| 基于深度学习网络的乳腺癌图片分类研究 |
| |
| 引用本文： | 程 年,俞 晨,宁静艳.基于深度学习网络的乳腺癌图片分类研究[J].教育技术导刊,2019,18(8):26-28. |
| |
| 作者姓名： | 程 年 俞 晨 宁静艳 |
| |
| 作者单位： | 1. 东南大学 自动化学院；2. 江苏省肿瘤医院，江苏 南京210096 |
| |
| 摘 要： | 利用深度学习网络对组织病理图片进行分类，以减少病理学家工作量，达到利用计算机辅助治疗的效果。提出利用在两种预训练好的框架下提取的特征进行训练，并研究了多级分类，该成果有利于癌症后期治疗，更加方便临床医学应用；利用迁移学习能够减少训练时间，并解决数据集不足的问题；通过数据增强的方法，可有效提高分类准确度。
|
| 关 键 词： | 深度学习 乳腺癌病理图片 数据增强 迁移学习 |
| 收稿时间： | 2018-12-11 |
Classification of Breast Couldcer Images Based on Deep Learning Network |
| |
| CHENG Nian,YU Cheng,NING Jing-yang.Classification of Breast Couldcer Images Based on Deep Learning Network[J].Introduction of Educational Technology,2019,18(8):26-28. |
| |
| Authors: | CHENG Nian YU Cheng NING Jing-yang |
| |
| Institution: | 1. School of Automated，Southeast University；2. Jiangsu couldcer Hospital，Southeast University，Nangjing 210096，China |
| |
| Abstract: | The purpose was that the method of deep learning was used to classify histopathological images， reducing the heavy work of pathologists and achieving the effect of using computer-assisted medical care. The way that the training of features extracted under two pre-trained frameworks was proposed. At the same time， multiple classes were also studied. Classification could be beneficial to the later treatment of couldcer， which was more convenient for clinical medicine applications. The use of migration learning could preferentially reduce the training time and data set； and the data enhancement method could improve the classification accuracy. |
| |
| Keywords: | deep learning breast couldcer histopathological image data augmentation transfer learning |
|
| 点击此处可从《教育技术导刊》浏览原始摘要信息 |
| 点击此处可从《教育技术导刊》下载免费的PDF全文 |
