%0 Journal Article
%T 基于机器学习的肺结节良恶性分类研究进展<br>Advances in the Classification of Benign and Malignant Pulmonary Nodules Based on Machine Learning
%A 杨杨
%A 李晓琴
%A 侯智超
%A 高斌
%J Biophysics
%P 43-56
%@ 2330-1694
%D 2021
%I Hans Publishing
%R 10.12677/BIPHY.2021.92006
%X 简要分析了计算机断层扫描(computed tomography, CT)与其他肺结节成像方式的优缺点。随后介绍了目前研究者们运用最多的两个肺部计算机断层扫描影像数据库以及其文件组成。最后从传统人工特征、深度学习、数据集的选择以及多分类这四个角度出发，重点介绍了肺结节良恶性分类的具体应用及进展，并加以讨论和展望。<br />
The advantages and disadvantages of computed tomography (CT) and other imaging methods of pulmonary nodules are analyzed. Then, the two CT image databases and their file composition that researchers use most at present are introduced. Finally, from the four perspectives of traditional artificial features, deep learning, database selection and multi-classification, the specific application and progress of benign and malignant classification of pulmonary nodules are mainly discussed and prospected.
%K 机器学习，深度学习，计算机辅助诊断，肺结节，金标准，结节分类
%K Machine Learning
%K Deep Learning
%K Computer Aided Diagnosis
%K Pulmonary Nodule
%K Golden Standard
%K Nodule Classification
%U http://www.hanspub.org/journal/PaperInformation.aspx?PaperID=42147