北京生物医学工程2025，Vol.44Issue(3)：308-314,7.DOI:10.3969/j.issn.1002-3208.2025.03.014
U-Net改进及其在肺部CT图像分割上的应用
U-Net improvement and its application on lung ct image segmentation
摘要
Abstract
Lung cancer and COVID-19 have become major hidden dangers threatening human life and health.Regular detection of human lungs by computed tomography(CT)is the main means of early detection of lung diseases.CT imaging can provide doctors with a clear lung condition,help to find the infected area of the focus,and play an important role in the screening of lung cancer and COVID-19.It is often used in medicine to assist in the diagnosis of various lung diseases.The improved U-Net model is used to segment the focal area on CT images,which is helpful to improve the diagnosis speed of doctors and relieve the pressure of domestic medical treatment.Therefore,using the improved U-Net model to segment CT images is a research hotspot in the field of computer-aided diagnosis,which is of great significance for the subsequent diagnosis and treatment of lung diseases.This review compares and analyzes the improved strategies of U-Net model and its application in pulmonary nodules and COVID-19 diseases.Firstly,the method of improving the structure of U-Net model is discussed,and the advantages and disadvantages of related improved models are analyzed.Secondly,taking the CT image of lung nodule and COVID-19 as an example,the research progress of U-Net improved model in the field of lung nodule and COVID-19 image segmentation is analyzed.Finally,the challenge and future development direction of deep learning in the field of lung image segmentation are discussed,which is of great significance to improve the speed of clinical diagnosis and early prevention of lung disease.关键词
肺结节/COVID-19/深度学习/U-Net/图像分割Key words
pulmonary nodule/COVID-19/deep learning/U-Net/image segmentation分类
基础医学引用本文复制引用
杨超,葛明锋,董文飞..U-Net改进及其在肺部CT图像分割上的应用[J].北京生物医学工程,2025,44(3):308-314,7.基金项目
中科院装备研制项目(YJKYYQ20210032)资助 （YJKYYQ20210032）
