秩一分解网络用于低剂量CT噪声伪影抑制  

Rank One Decomposition Network for Noise Artifact Suppression of LDCT

作　　者：奉刚 张雄[1] 任慧莹 王悦 FENG Gang;ZHANG Xiong;REN Hui-ying;WANG Yue(School of Electronics and Information Engineering,Taiyuan University of Science and Technology,Taiyuan 030024,China)

机构地区：[1]太原科技大学电子信息工程学院,太原030024

出　　处：《太原科技大学学报》2025年第1期20-26,共7页Journal of Taiyuan University of Science and Technology

基　　金：山西省基础研究计划(20210302124265);太原科技大学研究生教育创新项目(SY2022026)。

摘　　要：基于深度学习的LDCT图像降噪方法已成为该领域的研究热点。然而,现有的基于深度学习的方法生成的CT降噪图像会出现过度平滑的现象,对有益于医学诊断准确性的部分信息造成一定程度的破坏。为了使CT图像中的有用信息得到有效保留,提出了一个秩一分解重构网络,该网络能够将CT图像的内容投影到两个不同的域:低秩域和残差域。通过对两个域内的不同语义信息进行相应的处理,在尽可能抑制伪影与噪声同时,有效保留了7CT图像中有用的组织结构信息。实验结果表明,所提出的方法在PSNR和SSIM指标值方面与对比算法中性能最好的一种相比分别提高了0.908 6 dB和0.035 6 dB.The LDCT image denoising method based on deep learning has become a research hotspot in this field.However,the noise-reduced CT images generated by the existing deep learning-based methods is always over-smoothing,which will cause some damage to some information that is beneficial to the accuracy of medical diagnosis.In order to effectively preserve the useful information in CT images,a rank-one decomposition reconstruction network is proposed,which is able to project the content of CT images into two different domains:low-rank domain and residual domain.And then process the different semantic information in the two domains accordingly.In this way,the artifacts and noises are suppressed as much as possible,and the useful tissue structure information in the CT image is effectively preserved.The experimental results show that compared with the best one in the comparison algorithms the proposed method improves the PSNR and SSIM index values by 0.9086 dB and 0.0356 dB,respectively.

关 键 词：低剂量 深度学习 秩一分解 CT 图像修复 

分 类 号：TP389.1[自动化与计算机技术—计算机系统结构]