复杂度得以降低,提出的方法具有较好的去噪效果,提高了重建图像的质量。Computed tomography （CT） image reconstruction is an important research subject in fie ld of m edicalimaging.But as the heavily influence of the noise in medical CT image, We must choose appropriate denoisingmethod for image preprocessing to get the lowest noise images, w hile without sacrificing image precisionand spatial resolution. The thesis according to the sparsity of projection m atrix on CT image, proposinga new CT image reconstruction algorithm, the reconstruction process has two steps： fisdy, overcompletedictionaries for sparse representation method （K - SVD） which is applied to image denoising, then a lo w -rank decomposition of m atrix which is used to update CT images. Experimental results showed that the proposedmethod had strong ability to keep the details of the CT images, the characteristics of low -rank matrixto sim plify the calculation process, and to reduce the complexity of the algorithm, and the denoisingmethod had good denoising effect, improving the accuracy of the reconstructed image, and ultim atelyachieved a high quality image.

关 键 词：CT图像重建 过完备字典稀疏表示法 低秩矩阵分解 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]