无监督混阶栈式稀疏自编码器的图像分类学习

doi:10.11772/j.issn.1001-9081.2019061107

计算机应用 ›› 2019, Vol. 39 ›› Issue (12): 3420-3425.DOI: 10.11772/j.issn.1001-9081.2019061107

• 第十七届中国机器学习会议(CCML 2019)论文 • 上一篇下一篇

无监督混阶栈式稀疏自编码器的图像分类学习

杨东海^1,2, 林敏敏^1,2, 张文杰^1,2, 杨敬民^1,2

1. 闽南师范大学计算机学院, 福建漳州 363000;
2. 福建省粒计算及其应用重点实验室(闽南师范大学), 福建漳州 363000

收稿日期:2019-04-29 修回日期:2019-06-25 发布日期:2019-08-26 出版日期:2019-12-10
作者简介:杨东海(1988-),男,福建漳州人,硕士研究生,CCF会员,主要研究方向:计算机视觉;林敏敏(1994-),女,福建州人,硕士研究生,CCF会员,主要研究方向:机器学习、无线通信;张文杰(1984-),男,福建漳州人,博士,教授,CCF会员,主要研究方向:认知无线电、计算机网络体系结构;杨敬民(1980-),男,福建漳州人,硕士,副教授,CCF会员,主要研究方向:认知无线电、计算机网络体系结构、机器学习。
基金资助:
国家自然科学基金资助项目（61701213）；福建省省属高校科研专项资助项目（JK2017031）；教育部产学合作协同育人项目（201702098015，201702057020）；漳州市自然科学基金资助项目（ZZ2018J21）。

Image classification learning via unsupervised mixed-order stacked sparse autoencoder

YANG Donghai^1,2, LIN Minmin^1,2, ZHANG Wenjie^1,2, YANG Jingmin^1,2

1. School of Computer Science, Minnan Normal University, Zhangzhou Fujian 363000, China;
2. Fujian Key Laboratory of Granular Computing and Application(Minnan Normal University), Zhangzhou Fujian 363000, China

Received:2019-04-29 Revised:2019-06-25 Online:2019-08-26 Published:2019-12-10
Contact: 杨敬民
Supported by:
This work is partially supported by the National Natural Science Foundation of China (61701213), the Special Research Fund for Higher Education of Fujian (JK2017031), the Cooperative Education Project of Ministry of Education (201702098015, 201702057020), the Natural Science Foundation of Zhangzhou (ZZ2018J21).

摘要/Abstract

摘要： 目前多数图像分类的方法是采用监督学习或者半监督学习对图像进行降维，然而监督学习与半监督学习需要图像携带标签信息。针对无标签图像的降维及分类问题，提出采用混阶栈式稀疏自编码器对图像进行无监督降维来实现图像的分类学习。首先，构建一个具有三个隐藏层的串行栈式自编码器网络，对栈式自编码器的每一个隐藏层单独训练，将前一个隐藏层的输出作为后一个隐藏层的输入，对图像数据进行特征提取并实现对数据的降维。其次，将训练好的栈式自编码器的第一个隐藏层和第二个隐藏层的特征进行拼接融合，形成一个包含混阶特征的矩阵。最后，使用支持向量机对降维后的图像特征进行分类，并进行精度评价。在公开的四个图像数据集上将所提方法与七个对比算法进行对比实验，实验结果表明，所提方法能够对无标签图像进行特征提取，实现图像分类学习，减少分类时间，提高图像的分类精度。

关键词: 无监督学习, 栈式自编码器, 降维, 混阶特征, 图像分类

Abstract: Most of the current image classification methods use supervised learning or semi-supervised learning to reduce image dimension. However, supervised learning and semi-supervised learning require image carrying label information. Aiming at the dimensionality reduction and classification of unlabeled images, a mixed-order feature stacked sparse autoencoder was proposed to realize the unsupervised dimensionality reduction and classification learning of the images. Firstly, a serial stacked sparse autoencoder network with three hidden layers was constructed. Each hidden layer of the stacked sparse autoencoder was trained separately, and the output of the former hidden layer was used as the input of the latter hidden layer to realize the feature extraction of image data and the dimensionality reduction of the data. Secondly, the features of the first hidden layer and the second hidden layer of the trained stacked autoencoder were spliced and fused to form a matrix containing mixed-order features. Finally, the support vector machine was used to classify the image features after dimensionality reduction, and the accuracy was evaluated. The proposed method was compared with seven comparison algorithms on four open image datasets. The experimental results show that the proposed method can extract features from unlabeled images, realize image classification learning, reduce classification time and improve image classification accuracy.

Key words: unsupervised learning, stacked sparse autoencoder, dimensionality reduction, mixed-order feature, image classification

中图分类号:

TP391.41

杨东海, 林敏敏, 张文杰, 杨敬民. 无监督混阶栈式稀疏自编码器的图像分类学习[J]. 计算机应用, 2019, 39(12): 3420-3425.

YANG Donghai, LIN Minmin, ZHANG Wenjie, YANG Jingmin. Image classification learning via unsupervised mixed-order stacked sparse autoencoder[J]. Journal of Computer Applications, 2019, 39(12): 3420-3425.

参考文献

[1] LEE J, CHOE Y. Robust PCA based on incoherence with geometrical interpretation[J]. IEEE Transactions on Image Processing, 2018, 27(4):1939-1950.
[2] LIU C, JIN T, HOI S C H, et al. Collaborative topic regression for online recommender systems:an online and Bayesian approach[J]. Machine Learning, 2017, 106(5):651-670.
[3] RAFAILIDIS D, MANOLOPOULOU S, DARAS P. A unified framework for multimodal retrieval[J]. Pattern Recognition, 2013, 46(12):3358-3370.
[4] JIN X, HAN H, DAI Q. Plenoptic image coding using macropixel-based intra prediction[J]. IEEE Transactions on Image Processing, 2018, 27(8):3954-3968.
[5] BERRY T, HARLIM J. Iterated diffusion maps for feature identification[J]. Applied and Computational Harmonic Analysis, 2018, 45(1):84-119.
[6] 焦斌亮,张可.基于SPE的无线传感器网络定位算法[J].小型微型计算机系统,2013,34(2):269-271.(JIAO B L, ZHANG K. Localization algorithm based on stochastic proximity embedding in wireless sensor networks[J]. Journal of Chinese Computer Systems, 2013, 34(2):269-271.)
[7] ZABALZA J, REN J, ZHENG J, et al. Novel segmented stacked autoencoder for effective dimensionality reduction and feature extraction in hyperspectral imaging[J]. Neurocomputing, 2016, 185:1-10.
[8] FAN Z, BI D, HE L, et al. Low-level structure feature extraction for image processing via stacked sparse denosing autoencoder[J]. Neurocomputing, 2017, 243:12-20.
[9] 张春雨,韩立新,徐守晶.基于栈式自动编码的图像哈希算法[J].电子测量技术,2016,39(3):46-49,69.(ZHANG C Y, HAN L X, XU S J. Image hashing algorithm based on stacked autoencoder[J]. Electronic Measurement Technology, 2016, 39(3):46-49, 69.)
[10] ZHANG Y, THING V L L. A semi-feature learning approach for tampered region localization across multi-format images[J]. Multimedia Tools and Applications, 2018, 77(19):25027-25052.
[11] MAO K M, TANG R J, WANG X Q, et al. Feature representation using deep autoencoder for lung nodule image classification[J]. Complexity, 2018, 2018:Article ID 3078374.
[12] DAI Y, WANG G. Analyzing tongue images using a conceptual alignment deep autoencoder[J]. IEEE Access, 2018, 6:5962-5972.
[13] ZHAO G, WANG X, NIU Y, et al. Segmenting brain tissues from chinese visible human dataset by deep-learned features with stacked autoencoder[J]. BioMed Research International, 2016, 2016:Article ID 5284586.
[14] 杨文元.多标记学习自编码网络无监督维数约简[J].智能系统学报,2018,13(5):808-817.(YANG W Y.Unsupervised dimensionality reduction via autoencoder networks[J]. CAAI Transactions on Intelligent Systems, 2018, 13(5):808-817.)
[15] GUAN R, WANG X, YANG M Q, et al. Multi-label deep learning for gene function annotation in cancer pathways[J]. Scientific Reports, 2018, 8(1):Article ID 267.
[16] 徐培,蔡小路,何文伟,等.基于深度自编码网络的运动目标检测[J].计算机应用,2014,34(10):2934-2937,2962.(XU P, CAI X L, HE W W, et al. Motion detection based on deep auto-encoder networks[J]. Journal of Computer Applications, 2014, 34(10):2934-2937, 2962.)
[17] ZHENG W, GONG S, XIANG T. Towards open-world person re-identification by one-shot group-based verification[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2016, 38(3):591-606.
[18] BRERETON R G, LLOYD G R. Support vector machines for classification and regression[J]. The Analyst, 2010, 135(2):230-267.
[19] VAN DER MAATEN L. Accelerating t-SNE using tree-based algorithms[J]. Journal of Machine Learning Research, 2014, 15(1):3221-3245.
[20] YUAN S, WU X, XIANG Y. SNE:Signed Network Embedding[C]//Proceedings of the 2017 Pacific-Asia Conference on Knowledge Discovery and Data Mining, LNCS 10235. Cham:Springer, 2017:183-195.
[21] ZHANG S, WANG J, TAO X, et al. Constructing deep sparse coding network for image classification[J]. Pattern Recognition, 2017, 64:130-140.
[22] OTHMAN E, BAZI Y, ALAJLAN N, et al. Using convolutional features and a sparse autoencoder for land-use scene classification[J]. International Journal of Remote Sensing, 2016, 37(10):2149-2167.
[23] 沈承恩,何军,邓扬. 基于改进堆叠自动编码机的垃圾邮件分类[J].计算机应用,2016,36(1):158-162.(SHEN C E, HE J, DENG Y. Spam filtering based on modified stack auto-encoder[J]. Journal of Computer Applications, 2016, 36(1):158-162.)
[24] BADEM H, CALISKAN A, BASTURK A, et al. Classification of human activity by using a stacked autoencoder[C]//Proceedings of the 2017 Medical Technologies National Congress. Piscataway:IEEE, 2017:1-4.
[25] 杨帅,王鹃.基于堆栈降噪自编码器改进的混合推荐算法[J].计算机应用,2018,38(7):1866-1871.(YANG S, WANG J. Improved hybrid recommendation algorithm based on stacked denoising autoencoder[J]. Journal of Computer Applications, 2018, 38(7):1866-1871.)

无监督混阶栈式稀疏自编码器的图像分类学习

Image classification learning via unsupervised mixed-order stacked sparse autoencoder

PDF

可视化

摘要/Abstract

引用本文

使用本文

参考文献

相关文章 15

编辑推荐

Metrics

[1]	贾鹤鸣, 郎春博, 姜子超. 基于轻量级卷积神经网络的植物叶片病害识别方法[J]. 计算机应用, 2021, 41(6): 1812-1819.
[2]	刘睿珩, 叶霞, 岳增营. 面向自然语言处理任务的预训练模型综述[J]. 《计算机应用》唯一官方网站, 2021, 41(5): 1236-1246.
[3]	郭一村, 陈华辉. 在线哈希算法研究综述[J]. 计算机应用, 2021, 41(4): 1106-1112.
[4]	张阳, 王小宁. 基于Word2Vec词嵌入和高维生物基因选择遗传算法的文本特征选择方法[J]. 《计算机应用》唯一官方网站, 2021, 41(11): 3151-3155.
[5]	许学斌, 张佳达, 刘伟, 路龙宾, 赵雨晴. 融合空间和通道特征的高精度乳腺癌分类方法[J]. 计算机应用, 2021, 41(10): 3025-3032.
[6]	郑延斌, 韩梦云, 樊文鑫. 基于二维主成分分析与卷积神经网络的手写体汉字识别[J]. 计算机应用, 2020, 40(8): 2465-2471.
[7]	汪志远, 降爱莲, 奥斯曼·穆罕默德. 基于正则互表示的无监督特征选择方法[J]. 计算机应用, 2020, 40(7): 1896-1900.
[8]	边小勇, 江沛龄, 赵敏, 丁胜, 张晓龙. 基于多分支神经网络模型的弱监督细粒度图像分类方法[J]. 计算机应用, 2020, 40(5): 1295-1300.
[9]	吴立人, 刘政浩, 张浩, 岑悦亮, 周维. 聚焦图像对抗攻击算法PS-MIFGSM[J]. 计算机应用, 2020, 40(5): 1348-1353.
[10]	徐戈, 肖永强, 汪涛, 陈开志, 廖祥文, 吴运兵. 基于视觉误差与语义属性的零样本图像分类[J]. 计算机应用, 2020, 40(4): 1016-1022.
[11]	王锦凯, 贾旭. 基于加权正交约束非负矩阵分解的车脸识别算法[J]. 计算机应用, 2020, 40(4): 1050-1055.
[12]	郭志强, 胡永武, 刘鹏, 杨杰. 基于特征融合的室外天气图像分类[J]. 计算机应用, 2020, 40(4): 1023-1029.
[13]	张凯琳, 阎庆, 夏懿, 章军, 丁云. 基于焦点损失的半监督高光谱图像分类[J]. 计算机应用, 2020, 40(4): 1030-1037.
[14]	李东博, 黄铝文. 重加权稀疏主成分分析算法及其在人脸识别中的应用[J]. 计算机应用, 2020, 40(3): 717-722.
[15]	张盼盼, 李其申, 杨词慧. 基于轻量级分组注意力模块的图像分类算法[J]. 计算机应用, 2020, 40(3): 645-650.