摘要
该研究运用生成对抗网络替代传统的数据增强方法进行糖网病图像生成,为解决糖网病分类训练数据不平衡问题提供了新思路。分别使用仿射变换方法和生成对抗网络进行数据增强,并将增强后的图像数据用于糖网病分类,结果证明该研究所用方法具有更高的准确率和更快的收敛速度。
The generative adversarial network is used to replace traditional data enhancement methods for diabetic retinopathy image generation,which provides a new idea to solve the problem of unbalanced training data of diabetic retinopathy classification.In this paper,the affine transformation method and the generative adversarial network are used for data enhancement,and the enhanced image are applied to the classification of diabetic retinopathy.The results show that the method used in this study has higher accuracy and faster convergence speed.
作者
余莹
刘颖
章浩伟
YU Ying;LIU Ying;ZHANG Haowei(School of Health Science and Engineering,University of Shanghai for Science and Technology,Shanghai,200093)
出处
《生物医学工程学进展》
CAS
2021年第4期187-190,共4页
Progress in Biomedical Engineering
基金
微创励志创新基金资助项目(182702173)。
关键词
糖网病
生成对抗网络
数据增强
diabetic retinopathy
generative adversarial network
data enhancement
