SCIENTIA SINICA Informationis, Volume 49, Issue 4: 464-485(2019) https://doi.org/10.1360/N112018-00196

## A watermarking algorithm for image content authentication in double-compression environment

• AcceptedDec 17, 2018
• PublishedApr 11, 2019
Share
Rating

### Abstract

Malicious tampering of digital image content may lead to serious consequences. As image resolution increases and its circulations on the network grows rapidly, images undergo double compression in the following environment: compression (JPEG/JPEG2000) for first release, decoding, application processing (conventional signal processing/security attacks/malicious content tampering), and compression (JPEG2000/JPEG) for rerelease before detection. Effectively determining whether the image content has been tampered and to locate the tampering position in a double-compression environment is an urgent problem. In this paper, a novel watermark expression method based on rotating vector and its modulation algorithm is proposed. Based on this method, a semi-fragile watermarking scheme is established for the image content authentication. The stability of watermarked data is analyzed theoretically, providing the watermark scheme with a theoretical basis. The authentication schemes with feature extraction and reconstruction, watermark embedding and extraction, and tamper detection and localization are mainly elaborated. The robustness of watermark, its security, and related detection performance are analyzed. Theoretical analysis and experiments show that the scheme features a good watermark transparency, good robustness, and stable distribution for different attacks and can effectively distinguish malicious tampering from content-preserving operators and locate tampering area under double-compression environment. In addition, the proposed scheme can resist watermark attacks, collage attacks, and forge attacks. Compared with related schemes, the proposed scheme has superior comprehensive performance and is suitable for content authentication in double-compression environment, which expands the application scope of watermark-based content authentication.

### References

[1] Lin C Y, Chang S F. Semi-fragile watermarking for authenticating JPEG visual content. In: Proceedings of SPIE, Security and Watermarking of Multimedia Contents II, San Jose, 2000. 140--151. Google Scholar

[2] Chang C C, Tsai P, Lin C C. SVD-based digital image watermarking scheme. Pattern Recognition Lett, 2005, 26: 1577-1586 CrossRef Google Scholar

[3] Li C, Huang J W. A Semi-Fragile Image Watermarking Resisting to JPEG. J Software, 2006, 17: 315-324 CrossRef Google Scholar

[4] Gong D F, Liu F L, Luo X Y. An image edge based robust watermarking algorithm. Sci Sin Inform, 2013, 43: 1410--1430. Google Scholar

[5] Ma L, Zhang X H. Characteristics of color images with watermark based on the relationship between non-void subspaces of inner space. Chinese J Comput, 2017, 40: 1203--1217. Google Scholar

[6] Maeno K, Sun Q B, Chang S F. New semi-fragile image authentication watermarking techniques using random bias and nonuniform quantization. IEEE Trans Multimedia, 2006, 8: 32-45 CrossRef Google Scholar

[7] Patra J C, Phua J E, Bornand C. A novel DCT domain CRT-based watermarking scheme for image authentication surviving JPEG compression. Digital Signal Processing, 2010, 20: 1597-1611 CrossRef Google Scholar

[8] Zhang X, Cui L, Shao L. A fast semi-fragile watermarking scheme based on quantizing the weighted mean of integer Haar wavelet coefficients. In: Proceedings of Symposium on Photonics and Optoelectronics, Shanghai, 2012. Google Scholar

[9] Preda R O. Semi-fragile watermarking for image authentication with sensitive tamper localization in the wavelet domain. Measurement, 2013, 46: 367-373 CrossRef Google Scholar

[10] Das C, Panigrahi S, Sharma V K. A novel blind robust image watermarking in DCT domain using inter-block coefficient correlation. AEU - Int J Electron Commun, 2014, 68: 244-253 CrossRef Google Scholar

[11] Al-Haj A, Amer A. Secured telemedicine using region-based watermarking with tamper localization.. J Digit Imag, 2014, 27: 737-750 CrossRef PubMed Google Scholar

[12] Guo J, Zheng P, Huang J. Secure watermarking scheme against watermark attacks in the encrypted domain. J Visual Communication Image Representation, 2015, 30: 125-135 CrossRef Google Scholar

[13] Qi X, Xin X. A singular-value-based semi-fragile watermarking scheme for image content authentication with tamper localization. J Visual Communication Image Representation, 2015, 30: 312-327 CrossRef Google Scholar

[14] Chen S T, Huang H N, Kung W M. Optimization-based image watermarking with integrated quantization embedding in the wavelet-domain. Multimed Tools Appl, 2016, 75: 5493-5511 CrossRef Google Scholar

[15] Imran S, Kumar D P, Tetsuya S. A semi-fragile watermarking method using slant transform and LU decomposition for image authentication. In: Proceedings of International Conference on Electrical, Computer and Communication Engineering, Cox's Bazar, 2017. 881--885. Google Scholar

[16] Wang Y, Pearmain A. Blind image data hiding based on self reference. Pattern Recognition Lett, 2004, 25: 1681-1689 CrossRef Google Scholar

[17] Hsu L Y, Hu H T. Robust blind image watermarking using crisscross inter-block prediction in the DCT domain. J Visual Communication Image Representation, 2017, 46: 33-47 CrossRef Google Scholar

[18] Parah S A, Sheikh J A, Loan N A. Utilizing neighborhood coefficient correlation: a new image watermarking technique robust to singular and hybrid attacks. Multidim Syst Sign Process, 2018, 29: 1095-1117 CrossRef Google Scholar

[19] Ali M, Ahn C W. An optimal image watermarking approach through cuckoo search algorithm in wavelet domain. Int J Syst Assur Eng Manag, 2018, 9: 602-611 CrossRef Google Scholar

[20] Verma V S, Jha R K, Ojha A. Digital watermark extraction using support vector machine with principal component analysis based feature reduction. J Visual Communication Image Representation, 2015, 31: 75-85 CrossRef Google Scholar

[21] Mehta R, Rajpal N, Vishwakarma V P. A robust and efficient image watermarking scheme based on Lagrangian SVR and lifting wavelet transform. Int J Mach Learn Cyber, 2017, 8: 379-395 CrossRef Google Scholar

[22] Holliman M, Memon N. Counterfeiting attacks on oblivious block-wise independent invisible watermarking schemes. IEEE Trans Image Process, 2000, 9: 432-441 CrossRef PubMed ADS Google Scholar

[23] Goljan M. Cryptanalysis of the Yeung-Mintzer fragile watermarking technique. J Electron Imag, 2002, 11: 262-274 CrossRef ADS Google Scholar

[24] Wang Z, Bovik A C, Sheikh H R. Image Quality Assessment: From Error Visibility to Structural Similarity. IEEE Trans Image Process, 2004, 13: 600-612 CrossRef Google Scholar

• Figure 1

(Color online) Watermarked vector ${{\boldsymbol~A}_w}$ formed through rotating the original vector (${\boldsymbol~A}$) by $\theta$. (a) Counterclockwise; (b) clockwise

• Figure 2

(Color online) The relationship between the dithered angle ($\theta_d$) of vector ${\boldsymbol~A}$ and the cumulative probability under perturbation for (a) test schemes 1$\sim$4, respectively, and (b) test scheme 5 and schemes 1$\sim$4

• Figure 3

(Color online) The probability of feature stability for individual attack in double compression environment for (a) test schemes 1$\sim$4, respectively, and (b) test scheme 5 and schemes 1$\sim$4

• Figure 4

(Color online) Function relationship between tampering proportion ($p_t$) and robustness of the tampering matrix ($P_{\rm~t\_Tr}$), false positive probability for localization ($P_{\rm~tl\_fp}$), false negative probability for localization ($P_{\rm~tl\_fn}$) and tampering localization accuracy ($P_{\rm~tl\_d}$), respectively, under double

• Figure 5

(Color online) Robustness of single compression. (a) JPEG; (b) JPEG2000

• Figure 6

(Color online) Robustness of different image attacks during double compression. (a)$\sim$(d) correspond to the double compression schemes 1$\sim$4, respectively, (e) is the uncompressed case, and (f) is the average of (a)$\sim$(d)

• Figure 7

(Color online) Detection mask images after content authentication for the content-preserving image attack in the double compression environment. Subtitles show the detailed process of signal processing; “DCi" stands for ith-type double compression scheme; “Ai" is the number of attacks in Subsection 2.4. (a) Jpeg(75); (b) Jp2k(8); (c) DC1; (d) DC2; (e) DC3; (f) DC4; (g) DC5; (h) DC1$\leftarrow$A3; (i) DC2$\leftarrow$A6; (j) DC3$\leftarrow$A8; (k) DC4$\leftarrow$A11; (l) DC5$\leftarrow$A2

• Figure 8

(Color online) Detection mask images after content authentication for malicious content tampering attack in the double compression environment. Subtitles show the detailed process of signal processing; “DCi" stands for ith-type double compression scheme; “Ai" is the number of attacks in Subsection 2.4; “T" and “Tc" respectively represent content tampering and collage attack. (a) Jpeg(75)TcA1; (b) Jp2k(8)Tc; (c) DC1$\leftarrow$Tc; (d) DC2$\leftarrow$Tc; (e) DC3$\leftarrow$T; (f) DC4$\leftarrow$T; (g) DC5$\leftarrow$T; (h) DC1$\leftarrow$TA3; (i) DC2$\leftarrow$TA6; (j) DC3$\leftarrow$TA8; (k) DC4$\leftarrow$TA11; (l) DC5$\leftarrow$TA2

• Table 1   Comparison of PSNR values
 Lena Sailboat Baboon Peppers Man Average (128 images) Ours 40.71 39.03 42.46 39.80 39.41 41.07 Qi 40.28 40.30 40.35 40.33 40.32 40.30 Wang 34.13 31.40 31.58 34.84 32.53 34.00 Hsu 38.69 37.38 37.78 38.45 37.80 38.19

a) The bold font represents the maximum of the corresponding column.

• Table 2   Comparison of SSIM values
 Lena Sailboat Baboon Peppers Man Average (128 images) Ours 0.9861 0.9885 0.9960 0.9837 0.9862 0.9897 Qi 0.9598 0.9707 0.9866 0.9609 0.9704 0.9701 Wang 0.9620 0.9527 0.9659 0.9653 0.9480 0.9627 Hsu 0.9563 0.9629 0.9826 0.9554 0.9631 0.9650

a) The bold font represents the maximum of the corresponding column.

• Table 3   Tampering detection results and comparison of theoretical and experimental data under double compression
 Sub-graph $p_t$ (%) $P_{\rm~t\_Tr}$ $P_{\rm~tl\_fp}$ $P_{\rm~tl\_fn}$ BER TCR Auth. Theory Experiment Theory Experiment Theory Experiment (a) 1.000 0.058 0.058 0.000 Y (b) 0.982 0.127 0.127 0.139 Y (c) 0.975 0.195 0.195 0.130 Y (d) 0.971 0.170 0.170 0.172 Y (e) 0.998 0.037 0.037 0.053 Y (I) (f) 0.952 0.180 0.182 0.266 Y? Figure 7 (g) 0.0 0.977 0.990 0.023 0.052 0.0 0.0 0.052 0.189 Y (h) 0.987 0.169 0.169 0.075 Y (i) 0.913 0.227 0.227 0.313 Y? (j) 1.000 0.046 0.046 0.000 Y (k) 0.969 0.175 0.175 0.179 Y (l) 0.957 0.218 0.218 0.197 Y Avg($\cdot$) $|\Delta|=0.002$ $|\Delta|=0.119$ $|\Delta|=0.0$ 0.142 Sub-graph $p_t$ (%) $P_{\rm~t\_Tr}$ $P_{\rm~tl\_fp}$ $P_{\rm~tl\_fn}$ BER TCR Auth. Theory Experiment Theory Experiment Theory Experiment (a) 10.55 0.939 0.917 0.036 0.031 0.051 0.054 0.226 0.368 N (b) 16.41 0.913 0.894 0.043 0.050 0.076 0.107 0.259 0.411 N (c) 20.31 0.890 0.894 0.049 0.041 0.096 0.138 0.283 0.376 N (d) 24.90 0.857 0.868 0.057 0.067 0.121 0.185 0.296 0.446 N (e) 11.82 0.935 0.943 0.037 0.008 0.056 0.069 0.146 0.389 N (II) (f) 17.19 0.908 0.870 0.044 0.062 0.081 0.104 0.278 0.467 N Figure 8 (g) 22.85 0.870 0.885 0.054 0.010 0.111 0.123 0.243 0.474 N (h) 27.34 0.843 0.837 0.060 0.043 0.131 0.153 0.333 0.490 N (i) 7.91 0.953 0.899 0.032 0.075 0.035 0.054 0.301 0.334 N (j) 29.30 0.829 0.891 0.063 0.003 0.141 0.187 0.268 0.409 N (k) 40.43 0.749 0.758 0.078 0.060 0.197 0.222 0.380 0.638 N (l) 50.78 0.671 0.649 0.090 0.069 0.253 0.227 0.442 0.792 NW Avg($\cdot$) $|\Delta|=0.023$ $|\Delta|=0.023$ $|\Delta|=0.027$
• 表 4表 1

Citations

• #### 0

Altmetric

Copyright 2020 Science China Press Co., Ltd. 《中国科学》杂志社有限责任公司 版权所有