Research article Special Issues

A new privacy attack network for remote sensing images classification with small training samples

  • Received: 21 February 2019 Accepted: 05 May 2019 Published: 21 May 2019
  • Solving overfitting problems of privacy attacks on small-sample remote sensing data is still a big challenge in practical application. We propose a new privacy attack network, called joint residual network (JRN), for deep learning based privacy objects classification of small-sample remote sensing images in this paper. Unlike the original residual network structure, which add the bottom feature map to top feature map, JRN fuses the bottom feature map with top feature map by matrix joint. It can reduce the possibility that convolution layers extract the noise of training set or consider the inherent attributes of training set as the whole sample attributes. A series benchmark experiments based on GoogleNet model have been enforced and finally, we compare the model process output and the classification accuracy on small-sample data sets. On the UCMLU data set, the GoogleNet-Feat model which is integrated with JRN is 1.66% higher of accuracy than the original GoogleNet model and 1.87% higher than the GoogleNet-R model; on the WHU-RS dataset, GoogleNet-Feat model is 1.04% higher than the GoogleNet model, and is 3.12% higher than the GoogleNet-R model. Compared with the contrast experiments, the classification accuracy of GoogleNet-Feat is the highest when facing the overfitting problems resulting from the small samples.

    Citation: Eric Ke Wang, Fan Wang, Ruipei Sun, Xi Liu. A new privacy attack network for remote sensing images classification with small training samples[J]. Mathematical Biosciences and Engineering, 2019, 16(5): 4456-4476. doi: 10.3934/mbe.2019222

    Related Papers:

  • Solving overfitting problems of privacy attacks on small-sample remote sensing data is still a big challenge in practical application. We propose a new privacy attack network, called joint residual network (JRN), for deep learning based privacy objects classification of small-sample remote sensing images in this paper. Unlike the original residual network structure, which add the bottom feature map to top feature map, JRN fuses the bottom feature map with top feature map by matrix joint. It can reduce the possibility that convolution layers extract the noise of training set or consider the inherent attributes of training set as the whole sample attributes. A series benchmark experiments based on GoogleNet model have been enforced and finally, we compare the model process output and the classification accuracy on small-sample data sets. On the UCMLU data set, the GoogleNet-Feat model which is integrated with JRN is 1.66% higher of accuracy than the original GoogleNet model and 1.87% higher than the GoogleNet-R model; on the WHU-RS dataset, GoogleNet-Feat model is 1.04% higher than the GoogleNet model, and is 3.12% higher than the GoogleNet-R model. Compared with the contrast experiments, the classification accuracy of GoogleNet-Feat is the highest when facing the overfitting problems resulting from the small samples.


    加载中


    [1] J. Sivic and A. Zisserman, Video Google: A Text Retrieval Approach to Object Matching in Videos, Proc. Ninth International Conf. Computer Vision, 2003.
    [2] S. Lazebnik, C. Schmid and J. Ponce, Beyond bags of features: Spatial pyramid matching for recognizing natural scene categories, Computer vision and pattern recognition, New York, USA, 2006.
    [3] E. K. Wang, Y. P. Li, Y. M. Ye, et al., A Dynamic Trust Framework for Opportunistic Mobile Social Networks, IEEE T. Netw. Serv., 15 (2018), 319–329.
    [4] L. Gueguen, Classifying compound structures in satellite images: A compressed representation for fast queries, IEEE T. Geosci. Remote, 53 (2015), 1803–1818.
    [5] A. S. Razavian, H. Azizpour, J. Sullivan, et al., CNN features off-the-shelf: an astounding base-line for recognition,Proceedings of the IEEE conference on computer vision and pattern recogni-tion workshops. Columbus, OH, USA, (2014), 806–813.
    [6] M. Oquab, L. Bottou, I. Laptev, et al., Learning and transferring mid-level image representations using convolutional neural networks, Proceedings of the IEEE conference on computer vision and pattern recognition. Columbus, OH, USA, 2014.
    [7] R. Girshick, J. Donahue, T. Darrell, et al., Rich feature hierarchies for accurate object detection and semantic segmentation,Proceedings of the IEEE conference on computer vision and pattern recognition. Columbus, OH, USA, 2014.
    [8] J. Deng, W. Dong, R. Socher, et al., Imagenet: A large-scale hierarchical image database, IEEE Conference on Computer Vision and Pattern Recognition, 2009.
    [9] R. Salakhutdinov, J. B. Tenenbaum and A. Torralba, Learning with hierarchical-deep models, IEEE T. Pattern Anal., 35 (2013), 1958–1971.
    [10] F. Hu, G. S. Xia, J. Hu, et al., Transferring deep convolutional neural networks for the scene classification of high-resolution remote sensing imagery, Remote Sens-Basel, 7 (2015), 14680–14707.
    [11] W. Hu, Y. Huang, L. Wei, et al., Deep convolutional neural networks for hyperspectral image classification, J. Sensors, 15 (2015), 1–12.
    [12] F. Zhang, B. Du and L. Zhang, Saliency-guided unsupervised feature learning for scene classifi-cation, IEEE T. Geosci. Remote , 53 (2015), 2175–2184.
    [13] O. Firat, G. Can and F. T. Y. Vural, Representation learning for contextual object and region detection in remote sensing,Conference on Pattern Recognition (ICPR), 2014.
    [14] Y. Bengio, Deep Learning of Representations for Unsupervised and Transfer Learning, Interna-tional Conference on ICML Unsupervised and Transfer Learning, 2012.
    [15] G. Mesnil, Y. Dauphin, X. Glorot, et al., Unsupervised and Transfer Learning Challenge: a Deep Learning Approach, International Conference on ICML Unsupervised and Transfer Learning, 2012.
    [16] Y. LeCun, L. Bottou, G. B. Orr, et al., Efficient backprop.In Neural Networks: Tricks of the Trade, Springer, 1998.
    [17] A. M. Saxe, J. L. McClelland and S. Ganguli, Exact solutions to the nonlinear dynamics of learning in deep linear neural networks, arXiv preprint arXiv:1312.6120, 2013.
    [18] K.He, X.Zhang, S.Ren, et al., Delving deep into rectifiers: Surpassing human-level performance on imagenet classification, IEEE international conference on computer vision, 2015.
    [19] C. Szegedy, W. Liu, Y. Jia, et al., Going deeper with convolutions, IEEE Conference on Computer Vision and Pattern Recognition, 2015.
    [20] J. Wang, L. C. K. Hui, S. M. Yiu, et al., A survey on cyber attacks against nonlinear state estima-tion in power systems of ubiquitous cities, Pervasive Mob. Comput., 39 (2017), 10–17.
    [21] K. Chatfield, K. Simonyan, A. Vedaldi, et al., Return of the devil in the details: Delving deep into convolutional nets, arXiv preprint arXiv:1405.3531, 2014.
    [22] M. D. Zeiler and R. Fergus, Visualizing and understanding convolutional networks,European conference on computer vision, 2014.
    [23] S. Ioffe and C. Szegedy, Batch normalization: Accelerating deep network training by reducing internal covariate shift, International Conference on Machine Learning, 2015.
    [24] K. He, X. Zhang, S. Ren, et al., Deep residual learning for image recognition,IEEE conference on computer vision and pattern recognition, 2016.
    [25] Y. Yang and S. Newsam, Bag-of-visual-words and spatial extensions for land-use classification, Proceedings of the 18th SIGSPATIAL international conference on advances in geographic infor-mation systems. ACM, San Jose, CA, USA, (2010), 270–279.
    [26] G. S. Xia, W. Yang, J. Delon, et al., Structural high-resolution satellite image indexing, ISPRS TC VII Symposium-100 Years ISPRS, 2010.
    [27] O. Russakovsky, J. Deng, H. Su, et al., Imagenet large scale visual recognition challenge, Int. J. Comput. Vision, 115 (2017), 211–252.
    [28] C. M. Chen, B. Xiang, Y. Liu, et al., A Secure Authentication Protocol for Internet of Vehicles, IEEE Access, 7 (2019), 12047–12057.
    [29] K. H. Wang, C. M. Chen, W. C. Fang, et al., On the security of a new ultra-lightweight authenti-cation protocol in IoT environment for RFID tags, J. Supercomput., 74 (2018), 65–70.
    [30] E. Wang, Y. Li, Z. Nie, et al., Deep Fusion Feature Based Object Detection Method for High Resolution Optical Remote Sensing Images, Appl. Sci., 9 (2019), 1130–1148.
    [31] A. Karati, S. H. Islam and M. Karuppiah, Provably Secure and Lightweight Certificateless Sig-nature Scheme for IIoT Environments, IEEE T. Ind. Inform, 18 (2018), 1–8.
    [32] J. Guan and E. Wang, Repeated review based image captioning for image evidence review, Signal Process-Image, 63 (2018), 141–148.
  • Reader Comments
  • © 2019 the Author(s), licensee AIMS Press. This is an open access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0)
通讯作者: 陈斌, bchen63@163.com
  • 1. 

    沈阳化工大学材料科学与工程学院 沈阳 110142

  1. 本站搜索
  2. 百度学术搜索
  3. 万方数据库搜索
  4. CNKI搜索

Metrics

Article views(4127) PDF downloads(526) Cited by(8)

Article outline

Figures and Tables

Figures(14)  /  Tables(7)

Other Articles By Authors

/

DownLoad:  Full-Size Img  PowerPoint
Return
Return

Catalog