基于深度学习的低剂量计算机断层扫描图像去噪算法的研究进展  

Research progress in low-dose computed tomography image denoising algorithm based on deep learning

作　　者：张圣钧 王春祥[1] Zhang Shengjun;Wang Chunxiang(Department of Radiology,Tianjin Children's Hospital(Children's Hospital,Tianjin University),Tianjin 300074,China)

机构地区：[1]天津市儿童医院(天津大学儿童医院)医学影像科,天津300074

出　　处：《国际生物医学工程杂志》2024年第6期616-621,共6页International Journal of Biomedical Engineering

基　　金：天津市医学重点学科(专科)建设项目基金(TJYXZDXK-040A)。

摘　　要：在计算机断层扫描(CT)检查时,较高的辐射剂量会引起健康问题,应在确保诊断要求的前提下,尽可能降低辐射剂量。但低剂量计算机断层扫描(LDCT)会增大图像噪声,降低图像质量,影响诊断的正确率。因此,LDCT图像去噪算法是研究的重点方向。综述了基于深度学习的LDCT图像的投影域预处理去噪、图像重建去噪、图像域后处理去噪以及这3个过程相组合的双域处理去噪算法的研究进展。总结了目前LDCT图像去噪算法所面临的问题,以促进深度学习在该算法中的应用。Higher doses of radiation from computed tomography(CT)scans can cause health problems.The radiation dose should be as low as possible while still meeting diagnostic needs.However,low-dose computed tomography(LDCT)increases image noise,reduces image quality and compromises diagnostic accuracy.Researchers have focused on the denoising technology for LDCT image.In this review,the research progress of projection domain pre-processing denoising,image reconstruction denoising,image domain post-processing denoising and dual-domain processing denoising algorithm based on deep learning for LDCT image were reviewed.The current problems of LDCT denoising were also summarized to promote the application of deep learning in LDCT image denoising.

关 键 词：计算机断层扫描 低剂量 深度学习 图像去噪 

分 类 号：TP3[自动化与计算机技术—计算机科学与技术]