深度学习图像重建算法在CT成像中的应用进展  

Application progress of deep learning image reconstruction algorithm in CT imaging

作　　者：阿依古丽·托合提 班允清[1] Ayiguli Tuoheti;BAN Yunqing(Imaging Center,The Fifth Affiliated Hospital of Xinjiang Medical University,Urumqi,Xinjiang 830011,China)

机构地区：[1]新疆医科大学第五附属医院影像中心,新疆乌鲁木齐830011

出　　处：《影像研究与医学应用》2024年第24期4-6,共3页Journal of Imaging Research and Medical Applications

摘　　要：自从CT技术问世以来,疾病的诊治得到了前所未有的发展。同时,CT检查过程中产生的辐射剂量问题是医患关注的热点,因此,如何在确保图像质量的前提下,最大程度地降低辐射剂量,已成为CT研究的重点。深度学习图像重建(DLIR)算法是基于卷积神经网络的一种新兴的CT重建技术,与传统的滤波反投影、迭代重建算法相比,可以降低辐射剂量和对比剂用量,优化图像质量。因此,本文对DLIR算法的技术演变史、基本原理、技术优势及临床应用做一综述报道。Since the introduction of CT technology,the diagnosis and treatment of diseases have been unprecedentedly developed.At the same time,the radiation dose generated during CT examination is a hotspot of concern for doctors and patients,so how to minimize the radiation dose under the premise of ensuring image quality has become the focus of CT research.Deep learning image reconstruction(DLIR)algorithm is an emerging CT reconstruction technique based on convolutional neural network,which can reduce the radiation dose and contrast dosage and optimize the image quality compared with the traditional filtered back projection,iterative reconstruction algorithms.Therefore,this paper reports a general overview of the technological evolution of the DLIR algorithm,its basic principles,technical advantages and clinical applications.

关 键 词：深度学习 图像重建算法 计算机体层成像 图像质量 

分 类 号：R445.3[医药卫生—影像医学与核医学]