基于双编码器双解码器GAN的低剂量CT降噪模型  

Low-dose CT denoising model based on dual encoder-decoder generative adversarial network

作　　者：上官宏 任慧莹[1] 张雄 韩兴隆[1] 桂志国 王燕玲[2,3] SHANGGUAN Hong;REN Huiying;ZHANG Xiong;HAN Xinglong;GUI Zhiguo;WANG Yanling(School of Electronic Information Engineering,Taiyuan University of Science and Technology,Taiyuan Shanxi 030024,China;State Key Laboratory of Dynamic Measurement Technology(North University of China),Taiyuan Shanxi 030051,China;School of Information,Shanxi University of Finance and Economics,Taiyuan Shanxi 030006,China)

机构地区：[1]太原科技大学电子信息工程学院,太原030024 [2]动态测试技术省部共建国家重点实验室(中北大学),太原030051 [3]山西财经大学信息学院,太原030006

出　　处：《计算机应用》2025年第2期624-632,共9页journal of Computer Applications

基　　金：国家自然科学基金资助项目(62001321);山西省基础研究计划项目(202103021224265,202103021223308);太原科技大学研究生教育创新项目(SY2022015,XCX212026)。

摘　　要：近年来,生成对抗网络(GAN)用于低剂量计算机断层成像(LDCT)图像降噪已经表现出显著的性能优势,成为该领域的研究热点。然而,GAN的生成器对LDCT图像中噪声和伪影分布的感知能力不足,导致网络的降噪性能受限。因此,提出一种基于双编码器双解码器生成对抗网络(DualED-GAN)的低剂量CT降噪模型。首先,提出由一对编解码器构成伪影像素级特征提取通道,用于估计LDCT中的伪影噪声;其次,提出由另外一对编解码器构成伪影掩码信息提取通道,用于估计伪影的强度和位置信息;最后,采用伪影图像质量标签图辅助估计伪影的掩码信息,可以为伪影像素级特征提取通道提供补充特征,进而提高GAN降噪网络对伪影噪声分布强度的敏感性。实验结果表明,在mayo测试集上与次优模型DESD-GAN(Dual-Encoder-Single-Decoder based Generative Adversarial Network)相比,所提模型的平均峰值信噪比(PSNR)提高了0.3387 dB,平均结构相似性度(SSIM)提高了0.0028。可见,所提模型在伪影抑制、结构保留与模型鲁棒性方面均有更好的表现。In recent years,Generative Adversarial Network(GAN)used for Low-Dose Computed Tomography(LDCT)image denoising has shown significant performance advantages,becoming a hot topic in the field.However,the insufficient perception ability of GAN generator for the noise and artifact distribution in LDCT images leads to limit the denoising performance.To address this issue,an LDCT denoising model based on a Dual Encoder-Decoder GAN(DualED-GAN)was proposed.Firstly,a pair of encoder-decoder was proposed to form an artifact pixel-level feature extraction channel for estimating the artifact noise in LDCT images.Secondly,another pair of encoder-decoder was proposed to form an artifact mask information extraction channel for estimating the intensity and location information of artifacts.Finally,the artifact image quality label maps were used to assist in estimating the mask information of artifacts,so that supplementary features were provided for the artifact pixel-level feature extraction channel,thereby enhancing the sensitivity of the GAN denoising network to the distribution intensity of artifact noise.Experimental results show that compared with the sub-optimal model DESD-GAN(Dual-Encoder-Single-Decoder based Generative Adversarial Network),the proposed model increases the average Peak Signal-to-Noise Ratio(PSNR)by 0.3387 dB,and the average Structural Similarity Index Measure(SSIM)by 0.0028 on mayo test set.It can be seen that the proposed model performs better in all terms of artifact suppression,structural preservation,and model robustness.

关 键 词：低剂量计算机断层成像 生成对抗网络 编码器 解码器 降噪 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]