Journal of Computer Applications ›› 2023, Vol. 43 ›› Issue (S1): 250-257.DOI: 10.11772/j.issn.1001-9081.2022081216

• Multimedia computing and computer simulation • Previous Articles    

Progress of U-Net applicaitons to lung nodule segmentation

Quanyou SHEN1, Xiaobo ZHANG1, Wenhao LI1, Lihan LI1, Rongde XU2, Daohua CHEN3, Jing LI4,5()   

  1. 1.School of Autumation,Guangdong University of Technology,Guangzhou Guangdong 510006,China
    2.Department of Interventional Radiology,Guangdong Provincial People’s Hospital(Guangdong Academy of Medical Sciences),Southern Medical University,Guangzhou Guangdong 510000,China
    3.The Second School of Clinical Medicine,Southern Medical University. Guangzhou Guangdong 510006,China
    4.Department of Pulmonary and Critical Care Medicine,The First People’s Hospital of Yunnan Province;(The Affiliated Hospital of Kunming University of Science and Technology),Kunming Yunnan 650032,China
    5.Department of Pulmonary and Critical Care Medicine,Guangdong Provincial People’s Hospital (Guangdong Academy of Medical Sciences),Southern Medical University,Guangzhou Guangdong 510000,China
  • Received:2022-08-17 Revised:2023-02-21 Accepted:2023-02-23 Online:2023-07-04 Published:2023-06-30
  • Contact: Jing LI

U-Net在肺结节分割中的应用进展

沈权猷1, 张小波1, 李文豪1, 李礼汉1, 许荣德2, 陈道花3, 李静4,5()   

  1. 1.广东工业大学 自动化学院,广州 510006
    2.南方医科大学附属广东省人民医院(广东省医学科学院) 微创介入科,广州 510000
    3.南方医科大学 第二临床医学院,广州 510515
    4.云南省第一人民医院(昆明理工大学附属医院) 呼吸与危重症医学科,昆明 650032
    5.南方医科大学附属广东省人民医院(广东省医学科学院) 呼吸与危重症医学科,广州 510000
  • 通讯作者: 李静
  • 作者简介:沈权猷(2001—),男,广东信宜人,主要研究方向:人工智能
    张小波(1977—),男,湖北随州人,讲师,博士,主要研究方向:人工智能
    李文豪(2000—),男,广东汕头人,主要研究方向:人工智能
    李礼汉(2001—),男,广东吴川人,主要研究方向:人工智能
    许荣德(1970—),男,广东潮阳人,主要研究方向:肿瘤介入
    陈道花(1989—),女,湖南永州人,硕士,主要研究方向:呼吸介入
    李静(1968—),女,广东广州人,主任医师,博士,主要研究方向:介入呼吸病学。dr.lijing@gdph.org.cn
  • 基金资助:
    云南省重大科技专项(202102AA100012)

Abstract:

It is of great clinical significance to achieve automatic and accurate segmentation of lung nodules in medicine. With the remarkable progress of computer vision, deep learning as a part of artificial intelligence has attracted more and more attention in the automatic segmentation of medical images. U-Net has been widely used in the field of medical image segmentation due to its good performance on small sample datasets. Researchers are currently trying to use different U-Net structures to improve the performance of computer-aided diagnosis systems in lung cancer screening of medical images. In this work, the datasets and evaluation metrics commonly used in lung nodule segmentation were first introduced, and the U-Net-based automatic segmentation techniques related to lung nodules were investigated. Then, U-Net models and improvements around codecs, skip connections and overall structure were analyzed and summarized. Finally, the challenges and limitations of deep learning-based automatic segmentation techniques were also discussed.

Key words: deep learning, U-Net, medical image processing, lung tumor segmentation, lung nodule

摘要:

医学上实现自动肺结节精准分割具有十分重要的临床意义。随着计算机视觉的显著进步,深度学习作为人工智能的一部分,在医学图像自动分割中引起了越来越多的关注。U-Net由于在小样本数据集上的良好表现,在医学图像分割领域得到广泛应用。目前,研究人员正在尝试使用不同的U-Net结构,以提高计算机辅助诊断系统在医学图像的肺癌筛查中的性能。首先,围绕肺结节分割任务介绍了当下常用的数据集和评价指标;其次,调查与肺结节相关的U-Net分割技术网络;另外,基于U-Net分别分析与归纳编解码器、跳跃连接和整体结构的改进;最后,还讨论了基于深度学习的自动分割技术的挑战和限制。

关键词: 深度学习, U-Net, 医学图像处理, 肺肿瘤分割, 肺结节

CLC Number: