《计算机应用》唯一官方网站 ›› 2022, Vol. 42 ›› Issue (5): 1424-1430.DOI: 10.11772/j.issn.1001-9081.2021050813

• 人工智能 • 上一篇    下一篇

基于神经网络架构搜索的肺结节分类算法

谢新林1,2, 肖毅3, 续欣莹3()   

  1. 1.太原科技大学 电子信息工程学院, 太原 030024
    2.先进控制与装备智能化山西省重点实验室(太原科技大学), 太原 030024
    3.太原理工大学 电气与动力工程学院, 太原 030024
  • 收稿日期:2021-05-17 修回日期:2021-09-26 接受日期:2021-11-26 发布日期:2022-03-08 出版日期:2022-05-10
  • 通讯作者: 续欣莹
  • 作者简介:谢新林(1990—),男,山西运城人,讲师,博士,CCF会员,主要研究方向:医学图像处理、粗糙集
    肖毅(1996—),男,江西赣州人,硕士研究生,主要研究方向:医学图像处理
    续欣莹(1979—),男,山西忻州人,教授,博士,主要研究方向:计算机视觉、粒计算。 xuxinying@tyut.edu.cn
  • 基金资助:
    国家自然科学基金资助项目(62006169);山西省自然科学基金资助项目(201901D211304)

Lung nodule classification algorithm based on neural network architecture search

Xinlin XIE1,2, Yi XIAO3, Xinying XU3()   

  1. 1.School of Electronic Information Engineering,Taiyuan University of Science and Technology,Taiyuan Shanxi 030024,China
    2.Shanxi Key Laboratory of Advanced Control and Equipment Intelligence (Taiyuan University of Science and Technology),Taiyuan Shanxi 030024,China
    3.College of Electrical and Power Engineering,Taiyuan University of Technology,Taiyuan Shanxi 030024,China
  • Received:2021-05-17 Revised:2021-09-26 Accepted:2021-11-26 Online:2022-03-08 Published:2022-05-10
  • Contact: Xinying XU
  • About author:XIE Xinlin, born in 1990,Ph. D.,lecturer. His research interests include medical image processing,rough set.
    XIAO Yi, born in 1996,M. S. candidate. His research interests include medical image processing.
    XU Xinying, born in 1979, Ph. D., professor. His research interests include computer vision,granular computing.
  • Supported by:
    National Natural Science Foundation of China(62006169);Natural Science Foundation of Shanxi Province(201901D211304)

摘要:

肺结节分类是早期肺癌诊断的重要任务。基于深度学习的肺结节分类方法虽然能够取得良好的分类精度,但存在模型复杂和可解释性差的问题。为此,提出了一种基于神经网络架构搜索的肺结节分类算法。首先,将注意力残差卷积cell作为搜索空间的基本单元,并使用偏序剪枝方法作为搜索策略来构建神经网络架构以搜索3D分类网络,从而达到网络性能和搜索速度的平衡。其次,在网络中构建了多尺度通道和空间注意力模块来提高特征描述和类别推理的可解释性。最后,采用堆叠法将搜索到的网络架构进行多模型的融合,从而获取精准的肺结节良恶性分类预测结果。实验结果表明,在肺结节分类常用数据集LIDC-IDRI上,所提算法与最新肺结节分类算法相比具有较好的分类性能和较快的收敛,且所提算法的特异性和精确率分别达到95.37%和93.42%,能够实现良恶性肺结节的准确分类。

关键词: 肺结节分类, 神经网络架构搜索, 注意力模块, 多模型融合, 深度学习

Abstract:

Lung nodule classification is an important task in the diagnosis of early-stage lung cancer. Although the lung nodule classification methods based on deep learning can achieve good classification accuracy, they have the problems of complex model and poor interpretability. Therefore, a lung nodule classification algorithm based on neural network architecture search was proposed. Firstly, the attention residual convolution cell was regarded as the basic unit of the search space, and the partial order pruning method was used as the search strategy to construct the neural network architecture for searching 3D classification network, thereby achieving the balance between network performance and search speed. Then, the multi-scale channels and spatial attention modules were constructed in the network to improve the interpretability of feature description and categorical inference. Finally, the stacking method was used to merge the searched network architectures with multiple models to obtain accurate prediction results of classification of benign and malignant lung nodules. Compared with the state-of-the-art lung nodule classification methods, the proposed algorithm has better classification performance and faster convergence on the widely-used lung nodule classification dataset LIDC-IDRI. Moreover, the proposed algorithm has the specificity and precision reached 95.37% and 93.42% respectively, showing it can achieve accurate classification of benign and malignant lung nodules.

Key words: lung nodule classification, neural network architecture search, attention module, multi-model fusion, deep learning

中图分类号: