Content Verification of Encrypted Images Transmitted ...

Content Verification of Encrypted Images Transmitted Over Wireless AWGN Channels Sabry S. Nassar, Nabil M. Ayad, Hamdy M. Kelash, Hala S. El-sayed, Mohsen A. M. El-Bendary, Fathi E. Abd El-Samie & Osama S. Faragallah Wireless Personal Communications An International Journal ISSN 0929-6212 Wireless Pers Commun DOI 10.1007/s11277-015-3142-3

1 23

Your article is protected by copyright and all rights are held exclusively by Springer Science +Business Media New York. This e-offprint is for personal use only and shall not be selfarchived in electronic repositories. If you wish to self-archive your article, please use the accepted manuscript version for posting on your own website. You may further deposit the accepted manuscript version in any repository, provided it is only made publicly available 12 months after official publication or later and provided acknowledgement is given to the original source of publication and a link is inserted to the published article on Springer's website. The link must be accompanied by the following text: "The final publication is available at link.springer.com”.

1 23

Author's personal copy Wireless Pers Commun DOI 10.1007/s11277-015-3142-3

Content Verification of Encrypted Images Transmitted Over Wireless AWGN Channels Sabry S. Nassar1 • Nabil M. Ayad1 • Hamdy M. Kelash2 Hala S. El-sayed3 • Mohsen A. M. El-Bendary4 • Fathi E. Abd El-Samie5 • Osama S. Faragallah2,6

•

Springer Science+Business Media New York 2015

Abstract In this paper, a content-based image verification scheme is presented. This scheme is suitable for verifying the integrity of images transmitted over insecure networks. The Discrete Cosine Transform is used to embed a block-based mark for each block in another block according to a specific algorithm. To achieve higher protection, the marked image is then encrypted using 2-D chaotic Baker map before being transmitted over the communication channel. The encrypted marked image is transmitted over a wireless Additive White Gaussian Noise (AWGN) channel without error correction codes. At the receiver side, a decryption process is performed, and the marks embedded in each block are extracted to detect suspicious forgery activities. Simulation results show the suitability of the proposed scheme for applications with sensitive data types such as military and nuclear applications. Keywords Information security Discrete Cosine Transform Encryption Chaotic Baker map

& Osama S. Faragallah [email protected]; [email protected] Hala S. El-sayed [email protected] 1

Nuclear Research Center, Atomic Energy Authority of Egypt, Cairo, Egypt

2

Department of Computer Science and Engineering, Faculty of Electronic Engineering, Menoufia University, Menouf 32952, Egypt

3

Department of Electrical Engineering, Faculty of Engineering, Menoufia University, Shebin El-kom 32511, Egypt

4

Faculty of Industrial Education, Helwan University, Cairo, Egypt

5

Department of Electronics and Electrical Communications, Faculty of Electronic Engineering, Menoufia University, Menouf 32952, Egypt

6

Department of Information Technology, College of Computers and Information Technology, Taif University, Al-Hawiyah 21974, Kingdom of Saudi Arabia

123

Author's personal copy S. S. Nassar et al.

1 Introduction Information security generally means the procedures and mechanisms taken to protect the confidentiality, integrity, and availability of data during transmission over insecure channel environments. Confidentiality aims to assure data privacy to allow authorized users only to access the data. Data integrity is related to the assurance that information has not been modified during transmission from source to receiver. Availability means that information is available for authorized users at any time. To verify image content integrity and prevent forgery, some techniques have emerged. These techniques should have enough capability to sense small tampering attacks on images. They can be used in several applications such; law, journalism, commerce, nuclear, and defense applications [1]. Digital watermarking can be used to verify image integrity by embedding either a visible or an invisible watermark. A watermark may be robust, fragile, or semi-fragile based on its resistivity to various types of manipulation. A robust watermark is mainly used to protect copyright [2–4]. On the other hand, a fragile or semi-fragile watermark is used to verify the content integrity and authenticity, where the distortion introduced due to watermark embedding is used as a measure of the content modification [5]. Traditional watermarking schemes require that the receiver knows the watermark used in order to compare with the extracted one as in [6–8]. This requirement implies that the watermark has to be designed separately and transmitted in a secure manner to the receiver. Li et al. [9] propose an image integrity and authentication verification scheme using content-based watermarks and a public key cryptosystem. A feature map of the underlying image is extracted on the transmitting side as a watermark, and it is partitioned into encrypted blocks. At the receiver side, the feature map from the received image is extracted and decrypted to verify the integrity and authenticity. This scheme is capable of detecting geometric transformations, removal or addition of objects, localizing tampering, and detecting cropping. Although this scheme does not require an additional secure channel to send extra information such as image size, look-up table, or private key, there is an extra processing overhead required in watermark extraction, embedding, and verification processes. Most published work in the literature for watermarking assumes noise free environments. This is not the real case in wireless communications, where noise and third parties exist. Some attempts have been presented for image authentication in the presence of random packet loss, but with large computational complexity [10, 11]. In this paper, we present a secure image integrity verification scheme that is characterized by content-based watermarking. In this scheme, there is no need for an additional secure communication channel. It has a high degree of security due to the encryption process implemented. Section 2 provides a brief overview on Discrete Cosine Transform (DCT) as the transform domain used with the proposed scheme. Section 3 explores briefly the used image encryption algorithm, and defines the AWGN channel. The proposed scheme is described systematically in Sect. 4. The simulation results are presented in Sect. 5. Section 6 gives the paper concluding remarks with future trends.

2 Discrete-Cosine Transform Embedding Unlike the spatial domain data embedding, transform domain embedding techniques embed the information into the transform coefficients of the cover image, and then the image is converted back into the spatial domain [12, 13].

123

Author's personal copy Content Verification of Encrypted Images Transmitted Over…

The transform used in this work is the DCT, which divides the image into high, middle, and low frequency components (FH, FM, and Fl) as shown in Fig. 1. In this work, the original image is divided to non-overlapping blocks of size 8 9 8, which then according to the proposed algorithm are mapped to frequency domain by applying DCT to each block. This will result in three frequency sub-bands; low frequency sub-band, mid frequency subband, and high frequency sub-band. Most of the signal energy lies at the low frequency sub-band which contains most important visual information of the image, while in high frequency sub-band, the high frequency components of the image are usually removed through compression and sometimes noise attacks. So the mid-band frequency is selected for watermarking purpose. A modification of mid-band frequency coefficients through embedding process will not affect the image visibility and the watermark cannot be removed by compression. The equation for a 2D (N by N square image) DCT is defined by the following equation [14]: " # N 1 N 1 X X ð2i þ 1Þup ð2i þ 1Þvp ð1Þ cos aðuÞ xi cos Cðu; vÞ ¼ aðvÞ 2N 2N i¼0 i¼0 where u, v = 0, 1, 2,…,N - 1.

3 Encryption Algorithm and AWGN Channel The encryption tool that is utilized in this work is the chaotic Baker map encryption. The Baker map is defined as a two dimensional chaotic map that transforms a square matrix into itself after operations similar to kneading operations which The Baker map can be considered as an efficient tool to randomize a square matrix of data. The discretized map can be represented by an M 9 M matrix as shown in Fig. 2, that represents a 2-D chaotic encryption of an 8 9 8 matrix [15]: The dependency of Baker map encryption on the ‘‘stretch-and-fold’’ concept which yields a great geometric transformation of plain image with adjacent pixels has become no longer relevant, which means that plain image is randomly distributed in the encrypted image. Wireless communication has become one of the most active areas of technology development, and a more prominent part of everyday life. There are three important and frequently used models to simulate wireless channels, which are Additive white Gaussian noise (AWGN), Rayleigh and Rician. AWGN channel model can be characterized as a

Fig. 1 Definition of DCT regions

123


Fig. 2 2-D chaotic encryption of an 8 9 8 matrix. a Original square matrix. b Chaotic encrypted matrix

wireless medium, which suffers linear addition of a white noise with constant spectral density, and normal Gaussian distribution. It does not account for fading, interference, nonlinearity or dispersion. The proposed algorithms are applied using this channel model to study their performance and practicality.

4 The Proposed Verification Algorithm The model investigated in this work is an attempt that can be used as a multi-task scheme for image content integrity verification, image confidence, and tampering areas detection. The model can be sub-divided into two models, where one model deals with marking process and the other deals with verification process. The two models are explained in a step-wise procedure below.

4.1 Marking Process This section presents the steps of marking and encrypting the data to be transmitted over AWGN wireless channel. This marking algorithm uses the confidential data such as image, as an input, and the output will be a signed block based image (Verified Data), which appears to be the same as the original image. The marked image is then encrypted to strength its security, and becomes ready to be transmitted over insecure medium. Assuming that, the original image is a standard gray-scale image as an example. The proposed algorithm is designed to be applied optimally on an image of square size (N 9 N) and (N/2) is divisible by 8. So in case of applied images of general sizes (N 9 M), they can be resized using image processing tools to be suitable for the proposed algorithm. The marking process is described in steps as follows: Step 1 Step 2

Input the original image (f), and then divide it into two equal halves (f1 and f2). Then both f1 and (f2) are divided to (8 9 8) non-overlapping blocks of pixels Working from left to right, top to bottom through f1, DCT is applied to each block

123


Step 3

Step 4 Step 5

A substitution process is performed for first row and column of each block in a half into last row and column of a corresponding block in the other half. This substitution process is performed with weights Working from left to right, top to bottom, the inverse DCT is performed for each block The final obtained image is encrypted using 2-D Baker map

4.2 Verification Process This model takes the encrypted marked image as an input, and implements the reverse process to reconstruct the true original image. The following steps describe the extraction process; Step 1 Step 2 Step 3 Step 4

Decrypt the image to obtain a received marked image Divide the marked image (z) into two halves (z1 and z2). Both (z1) and (z2) are divided to (8 9 8) non-overlapping blocks of pixels Working from left to right, top to bottom through (z1), DCT is applied to each block Embedded rows and columns are extracted and compared to their corresponding cones through a correlation analysis

5 Simulation Results Simulation experiments have been carried out on the proposed algorithm with Matlab on the Cameraman image of size 256 9 256 and the results are shown in Tables 1, 2, 3 and 4 and Figs. 3, 4, 5, 6, 7 and 8. An Additive White Gaussian Noise has been considered in some simulation experiments. The evaluation metrics used are the correlation coefficient between extracted rows or columns and the original ones for robustness of the mark or signature, and the mean square error (MSE) and peak signal-to-noise ratio (PSNR) for the quality of image after marking or embedding. The MSE is given by: Table 1 The marked image, encrypted marked image, decrypted marked image, extracted image Error free (No AWGN) channel implementation Marked Image

Encrypted marked

Decrypted marked

image

image

Extracted image

123

Author's personal copy S. S. Nassar et al. Table 2 Decrypted marked image, and extracted image at different SNRs AWGN channel implementation images SNR

Decrypted Marked Image

Extracted Image

5 dB

10 dB

20 dB

30 dB

MSE ¼

M X N 2 1 X 0 f ði; jÞ f ði; jÞ MN i¼1 j¼1

where f(i, j) is the original secret image and f 0 ði; jÞ is the marked image. The PSNR is given by: 2552 PSNR ðdBÞ ¼ 10 log MSE The results in Table 1 show robustness in the absence of noise.

123

ð2Þ

ð3Þ

Author's personal copy Content Verification of Encrypted Images Transmitted Over… Table 3 Block based correlation, and mean correlation at different SNRs AWGN channel implementation correlation SNR

Mean Block Based Correlation

Corre lation

5 dB

0.063

10 dB

0.28

0.29 20 dB

0.29 30 dB

The results in Fig. 4 show some deterioration after extraction. Correlation values in Fig. 5 are in favor of the suggested algorithm ensuring the possibility of mark verification.

123

Author's personal copy S. S. Nassar et al. Table 4 Image quality metrics at different SNRs

SNR (dB)

AWGN channel implementation results MSE

PSNR

5

1.8741 9 1006

14.5631

10

6.481 9 1003

10.0481

20

6.480 9 1003

10.0488

50

6.480 9 1003

10.0488

Fig. 3 Original image

Fig. 4 Extracted image in case of marked and not marked input

The obtained results with noise show the possibility of working with noise. Figure 6 shows a tampered image and Fig. 7 shows the extracted image with tampering. The correlation results with tampering are shown in Fig. 8. These results prove the tampering detection capability of the proposed algorithm.

123


Fig. 5 Block based correlation

Fig. 6 Tampered image

Fig. 7 Extracted image

123


Fig. 8 Block-based correlation with tampering

6 Conclusion The paper presented a high-fidelity image marking and encryption algorithm. Simulation results revealed the possibility of mark verification in addition to the robustness to noise. Furthermore, tampering detection experiments have shown good success results. In summary, we can say that the proposed algorithm can be used for confidential data transmission over wireless channels to guarantee the originality of data and detect all forensic trial.

References 1. Lou, D. C., Liu, J. L., & Li, C. T. (2003). Digital signature-based image authentication. In C. S. Lu (Ed.), Multimedia security: Steganography and digital watermarking techniques for protection of intellectual property. Hershey: Idea Group Inc. 2. Petrovic, R. (2005). Digital watermarks for audio integrity verification. Serbia and Montenegro, Nis, pp. 28–30. 3. Chang, W. H., & Chang, L. W. (2010). Semi-fragile watermarking for image authentication, localization, and recovery using Tchebichef moments. In Communications and Information Technologies (ISCIT). 4. Radharani, S., & Valarmathi, M. L. (2010). A study of watermarking scheme for image authentication. International Journal of Computer Applications, 2(4), 24–32. 5. Li, C.-T., & Yang, F.-M. (2003). One-dimensional neighborhood forming strategy for fragile watermarking. Journal of Electronic Imaging, 12(2), 284–291. 6. Barni, M., Bartolini, F., Cappellini, V., & Piva, A. (1998). A DCT-domain system for robust image watermarking. Signal Processing, 66, 357–372. 7. Cox, I. J., Kilian, J., Leighton, F. T., & Shamoon, T. (1997). Secure spread spectrum watermarking for multimedia. IEEE Transactions on Image Processing, 6(12), 1673–1687. 8. Hartung, F., & Girod, B. (1998). Watermarking of uncompressed and compressed video. Signal Processing, 66, 283–301. 9. Li, C.-T., Lou, D.-C., & Liu, J.-L. (2000). Image authentication and integrity verification via contentbased watermarks and a public key cryptosystem. International Conference on Image Processing, 3, 694–697. 10. www.ijcta.com/documents/…/ijcta2011020544.pdf 11. Sultana, F., Charles, S., & Govardhan, A. (2013). A tamper proof noise resilient end to end image based authentication system over wireless transmission with AWGN channel using wavelet based templates and AES. IJCSNS International Journal of Computer Science and Network Security, 13(5), 41–48. 12. Shoemaker, C. (2002). Rudko. Hidden bits: A survey of techniques for digital watermarking. Independent Study EER-290 Prof Rudko, Spring.

123

Author's personal copy Content Verification of Encrypted Images Transmitted Over… 13. Tewfik, A. H. (2000). Digital watermarking. San Mercury News. 14. Goel, S., Rana, A., Kaur, M. (2013). A review of comparison techniques of image steganography. IOSR Journal of Electrical and Electronics Engineering (IOSR-JEEE), 6(1), 41–48. ISSN: 2278-1676, p-ISSN: 2320-3331. 15. El-Bendary, M. A. M., Abou-El-azm, A. E., El-Fishawy, N. A., Shawki, F., El-Tokhy, M. A. R., Kazemian, H. B. (2012). Performance of the audio signals transmission over wireless networks with the channel interleaving considerations. EURASIP Journal on Audio, Speech, and Music Processing, 1–14. Sabry S. Nassar is an Assistant Lecturer and Network Security Specialist in Department of Reactors, Nuclear Research Center (NRC), Inshas, Egypt. He received his Master of Computer Science and Engineering in 2011 from Faculty of Electronic Engineering, Menufia University, Menouf, Egypt. Currently, he is a PhD student in the same college. His research interests are in encryption techniques, steganography, watermarking, forensics, and wireless networks.

Nabil M. Ayad received Ph.D degree in CSE from Cairo University, in 1984. He is working as vice chairman for reactors division, Nuclear Research Center, Atomic Energy Authority-Egypt. He is a member of IEEE. His main research interests database and networks.

Hamdy M. Kelash received the Eng. Degree from the Institute of Electronic, Egypt in 1971, MSc degree from Faculty of Engineering Technology, Helwan University, Egypt, in 1979, and the PhD degree from Institute National Polytechnique (INP), France in 1984. He has been lecturer in 1984 at the Electronic Industry department, faculty of Electronic Engineering, also a lecturer in 1987 at the Computer Sciences and Engineering department, and an Assistant Professor in 1993 and the Head of Computer Sciences and Engineering department, faculty of Electronic Engineering, Menoufia university from 2001 to 2007. Emeritus Prof.Dr.Eng from 2007 until Now. His main research interests include optical computing, artificial intelligence, network security, image processing, digital systems and parallel computing.

123

Author's personal copy S. S. Nassar et al. Hala S. El-sayed received the B.Sc. (Hons.), M.Sc., and Ph.D. degrees in Electrical Engineering from Menoufia University, Shebin El-kom, Egypt, in 2000, 2004, and 2010, respectively. She is currently Assistant Professor with the Department of Electrical Engineering, Faculty of Engineering, Menoufia University, where she was a Demonstrator from 2002 to 2004 and has been Assistant Lecturer from 2004 to 2010 and since 2010 she has been a Teaching Staff Member with the Department of Electrical Engineering, Faculty of Engineering, Menoufia University. She is a coauthor of about 40 papers in international journals and conference proceedings, and one textbook. Her research interests cover database Security, network security, data hiding, image encryption, wireless sensor network, secure building automation systems, and biometrics.

Mohsen A. M. El-Bendary received the B.Sc. (Honors), M.Sc., and PhD. from the Faculty of Electronic Engineering, Menoufia University, Menouf, Egypt, in 1998, 2008, and 2012, respectively. He joined the teaching staff of the Department of Electronics Technology, Faculty of Industrial Education, Helwan University, Cairo, Egypt, in 2012. He is a co-author of about 55 papers in international conference proceedings and journals. He is the author of the book ‘‘Developing Security Tools of WSN and WBAN Networks Applications’’ from Springer in 2014. His current research areas of interest include image enhancement, image processing, channel coding, WT applications, Wireless Sensor Networks, data hiding, multimedia communications, medical image processing, and digital communications. Also, He has more than 15 years of experience in the light current projects, fire Alarm systems, Firefighting, Access Control, and etc.

Fathi E. Abd El-Samie received the B.Sc. (Hons.), M.Sc., and Ph.D. degrees from Menoufia University, Menouf, Egypt, in 1998, 2001, and 2005, respectively. Since 2005, he has been a Teaching Staff Member with the Department of Electronics and Electrical Communications, Faculty of Electronic Engineering, Menoufia University. He is a coauthor of about 200 papers in international conference proceedings and journals, and four textbooks. His current research interests include image enhancement, image restoration, image interpolation, superresolution reconstruction of images, data hiding, multimedia communications, medical image processing, optical signal processing, and digital communications. Dr. Abd El-Samie was a recipient of the Most Cited Paper Award from the Digital Signal Processing journal in 2008.

123

Author's personal copy Content Verification of Encrypted Images Transmitted Over… Osama S. Faragallah received the B.Sc. (Hons.), M.Sc., and Ph.D. degrees in Computer Science and Engineering from Menoufia University, Menouf, Egypt, in 1997, 2002, and 2007, respectively. He is currently Associate Professor with the Department of Computer Science and Engineering, Faculty of Electronic Engineering, Menoufia University, where he was a Demonstrator from 1997 to 2002 and has been Assistant Lecturer from 2002 to 2007 and since 2007 he has been a Teaching Staff Member with the Department of Computer Science and Engineering, Faculty of Electronic Engineering, Menoufia University. He is a coauthor of about 100 papers in international journals and conference proceedings, and two textbooks. His current research interests include network security, cryptography, internet security, multimedia security, image encryption, watermarking, steganography, data hiding, medical image processing, and chaos theory.

123