﻿ 基于后验概率和滤色阵列特性的图像篡改检测算法 An Image Tampering Detection Algorithm Based on the Posterior Probability and Color Filter Array Artifacts

Computer Science and Application
Vol.07 No.09(2017), Article ID:22009,8 pages
10.12677/CSA.2017.79097

An Image Tampering Detection Algorithm Based on the Posterior Probability and Color Filter Array Artifacts

Jinqiao Wei, Ying Wang

Electronic Information College, Qingdao University, Qingdao Shandong

Received: Aug. 29th, 2017; accepted: Sep. 9th, 2017; published: Sep. 13th, 2017

ABSTRACT

Focused on the artifacts between the three primaries of an image introduced by the interpolation algorithm during its acquisition process, an image tampering detection algorithm based on the posterior probability and the color filter array artifacts is proposed. Firstly, the green channel component of the image is extracted, and the two-dimensional predictive filter is used to construct the predictive error function. Then the histograms’ character of original and tampering images is analyzed, and then the Gaussian mixture statistical model is established. EM algorithm is applied to estimate the model parameters. Then the posterior probability of each sub-block as an original block is calculated, and the feature likelihood is defined and it is applied to every sub-block, so that the tampering-area map can be obtained to complete the detection. The simulation results show that the algorithm has strong robustness and can locate the image’s tampered region more accurately.

Keywords:Posterior Probability, Color Filter Array Artifacts, Gaussian Mixture Model, Likelihood Ratio, Image Tempering Detection

1. 引言

T. Bianch等 [2] 对JPEG图像压缩过程进行建模，使用后验概率法估计量化步长，最后通过篡改区域和未篡改区域量化表的不一致性来定位篡改区域。Swaminathan等 [3] 提出了一种通过估计CFA模式和插值内核进行识别相机型号的方法，紧接着又在文献 [4] 中将去马赛克模型中估计出的参数之间的不一致性作为篡改检测的依据。H. Farid等 [5] 提出用期望最大化算法来估计去马赛克算法内插核参数，实现了借助每个像素与其相邻像素相关性的概率图来检测篡改区域，但这种方法需要预先知道图像篡改区域的尺寸。S. Vinoth等 [6] 提出将反向传播神经网络作为非线性模型来描述CFA插值特性并进行分类的方法。Peng Shuang等 [7] 提出一种利用自然图像颜色通道之间的相关性来检测图像篡改操作的方法，但对经模糊篡改的检测存在一定的误差。

Figure 1. Digital camera general imaging system

2. 滤色阵列

Bayer CFA传感器的原始输出图像的每个像素只有红、绿、蓝中的一种颜色分量，俗称“马赛克图像”，因此，必须采用插值算法将单色的马赛克图像转换成逼真的全彩色图像，这个过程通常被称为“去马赛克(Demosaicing)”。插值通常在缺失值的邻域应用一个加权矩阵(核)来实现，常见的CFA插值算法分两类：一类是以双线性和双三次插值 [9] 为代表的非自适应插值算法，另一类则是以梯度插值 [10] 为代表的自适应插值算法。但无论采用哪种插值算法，数码相机拍摄出的原始自然图像往往都会存在一定的滤色阵列特性，也称“去马赛克效应”，而被修改过的图像区域则不具备这些特性。

3. 算法介绍

(a) (b)

Figure 2. Bayer CFA and distribution of acquired and interpolated pixels of green channel

Figure 3. Detecting process of the proposed algorithm

3.1. 特征提取

(1)

(2)

(3)

(4)

(5)

(6)

(7)

3.2. 建立特征模型

(8)

(9)

(a) (b)

Figure 4. Histograms of proposed features. (a) feature histogram of original image; (b) feature histogram of tampered image

3.3. 生成篡改映射图

(10)

(11)

(12)

(11)式和(12)式具有相同的统计意义。将(12)式作用于图像的每个子块中，则可得到一映射图，图中的每个像素值分别对应于一个块的似然率，且似然率越低，则该子块为篡改块的概率越高。然而，似然率图往往包含通道噪声，为了消除这些噪声，需在更大的块上计算其累计特征值，块大小为，且满足。假设之间条件独立，则可定义累计似然率为：

(13)

4. 仿真结果及分析

5. 结语

(a) (b) (c) (d)

Figure 5. Localization of tampered image area. (a) original image; (b) tampered image; (c) detecting result of proposed algorithm; (d) detecting result of algorithm in [11]

(a) (b) (a) (b)

Figure 6. Localization of tampered image area. (a) original image; (b) tampered image; (c) detecting result of proposed algorithm; (d) detecting result of algorithm in [11]

An Image Tampering Detection Algorithm Based on the Posterior Probability and Color Filter Array Artifacts[J]. 计算机科学与应用, 2017, 07(09): 850-857. http://dx.doi.org/10.12677/CSA.2017.79097

1. 1. 周琳娜, 王东明. 数字图像取证技术[M]. 北京: 北京邮电大学出版社, 2008: 1-6.

2. 2. Bianchi, T. and Piva, A. (2012) Image Forgery Localization via Block-Grained Analysis of JPEG Artifacts. IEEE Transactions on Information Forensics and Security, 7, 1003-1017. https://doi.org/10.1109/TIFS.2012.2187516

3. 3. Swaminathan, A., Wu, M. and Liu, K.R. (2007) Nonintrusive Component Forensics of Visual Sensors Using Output Images. IEEE Transactions on Information Forensics and Security, 2, 91-106. https://doi.org/10.1109/TIFS.2006.890307

4. 4. Swaminathan, A., Wu, M. and Liu, K.J.R. (2008) Digital Image Forensics via Intrinsic Fingerprints. IEEE Transactions on Information Forensics and Security, 3, 101-117. https://doi.org/10.1109/TIFS.2007.916010

5. 5. Popescu, A.C. and Farid, H. (2005) Exposing Digital Forgeries in Color Filter Array Interpolated Images. IEEE Transactions on Signal Processing, 53, 3948-3959. https://doi.org/10.1109/TSP.2005.855406

6. 6. Vinoth, S., et al. (2013) Neural Network Modeling of Color Array Filter for Digital Forgery Detection Using Kernel LDA. Procedia Technology, 10, 498-504.

7. 7. 彭双, 彭圆圆, 肖长燕. 基于CFA插值的图像篡改检测算法[J]. 传感器与微系统, 2015, 32(6): 0141-0144.

8. 8. 张雯, 李学明. 改进的基于颜色滤波陈列特性的篡改检测[J]. 计算机工程与应用, 2009, 45(6): 176-179.

9. 9. Ng, T.T., Chang, S.F. and Sun, Q. (2004) Blind Detection of Photomontage Using Higher Order Statistics. Proceeding of the 2004 International Symposium on Circuits and Systems, Columbia, 23-26 May 2004, 688-691.

10. 10. Ng, T.T. and Chang, S.F. (2004) A Model for Image Splicing. Proceedings on 2004 International Conference on Image Processing, Singapore, 24-27 October 2004, 1169-1172.

11. 11. Ferrara, P., Bianchi, T., Rosa, D.A., et al. (2012) Image Forgery Localization via Fine-Grained Analysis of CFA Artifacts. IEEE Transactions on Information Forensics and Security, 7, 1566-1577. https://doi.org/10.1109/TIFS.2012.2202227