浏览全部资源
扫码关注微信
西北工业大学 电子信息学院, 西安 710072
[ "徐 哲(1998-),男,陕西汉中人,硕士研究生,2020年于西北工业大学获得学士学位,主要从事深度学习与图像处理的研究。Email: xuzhenwpu@126.com" ]
[ "耿 杰(1990-),男,山西晋中人,副教授,硕士生导师,2013年于大连理工大学获得学士学位,2018年于大连理工大学获得博士学位,主要从事深度学习与图像处理的研究,Email: gengjie@nwpu.edu.cn" ]
收稿日期:2020-11-04,
修回日期:2021-01-04,
纸质出版日期:2021-05-15
移动端阅览
徐哲,耿杰,蒋雯等.联合训练生成对抗网络的半监督分类方法[J].光学精密工程,2021,29(05):1127-1135.
XU Zhe,GENG Jie,JIANG Wen,et al.Co-training generative adversarial networks for semi-supervised classification method[J].Optics and Precision Engineering,2021,29(05):1127-1135.
徐哲,耿杰,蒋雯等.联合训练生成对抗网络的半监督分类方法[J].光学精密工程,2021,29(05):1127-1135. DOI: 10.37188/OPE.20212905.1127.
XU Zhe,GENG Jie,JIANG Wen,et al.Co-training generative adversarial networks for semi-supervised classification method[J].Optics and Precision Engineering,2021,29(05):1127-1135. DOI: 10.37188/OPE.20212905.1127.
深度神经网络需要大量数据进行监督训练学习,而实际应用中往往难以获取大量标签数据。半监督学习可以减小深度网络对标签数据的依赖,基于半监督学习的生成对抗网络可以提升分类效果,但仍存在训练不稳定的问题。为进一步提高网络的分类精度并解决网络训练不稳定的问题,本文提出一种基于联合训练生成对抗网络的半监督分类方法,通过两个判别器的联合训练来消除单个判别器的分布误差,同时选取无标签数据中置信度高的样本来扩充标签数据集,提高半监督分类精度并提升网络模型的泛化能力。在CIFAR-10和SVHN数据集上的实验结果表明,本文方法在不同数量的标签数据下都获得更好的分类精度。当标签数量为2 000时,在CIFAR-10数据集上分类精度可达80.36%;当标签数量为10时,相比于现有的半监督方法,分类精度提升了约5%。在一定程度上解决了GAN网络在小样本条件下的过拟合问题。
Deep neural networks require a large amount of data for supervised learning; however, it is difficult to obtain enough labeled data in practical applications. Semi-supervised learning can train deep neural networks with limited samples. Semi-supervised generative adversarial networks can yield superior classification performance; however, they are unstable during training in classical networks. To further improve the classification accuracy and solve the problem of training instability for networks, we propose a semi-supervised classification model called co-training generative adversarial networks (CT-GAN) for image classification. In the proposed model, co-training of two discriminators is applied to eliminate the distribution error of a single discriminator and unlabeled samples with higher confidence are selected to expand the training set, which can be utilized for semi-supervised classification and enhance the generalization of deep networks. Experimental results on the CIFAR-10 dataset and the SVHN dataset showed that the proposed method achieved better classification accuracies with different numbers of labeled data. The classification accuracy was 80.36% with 2000 labeled data on the CIFAR-10 dataset, whereas it improved by about 5% compared with the existing semi-supervised method with 10 labeled data. To a certain extent, the problem of GAN overfitting under a few sample conditions is solved.
WANG W , YANG Y , WANG X , et al . Development of convolutional neural network and its application in image classification: a survey [J]. Optical Engineering , 2019 , 58 ( 4 ): 040901 .
LOWE D G . Distinctive image features from scale-invariant keypoints [J]. International journal of computer vision , 2004 , 60 ( 2 ): 91 - 110 .
DALAL N , TRIGGS B . Histograms of oriented gradients for human detection [C]. 2005 IEEE computer society conference on computer vision and pattern recognition (CVPR' 05 ). IEEE , 2005, 1 : 886 - 893 .
HUANG F J , LECUN Y . Large-scale learning with svm and convolutional netw for generic object recognition [C]. 2006 IEEE Computer Society Conference on Computer Vision and Pattern Recognition , 2006 , 10 .
陈科峻 , 张叶 . 循环神经网络多标签航空图像分类 [J]. 光学 精密工程 , 2020 , 28 ( 6 ): 1404 - 1413 .
CHEN K J , ZHANG Y . Recurrent neural network multi-label aerial images classification [J]. Opt. Precision Eng. , 2020 , 28 ( 6 ): 1404 - 1413 . (in Chinese)
黄鸿 , 唐玉枭 , 段宇乐 . 半监督多图嵌入的高光谱影像特征提取 [J]. 光学 精密工程 , 2020 , 28 ( 2 ): 443 - 456 .
HUANG H , TANG Y X , DUAN Y L . Feature extraction of hyperspectral image with semi-supervised multi-graph embedding [J]. Opt. Precision Eng. , 2020 , 28 ( 2 ): 443 - 456 . (in Chinese)
赵浩光 , 王平 , 董超 , 等 . 结合多尺度视觉显著性的舰船目标检测 [J]. 光学 精密工程 , 2020 , 28 ( 6 ): 1395 - 1403 .
ZHAO H G , WANG P , DONG CH , et al . Ship detection based on the multi-scale visual saliency model [J]. Opt. Precision Eng. , 2020 , 28 ( 6 ): 1395 - 1403 . (in Chinese)
陈炜炜 , 严群 , 姚剑敏 . 基于AttentionGAN和形态学重建的TFT阵列缺陷检测 [J]. 液晶与显示 , 2020 , 35 ( 12 ): 1270 - 1277 .
CHEN W W , YAN Q , YAO J M . TFT array defect detection based on AttentionGAN and morphological [J]. Chinese Journal of Liquid Crystals and Displays , 2020 , 35 ( 12 ): 1270 - 1277 . (in Chinese)
GOODFELLOW I , POUGET-ABADIE J , MIRZA M , et al . Generative adversarial nets [C]. Advances in neural information processing systems , 2014 : 2672 - 2680 .
RADFORD , ALEC , METZLUKE , et al . Unsupervised representation learning with deep convolutional generative adversarial networks . arXiv preprint arXiv : 1511 . 06434 ( 2015 ).
ARJOVSKY M , CHINTALA S , BOTTOU L . Wasserstein gan [J]. arXiv preprint arXiv: 1701.07875 , 2017 .
SHAHAM T R , DEKEL T , MICHAELI T . Singan: Learning a generative model from a single natural image [C]. Proceedings of the IEEE International Conference on Computer Vision , 2019 : 4570 - 4580 .
KARNEWAR A , WANG O . Msg-gan: Multi-scale gradients for generative adversarial networks [C]. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , 2020 : 7799 - 7808 .
ZHANG H , GOODFELLOW I , METAXAS D , et al . Self-attention generative adversarial networks [C]. International conference on machine learning. PMLR , 2019 : 7354 - 7363 .
ODENA A . Semi-supervised learning with generative adversarial networks [J]. arXiv preprint arXiv: 1606.01583 , 2016 .
LI W , WANG Z , LI J , et al . Semi-supervised learning based on generative adversarial network: a comparison between good GAN and bad GAN approach [C]. CVPR Workshops , 2019 .
Salimans T , Goodfellow I , Zaremba W , et al . Improved techniques for training gans [C]. Advances in neural information processing systems , 2016 : 2234 - 2242 .
0
浏览量
474
下载量
4
CSCD
关联资源
相关文章
相关作者
相关机构