《计算机应用研究》|Application Research of Computers

基于邻域差分滤波生成式对抗网络的数据增强方法

Data augmentation based on neighboring difference filtering generative adversarial network

免费全文下载 (已被下载 次)  
获取PDF全文
作者 杜卉然,许亮,吕帅
机构 广东工业大学 自动化学院,广州 510006
统计 摘要被查看 次,已被下载
文章编号 1001-3695(2020)06-061-1895-05
DOI 10.19734/j.issn.1001-3695.2019.02.0031
摘要 针对工业产品样本缺乏且特征不明显而难以用于深度学习训练的问题,提出一种邻域差分滤波生成式对抗网络数据增强(NDF-GAN)方法。将邻域差分滤波器融合到生成对抗网络中,从样本中提取特征并进行样本重建,对样本进行数据增强。实验表明,该方法所生成的样本比现有两种模型质量更高,与真实样本混合训练分类模型后获得更好的分类性能。因此,提出的NDF-GAN实现了对工业产品样本的数据增强。
关键词 生成式对抗网络; 邻域差分; 特征提取; 数据增强
基金项目 国家自然科学基金资助项目(21376091)
本文URL http://www.arocmag.com/article/01-2020-06-061.html
英文标题 Data augmentation based on neighboring difference filtering generative adversarial network
作者英文名 Du Huiran, Xu Liang, Lyu Shuai
机构英文名 School of Automation,Guangdong University of Technology,Guangzhou 510006,China
英文摘要 Deep learning is hard to train with industrial product samples due to their features are non-obvious and amount is extremely rare. This paper proposed a neighboring difference filtering generative adversarial network(NDF-GAN). The combination of NDF-GAN extracted more features from real data and generated more samples, which could augment the real data. Experimental results reveal that, the proposed method can effectively provide more new samples to train the classifier and improve classification performance. In conclusion, NDF-GAN can augment industrial product samples.
英文关键词 generative adversarial network; neighboring difference filtering; feature extraction; data augmentation
参考文献 查看稿件参考文献
 
收稿日期 2019/2/16
修回日期 2019/4/8
页码 1895-1899,1905
中图分类号 TP391.41
文献标志码 A