基于改进DTCWT和引导滤波的低剂量CT图像降噪  

Low-dose CT image denoising based on improved DTCWT and guided filter

作　　者：徐勇 张志伟[1] 朱宏禹 韩建宁[1] Xu Yong Zhang;Zhiwei;Zhu Hongyu;Han Jianning(School of Information and Communication Engineering,North University of China,Taiyuan 030051,China)

机构地区：[1]中北大学信息与通信工程学院,太原030051

出　　处：《国外电子测量技术》2023年第3期189-194,共6页Foreign Electronic Measurement Technology

基　　金：山西省基础研究计划(202103021224201);国家自然科学基金(61671414)项目资助。

摘　　要：针对低剂量计算机断层扫描(CT)图像噪点较多和空间分辨率低的问题,提出了一种基于改进双树复小波变换(DTCWT)和引导滤波的低剂量CT图像降噪方法。首先使用DTCWT对低剂量CT图像进行多尺度分解,对高频子带系数使用拟合优度检验框架进行去噪,对低频子带使用三边滤波进行降噪,逆DTCWT重构得到初始去噪图像;然后使用降噪并插值的低频子带图像作为导向图像对CT图像进行引导滤波;接着结合初始去噪图像和引导滤波的去噪结果生成细节残存图像;最后用3维块匹配滤波(BM3D)阈值算法对细节残存图像进行二次去噪,叠加两次去噪结果得到最终去噪图像。实验结果表明,该方法优于其他传统图像去噪算法,在抑制噪声的同时,良好地保留了图像的边缘轮廓和纹理特征,对低剂量CT图像有着显著的去噪效果。In order to solve the problem of many noises and low spatial resolution of low-dose CT images,a low-dose CT image denoising method based on improved dual-tree complex wavelet transform and guided filtering is proposed.Firstly,the low-dose CT image is decomposed by DTCWT,the high-frequency subband coefficients are denoised by goodnessof-fit test framework,the low-frequency subband is denoised by trilateral filtering,and the initial denoised image is obtained by inverse DTCWT reconstruction.Then,the low-frequency subband image which is denoised and interpolated is used as the guide image to guide and filter the CT image.Then,the detail residual image is generated by combining the denoising results of the initial denoising image and the guide filter.Finally,the BM3Dthreshold algorithm is used to denoise the remaining details of the image twice,and the final denoised image is obtained by superimposing the denoising results twice.The experimental results show that this method is better than other traditional image denoising algorithms,and the edge outline and texture features of the image are well preserved while suppressing noise,and has a remarkable denoising effect on low-dose CT images.

关 键 词：低剂量CT 双树复小波 引导滤波 拟合优度检验 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]