人工智能在医学CT图像重建中的研究进展  

Research and Progress of Artificial Intelligence in Medical CT Image Reconstruction

作　　者：李青 李润睿 强彦 成煜斌 王涛 LI Qing;LI Runrui;QIANG Yan;CHENG Yubin;WANG Tao(College of Information and Computer,Taiyuan University of Technology,Taiyuan 030024,China)

机构地区：[1]太原理工大学信息与计算机学院,太原030024

出　　处：《太原理工大学学报》2023年第1期1-16,共16页Journal of Taiyuan University of Technology

基　　金：国家自然科学基金资助项目(61872261);国家自然科学基金重大项目资助(U21A20469)。

摘　　要：计算机断层扫描成像(CT)是临床医学中广泛使用的一种医学图像,它可以清晰地可视化人体内部精细结构细节。在临床操作中,为防止患者暴露在高辐射X射线束下引起组织受损,通常最小化X射线以获得CT图像,但会导致成像质量严重下降。为解决上述矛盾,如何重建出符合临床需求的CT图像是国内外研究者广泛关注的、具有挑战性的难点问题。随着人工智能领域深度学习技术的蓬勃发展,在大数据驱动下,利用深度学习技术来提升CT重建质量成为当前研究热点。本文分析了CT图像重建机理;总结了现有重建模型并梳理了重建方法的优劣势,根据深度学习方法的成像过程,将现有方法分为4大类,并依次介绍4类方法的基本思想,总结了重建方法优缺点;归纳了目前公开的公共数据集以及增加训练样本方法,并对损失函数的多样性进行对比分析;讨论了该新兴领域目前仍然存在的问题,展望了后续研究中需要解决的关键问题,以便于相关研究人员了解CT重建领域的研究现状,促进该领域的长足发展。Computed tomography(CT) is a medical image widely used in clinical medicine that visualizes the fine structural details inside human body. In clinical practice, to prevent the tissue damage caused by patient exposure to high-radiation X-ray beams, X-rays are usually minimized to obtain CT images, which results in a severe deterioration in imaging quality. To solve the above contradictions, how to reconstruct CT images that meet clinical needs has become a challenging problem widely concerned by researchers at home and abroad. With the vigorous development of deep learning technology in the field of artificial intelligence, the use of deep learning technology to improve the quality of CT reconstruction has become a hot topic in the current research under the drive of big data. In this paper, the mechanism of CT image reconstruction was analyzed. According to the imaging process of deep learning methods, the existing methods were divided into 4 categories, the basic ideas of 4 types of method introduced in turn, and the advantages and disadvantages of reconstruction methods summarized. The currently published public data set and the method of increasing training samples were summarized, and the diversity of loss functions;The problems that still exist in this emerging field are discussed, and the key problems that need to be solved in the follow-up research are looked forward so that relevant researchers can understand the research status in the field of CT reconstruction and promote the rapid development of the field.

关 键 词：人工智能 计算机断层扫描 图像重建 深度学习 投影域 图像域 双域网络 

分 类 号：TP3-05[自动化与计算机技术—计算机科学与技术] TP389.1