AWTV和高斯注意力引导的LDCT图像去噪网络  

Adaptive Weighted Total Variational and Gaussian Attention-Guided LDCT Image Denoising Networks

作　　者：李志媛 刘祎[2,3] 张鹏程[2,3] 张丽媛 任时磊 芦婧 桂志国[2,3] LI Zhiyuan;LIU Yi;ZHANG Pengcheng;ZHANG Liyuan;REN Shilei;LU Jing;GUI Zhiguo(School of Software,North University of China,Taiyuan 030051,China;State Key Laboratory of Dynamic Testing Technology,North University of China,Taiyuan 030051,China;School of Information and Communication Engineering,North University of China,Taiyuan 030051,China;School of Computer Science and Technology,North University of China,Taiyuan 030051,China)

机构地区：[1]中北大学软件学院,太原030051 [2]中北大学省部共建动态测试技术国家重点实验室,太原030051 [3]中北大学信息与通信工程学院,太原030051 [4]中北大学计算机科学与技术学院,太原030051

出　　处：《计算机工程与应用》2025年第3期253-263,共11页Computer Engineering and Applications

基　　金：山西省回国留学基金委项目(2021-111);国家自然科学基金(61801438);山西省高等学校科技创新项目(2020L0282)。

摘　　要：低剂量CT(low-dose CT,LDCT)图像的去噪任务是一个高度复杂且不确定的逆问题。现有的基于CNN的方法虽然有效,但提升空间有限且计算成本高。相比之下,将图像先验知识与模型相结合来辅助图像去噪是一种更有效的方法。提出了一种名为AWTV_GANet的LDCT图像去噪框架。该框架利用自适应加权总变分(adaptive weighted total variation,AWTV)展开和高斯注意力引导的方法,通过端到端的CNN模型,将噪声优化模型、边缘检测模型和图像重建模型集成在一起。实验证明,AWTV_GANet能够准确地去除伪影噪声,并恢复出更精细的结构细节,与其他方法相比具有优异的性能。The task of denoising low-dose CT(LDCT)images is a highly complex and uncertain inverse problem.Although the existing CNN-based methods are effective,the improvement space is limited and the calculation cost is high.In contrast,it is a more effective method to combine image prior knowledge with model to assist image denoising.In this paper,an LDCT image denoising framework named AWTV_GANet is proposed.The framework uses adaptive weighted total variation(AWTV)expansion and Gaussian attention guidance to integrate the noise optimization model,edge detection model and image reconstruction model through the end-to-end CNN model.Experiments show that AWTV_GANet can accurately remove artifact noise and recover finer structural details,which has excellent performance compared with other methods.

关 键 词：低剂量CT 图像去噪 深度学习 自适应加权总变分 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]