计算机工程与应用 ›› 2022, Vol. 58 ›› Issue (18): 59-70.DOI: 10.3778/j.issn.1002-8331.2201-0397

• 热点与综述 • 上一篇    下一篇

深度学习在肺结节辅助诊断中的应用

冯妍妍,魏德健,倪伟   

  1. 1.山东中医药大学 智能与信息工程学院,济南 250355
    2.肥城市人民医院 信息管理科,山东 肥城 271600
  • 出版日期:2022-09-15 发布日期:2022-09-15

Application of Deep Learning in Auxiliary Diagnosis of Pulmonary Nodules

FENG Yanyan, WEI Dejian, NI Wei   

  1. 1.School of Intelligence and Information Engineering, Shandong University of Traditional Chinese Medicine, Jinan 250355, China
    2.Information Management Department, Feicheng People’s Hospital, Feicheng, Shandong 271600, China
  • Online:2022-09-15 Published:2022-09-15

摘要: 肺癌位居癌症死亡率首位,对其进行早期诊断和治疗可降低肺癌患者的死亡率。深度学习能够自动提取结节特征,并完成肺结节的良恶性及恶性等级分类,因此深度学习方法成为肺癌早期诊断的重要手段。对常用数据集进行介绍,系统阐述了栈式去噪自编码器(SDAE)、深度置信网络(DBN)、生成对抗网络(GAN)、卷积神经网络(CNN)、循环神经网络(RNN)和迁移学习技术在肺结节良恶性分类中的应用,阐述了深度卷积生成对抗网络(DCGAN)、多尺度卷积神经网络(MCNN)、U型网络(U-Net)和集成学习技术在肺结节恶性等级分类中的应用,针对肺结节分类的深度学习方法进行了综合分析,并对未来研究方向进行展望。

关键词: 肺结节, 深度学习, 良恶性分类, 恶性等级分类, 计算机辅助诊断

Abstract: Lung cancer ranks first in cancer mortality, and its early diagnosis and treatment can reduce the mortality rate of lung cancer patients. Deep learning can automatically extract nodule features, and complete the classification of benign, malignant and malignant grades of pulmonary nodules. Therefore, deep learning methods have become an important means of early diagnosis of lung cancer. This paper introduces the commomly used datasets, then systematically expounds the application of stacked denoising autoencoder(SDAE), deep belief network(DBN), generative adversarial network(GAN), convolutional neural network(CNN), recurrent neural network(RNN) and transfer learning techniques in the classification of benign and malignant pulmonary nodules. And then, the application of deep convolutional generative adversarial network(DCGAN), multi-scale convolutional neural network(MCNN), U-shaped network(U-Net) and ensemble learning techniques in the classification of malignant grades of pulmonary nodules is described. A comprehensive analysis of deep learning methods for pulmonary nodules is carried out, and the future research directions are prospected.

Key words: pulmonary nodules, deep learning, benign and malignant classification, malignant grade classification, computer-aided diagnosis