利用模块化残差网络的图像隐写分析

doi:10.19665/j.issn1001-2400.2019.01.013

摘要/Abstract

摘要：

为了提高图像隐写分析方法对小嵌入率隐写术检测的准确性,针对小嵌入率隐写术提出一种基于高度模块化网络结构的图像隐写分析方法。首先,通过重复残差网络单元来构建基础网络模型,以提取数字图像中的复杂统计特性;其次,增加分组卷积以提取残差图像通道信息,加强来自隐写信息的信号特征;最后,利用大量数据集对网络进行训练,得到了基于模块化残差网络的图像隐写分析方法。实验结果表明,所提方法相较于现有算法可以提取更有效的图像特征,从而得到更好的检测效果。同时,利用残差网络块作为模板,可以很容易地搭建网络模型,便于网络的调整和训练。

关键词: 隐写分析, 残差网络, 分组卷积, 模块化, 低嵌入率

Abstract:

In order to improve the detection accuracy of small embedding rate steganography, an image steganalysis method based on the highly modularized convolutional neural network is proposed. First, the fundamental network is built by repeating residual network units to extract the complex statistical properties of digital images. Then, extracting the channel information on the residual image by adding the group convolution, it is very good to strengthen the signal characteristics from the hidden information. Finally, a large number of datasets are used to train the network, and the image steganalysis method based on the modularized residual network is obtained. Experimental results show that compared with the existing methods, the proposed method has a better performance, and extracts more effective image features. Meanwhile, using the residual network module as the template, the network model can be easily built to facilitate adjustment and training.

Key words: steganalysis, residual network, group convolution, modularized, low embedding rate

中图分类号:

TP391.4

郭继昌,何艳红,魏慧文. 利用模块化残差网络的图像隐写分析[J]. 西安电子科技大学学报, 2019, 46(1): 79-85.

GUO Jichang,HE Yanhong,WEI Huiwen. Image steganalysis based on the modularized residual network[J]. Journal of Xidian University, 2019, 46(1): 79-85.

图/表 7

图1

图2

图3

表1

文中网络参数设置"

卷积层	输出尺寸	卷积核大小	特征维数
Conv1	128×128	$1 × 1,64 3 × 3,64, C = 4 1 × 1,64$ ×2	64
Conv2	64×64	$1 × 1,64 3 × 3,64, C = 4 1 × 1,128$ ×2	128
Conv3	32×32	$1 × 1,64 3 × 3,64, C = 4 1 × 1,256$ ×2	256

表1

图4

表2

表3

参考文献 15

[1]	FRIDRICH J, KODOVSKY J . Rich Models for Steganalysis of Digital Images[J]. IEEE Transactions on Information Forensics and Security, 2012,7(3):868-882. doi: 10.1109/TIFS.2012.2190402
[2]	QIAN Y L, DONG J, WANG W, et al. Deep Learning for Steganalysis via Convolutional Neural Networks [C]//Proceedings of the SPIE: 9409. Bellingham: SPIE, 2015: 94090J.
[3]	XU G S, WU H Z, SHI Y Q . Structural Design of Convolutional Neural Networks for Steganalysis[J]. IEEE Signal Processing Letters, 2016,23(5):708-712. doi: 10.1109/LSP.2016.2548421
[4]	WU S T, ZHONG S H, LIU Y . Deep Residual Learning for Image Steganalysis[J]. Multimedia Tools and Applications, 2018,77(9):10437-10453. doi: 10.1007/s11042-017-4440-4
[5]	CAI B L, XU X M, JIA K , et al. Dehazenet: An End-to-end System for Single Image Haze Removal[J]. IEEE Transactions on Image Processing, 2016,25(11):5187-5198. doi: 10.1109/TIP.2016.2598681 pmid: 28873058
[6]	HE K M, ZHANG X Y, REN S Q, et al. Deep Residual Learning for Image Recognition [C]//Proceedings of the 2016 IEEE Computer Society Conference on Computer Vision and Pattern Recognition. Washington: IEEE Computer Society, 2016: 770-778.
[7]	YE J, NI J, YI Y . Deep Learning Hierarchical Representations for Image Steganalysis[J]. IEEE Transactions on Information Forensics and Security, 2017,12(11):2545-2557. doi: 10.1109/TIFS.2017.2710946
[8]	KODOVSKY J, FRIDRICH J, HOLUB V . Ensemble Classifiers for Steganalysis of Digital Media[J]. IEEE Transactions on Information Forensics and Security, 2012,7(2):432-444. doi: 10.1109/TIFS.2011.2175919
[9]	XIE S N, GIRSHICK R, DOLLAR P, et al. Aggregated Residual Transformations for Deep Neural Networks [C]//Proceedings of the 2017 30th IEEE Conference on Computer Vision and Pattern Recognition. Piscataway: IEEE, 2017: 5987-5995.
[10]	刘逸, 慕彩红, 刘敬 . 结合邻域信息粒子群聚类用于 SAR 图像变化检测[J]. 西安电子科技大学学报, 2015,42(1):187-193. doi: 10.3969/j.issn.1001-2400.2015.01.030
	LIU Yi, MU Caihong, LIU Jing . Change Detection for SAR Images Based on the Particle Swarm Clustering Algorithm Using Neighborhood in Formation[J]. Journal of Xidian University, 2015,42(1):187-193. doi: 10.3969/j.issn.1001-2400.2015.01.030
[11]	SEDIGHI V, COGRANNE R, FRIDRICH J . Content-adaptive Steganography by Minimizing Statistical Detectability[J]. IEEE Transactions on Information Forensics and Security, 2016,11(2):221-234. doi: 10.1109/TIFS.2015.2486744
[12]	LI B, WANG M, LI X L , et al. A Strategy of Clustering Modification Directions in Spatial Image Steganography[J]. IEEE Transactions on Information Forensics and Security, 2015,10(9):1905-1917. doi: 10.1109/TIFS.2015.2434600
[13]	KODOVSKY J, FRIDRICH J. Steganalysis of JPEG Images Using Rich Models [C]//Proceedings of the SPIE: 8303. Bellingham: SPIE, 2012: 83030A.
[14]	刘如意, 宋建锋, 权义宁 , 等. 一种自动的高分辨率遥感影像道路提取方法[J]. 西安电子科技大学学报, 2017,44(1):100-105. doi: 10.3969/j.issn.1001-2400.2017.01.018
	LIU Ruyi, SONG Jianfeng, QUAN Yining , et al . Automatic Road Extraction Method for High-resolution Remote Sensing Images[J]. Journal of Xidian University , 2017,44(1):100-105. doi: 10.3969/j.issn.1001-2400.2017.01.018
[15]	DENEMARK T, SEDIGHI V, HOLUB V, et al. Selection-channel-aware Rich Model for Steganalysis of Digital Images [C]//Proceedings of the 2014 IEEE International Workshop on Information Forensics and Security. Piscataway: IEEE, 2014: 48-53.

隐写方法	检测错误率/%
隐写方法	文献[3]	文献[7]	文献[1]	跨层结构
S_UNIWARD	43.24	40.40	40.60	41.33

隐写分析方法	检测错误率/%
	HILL(嵌入比特数/像素)		WOW(嵌入比特数/像素)		S_UNIWARD(嵌入比特数/像素)		MiPOD(嵌入比特数/像素)
	0.1	0.2	0.1	0.2	0.1	0.2	0.1	0.2
文献[3]	42.47	33.25	43.53	34.53	43.24	32.33	43.17	34.98
文献[7]	41.75	34.94	38.88	30.48	40.40	33.18	43.66	33.18
文献[1]	43.64	35.39	40.49	31.82	40.60	31.75	42.65	34.51
跨层结构	42.04	33.47	40.57	31.67	41.33	32.69	42.87	34.69
文中方法	41.73	31.44	39.08	30.08	38.20	31.07	41.23	32.67