Page 290 - 《软件学报》2021年第9期

2914 Journal of Software 软件学报 Vol.32, No.9, September 2021

4 结束语

本文针对现有空域隐写分析方法在低嵌入率下难以区分的问题,通过分析现有基于卷积神经网络的隐写
分析方法的特点,构造了一个新的网络结构 Shen-Net.实验结果证明:新提出的网络结构在对 WOW,S-
UNIWARD 和 HILL 这 3 种常见空域内容自适应隐写算法进行隐写分析时,准确率得到了较高的提升.在嵌入率
较低的情况下,现有网络结构无法收敛或准确性很低,而本文设计的网络结构仍能够取得较为理想的检测准确
率.此外,本文还通过采用逐步迁移学习的方法进一步提高了对低嵌入容量的检测准确率.由于在现实生活中,
JPEG 格式图像的使用更为常见,下一步我们将对 JPEG 格式图像的隐写分析方法进行深入研究.

References:
[1] Huang W, Zhao XF, Feng DG, Sheng RN. JPEG steganalysis based on feature fusion by principal component analysis. Ruan Jian
Xue Bao/Journal of Software, 2012,23(7):1869−1879 (in Chinese with English abstract). http://www.jos.org.cn/1000-9825/4107.
htm [10.3724/SP.J.1001.2012.04107]
[2] Zhang YW, Zhang WM, Yu NH. Specific testing sample steganalysis. Ruan Jian Xue Bao/Journal of Software, 2018,29(4):
987−1001 (in Chinese with English abstract). http://www.jos.org.cn/1000-9825/5411.htm [doi: 10.13328/j.cnki.jos.005411]
[3] Huang DZ, Zhang JF, Zhang R, Li PC, Guo YB. New system of multi-modal information hiding based on big data environment.
Journal of Electronic, 2017,45(2):477−484 (in Chinese with English abstract).
[4] Tang GM, Sun Y, Xu XY, Wang Y. Adaptive JPEG steganography based on distortion cost updating. Journal on Communications,
2017,38(9):1−8 (in Chinese with English abstract).
[5] Holub V, Fridrich J. Designing steganographic distortion using directional filters. In: Proc. of the IEEE Int’l Workshop on
Information Forensics and Security. IEEE, 2012. 234−239.
[6] Holub V, Fridrich J, Denemark T. Universal distortion function for steganography in an arbitrary domain. EURASIP Journal on
Information Security, 2014,2014(1):1.
[7] Li B, Wang M, Huang J, Li X. A new cost function for spatial image steganography. In: Proc. of the IEEE Int’l Conf. on Image
Processing. Paris: IEEE, 2014. 4206−4210.
[8] Fridrich J, Kodovsky J. Rich models for steganalysis of digital images. IEEE Trans. on Information Forensics and Security, 2012,
7(3):868−882.
[9] Zhai LM, Jia J, Ren WX, Xu YB, Wang LN. Recent advances in deep learning for image steganography and steganalysis. Journal
of Information Security, 2018,3(6):2−12 (in Chinese with English abstract).
[10] Pibre L, Pasquet J, Ienco D. Deep learning is a good steganalysis tool when embedding key is reused for different images, even if
there is a cover sourcemismatch. Electronic Imaging, 2016,2016(8):1−11.
[11] Salomon M, Couturier R, Guyeux C, Coucho JF, Bahi JM. Steganalysis via a convolutional neural network using large convolution
filters for embedding process with same stego key: A deep learning approach for telemedicine. European Research in Telemedicine/
La Recherche Européenne en Télémédecine, 2017,6(2):79−92.
[12] Pevný T, Filler T, Bas P. Using high-dimensional image models to perform highly undetectable steganography. In: Proc. of the Int’l
Workshop on Information Hiding. Berlin: Springer-Verlag, 2010. 161−177.
[13] Gao PX, Wei LX, Liu J, Liu MM. Image steganalysis based on convolution neural network. Application Research of Computers,
2019,36(1):54−288 (in Chinese with English abstract).
[14] Xu G, Wu HZ, Shi YQ. Structural design of convolutional neural networks for steganalysis. IEEE Signal Processing Letters, 2016,
23(5):708−712.
[15] Tan S, Li B. Stacked convolutional auto-encoders for steganalysis of digital images. In: Proc. of the Signal and Information
Processing Association Annual Summit and Conf. Chiang Mai: Asia-Pacific, 2014. 1−4.
[16] Qian Y, Dong J, Wang W, Tan T. Deep learning for steganalysis via convolutional neural networks. In: Alattar AM, et al, eds. Proc.
of SPIE-IS&T, 2015. 9409: 94090J.
[17] Qian Y, Dong J, Wang W, Tan T. Learning and transferring representations for image steganalysis using convolutional neural
network. In: Proc. of the IEEE Int’l Conf. on Image Processing. 2016. 2752−2756.