摘要
在计算机断层扫描(Computed Tomography,CT)成像领域,全变分(total variation,TV)重建算法可以从稀疏角度投影数据中重建出高质量的图像而不引入显著的伪影.为了进一步改善算法的性能,本文提出了一种结合邻域信息的TV正则化稀疏角度重建算法.首先通过像素邻域信息的均值和均方差构建了一个自适应权重函数,然后引入到TV模型中以此利用图像的各向异性边缘属性.本文算法可以自适应调节图像局部信息进一步改善了图像的稀疏性,可以更好的重建图像.应用此算法对Shepp-Logan仿真模型和真实的核桃投影数据进行重建,实验结果表明,该算法在抑制伪影和保留边缘结构细节信息方面能够取得更好的性能.
In the field of Computed Tomography (CT) imaging,total variation (TV) reconstruction algorithm can be used to reconstruct high quality images from sparse-view projection data without introducing significant artifacts. To further improve the algorithm’s performance,we propose a total variation regularization sparse-view reconstruction algorithm combined with neighborhood information in this paper. The proposed method is constructing into adaptive-weighted function by the mean and mean square error of pixel neighborhood information. Then it is introduced into the TV model to make use of the anisotropic edge property of the image. This method can adjust adaptively the image local information to further improve the image sparsity and reconstruct the image better. The algorithm is used to reconstruct the Shepp-Logan simulation model and the real walnut projection data. Experimental results showthat the proposed algorithm has achieve a better performance in artifacts suppression and edge structure details preservation.
作者
齐泽瑶
王远军
QI Ze-yao;WANG Yuan-jun(School of Medical Instrument and Food Engineering, University of Shanghai for Science and Technology, Shanghai 200093 , China)
出处
《小型微型计算机系统》
CSCD
北大核心
2019年第8期1745-1749,共5页
Journal of Chinese Computer Systems
基金
上海市自然科学基金项目(18ZR1426900)资助
