JOURNAL OF COMPUTER SCIENCE AND ENGINEERING, VOLUME 3, ISSUE 2, OCTOBER 2010 10

A Novel Blind Watermarking Scheme Based on Fuzzy Inference System S. Oueslati, A. Cherif and B. Solaiman Abstract—Digital image watermarking has been proposed as a method to enhance medical data security, confidentiality and integrity. Medical image watermarking requires extreme care when embedding additional data within the medical images because the additional information must not affect the image quality. In this paper, a novel image watermarking scheme in DCT domain based on the human visual system (HVS) model and fuzzy inference system technique is proposed. In order to efficiently extract the masking information, while taking into account the local characteristics of the image, the FIS is utilized to compute the optimum watermark weighting function that would enable the embedding of the maximum-energy and imperceptible watermark. In this work we use DICOM data as a watermark to embed in medical images. Image quality is measured with metrics which are used in image processing such as PSNR and MSE. Our results show good accuracy in the watermark extraction process. Index Terms—Digital Watermark, Fuzzy Inference System, Human Visual System, Medical imaging.

——————————
——————————

1 INTRODUCTION

T

he evolution of medical information systems, supported by advances in information technology, enables information to be shared between distant health professionals and manipulated and managed more easily [13],[15],[17]. However, at the same time, more attention should be paid to information protection. Digital image watermarking has been proposed as a method to enhance medical data security, confidentiality and integrity [1]. Medical image watermarking requires extreme care when embedding additional data within the medical images because the additional information must not affect the image quality. Medical images are stored for different purposes such as diagnosis, long time storage and research [11]. In the medical field the importance of the medical data security has been emphasized, especially with respect to the information referring to the patients (personal data, studies and diagnosis) [6], [7]. On the one hand the amount of digital medical images transmitted over the internet has increased rapidly, on the other hand the necessity of fast and secure diagnosis is important in the medical field, i.e. telemedicine, making watermarking the answer to more secure image transmissions. Anand et al. [9], proposed to insert an encrypted version of the electronic patient record (EPR) in the LSB (Least Significant Bit) of the gray scale levels of a medical image. Although ————————————————

• S. Oueslati is with the Faculty of Sciences of Tunis, Department of Physics, Laboratory of Signal Processing, Tunis, 1060, Tunisia. • Pr. A. Cherif is with the Department of Physics, Laboratory of Signal Processing, Faculty of Sciences of Tunis, 1060, Tunisia. • Pr. B. Solaiman is with the Higher National School of Telecommunication of Bretagne, Department: Image and Information Processing, Technopole of Brest Iroise, 29285 Brest, France.

the degradation in the image quality is minimum, the limitations and fragility of LSB watermarking schemes is well-known. Miaou et al. [23] proposed a method to authenticate the origin of the transmission, the message embedded is an ECG, the diagnosis report and physician’s information. Macq and Dewey [24] insert information in the headers of medical images. These approaches are not robust against attacks such as filtering, compression, additive noise, etc. neither to geometrical attacks such as rotation or scaling transformations. In this work, we propose the use of DICOM metadata as a watermark to embed in medical images extracted from the DICOM file. Our embedding and extracting schemes for digital watermarking of medical images with hidden patient information is proposed as a way to effectively reduce memory requirements, provide protection of information and reduce time and cost of transmission. We present this new method with more details. The proposed is an adaptive image watermarking algorithm based on a HVS model and a FIS. The FIS and the HVS combined are used to adjust the watermarking strength. If a watermark is applied at equal strength throughout an image it will tend to be more visible in texturally flat blocks, and less visible in busier blocks. In order for the embedded watermark to be even more robust against different types of attacks, it is essential to add as powerful invisible watermark as possible. In other words, users would like to insert the watermark with maximum strength before it becomes noticeable to the HVS. Based on the above reasons, the proposed scheme in this paper divides the original image into some 8×8 blocks, and the FIS according to different textural features and luminance of each block decide adaptively different embedding strengths. The watermark is embedded into

© 2010 JCSE http://sites.google.com/site/jcseuk/

JOURNAL OF COMPUTER SCIENCE AND ENGINEERING, VOLUME 3, ISSUE 2, OCTOBER 2010 11

the mid-band frequency range of the image after being transformed by the Discrete Cosine Transform (DCT). As a result, the watermark is more robust and imperceptible. The remainder of the paper is organized as follows. Section 2 provides a detailed a description of the HVS model and FIS, Section 3 describes the method for embedding and extraction watermark. In Section 4, the experimental results and comparisons are shown. The conclusions of our study are stated in Section 5.

2 HUMAN VISUAL SYSTEM FEATURES SELECTION

AND

TEXTURAL

2.1 The Human Visual System (HVS) In order to resist the normal signal processing and other different attacks, we wish the embedding strength to be as high as possible [22], [10]. However, because the watermark directly affects the original image, it is obvious that the higher the embedding strength, the lower the quality of the watermarked image will be. In other words, the robustness and the imperceptibility of the watermark are contradictory to each other. The best method is to take the human visual system (HVS) into account when embedding watermark. In this way, the strength of watermark is adaptive to the features of the original image to guarantee the maximum-possible imperceptivity of watermark. The local characteristics of the image are subsequently extracted using the model of the HVS. Barni [4] summarized the observations and experiments and gave the following three rules: (1) Disturbs are much less visible in highly textured regions than in uniform areas. (2) Contours are more sensitive to noise addition than highly textured regions but less than flat areas. (3) Disturbs are less visible over dark and bright regions. 2.2 Co-occurrence Matrix A general procedure for extracting textural properties of image was presented by Haralick et al. [2]. Each textural feature was computed from a set of COM probability distribution matrices for a given image. The COM measures the probability that a pixel of a particular grey level occurs at a specified direction and a distance from its neighbouring pixels. Co-occurrence matrix Co (i, j ) is represented by the function P (i, j , d ,θ ) , where i represent the grey level at location of coordinate ( x, y ) , j represents the grey level of its neighbouring pixel at a distance d and a direction h from a location ( x, y ) . The eight nearest-neighbour resolution cells (3 by 3 matrix), which define the surrounding image pixels, were expressed in terms of their spatial orientation to the central pixel (i, j ) called a reference cell [3]. The eight neighbours represent all the image pixels at a distance of 1. For example, resolution cells (i + 1, j ) and (i − 1, j )

are the nearest neighbours to the central cell (i, j ) in the horizontal direction (θ = 00 ) and at a distance ( d = 1) . This concept is extended to the three additional directions (θ = 45,90,1350 ) as well as when a distance equals 2, 3 and so on. The COM is scale invariant, i.e. the matrices show the relative frequency distributions of grey levels and describe how often one grey level will appear in a specified spatial relationship to another grey level on each image region [16]. In the current application, we used the following conditions to generate gray level co-occurrence matrices: (a) Number of gray levels: In order to save computation time, we compress the gray levels to 16 in this study. (b) Direction: The gray level co-occurrence matrices from 0°, 45°, 90°, and 135° directions are used. (c) Distance: The length of 3 pixels was used, because we experimentally found that value to be optimal. A total of 14 statistical features for each image can be calculated. We experimentally evaluated all of these features on their ability to discriminate between normal and abnormal cases. Of the 14 features, we found that the following 4 have the most powerful discrimination ability as texture features of the composite images: angular second moment (ASM), contrast (CON), correlation (COR), and entropy (ENT). Using the formulas of the textural features, the angular second moment, contrast, correlation and entropy are calculated as follows: •

The angular second moment (ASM) 2

ASM = ∑∑ {Co(i, j )} i

(1)

j

Angular second moment gives a strong measurement of uniformity. Higher non-uniformity values provide evidence of higher structural variations. Indeed the higher value of this feature indicates that the intensity varies less in an image. •

The contrast

CON = ∑∑ (i, j ) 2 Co(i, j ) i

(2)

j

‘Contrast' measures local variation in an image. A high contrast value indicates a high degree of local variation. The higher the values of contrast are, the sharper the structural variations in the image are.

© 2010 JCSE http://sites.google.com/site/jcseuk/

JOURNAL OF COMPUTER SCIENCE AND ENGINEERING, VOLUME 3, ISSUE 2, OCTOBER 2010 12

The correlation

•

∑ ijCo(i, j ) − m .m x

y

i, j

COR =

S x .S y

(3)

Where :

mx = ∑ i ∑ Co(i, j ) i

(4)

j

m y = ∑ j ∑ Co(i, j ) j

(5)

i

S 2 x = ∑ i 2 ∑ Co(i, j ) i

(6)

j

S 2 y = ∑ j 2 ∑ Co(i, j ) j

(7)

i

consuming defuzzification procedure. The fuzzy rules and membership functions were developed using intuitive logic and the characteristics of the human visual system. In order to efficiently extract the masking information, while taking into account the local characteristics of the image. In what follows we present the insertion algorithm whose steps are detailed. In this work, the FIS is utilized to compute the optimum watermark weighting function that would enable the embedding of the maximum-energy and imperceptible watermark. This FIS is therefore ideal to model the watermark weighting function, as it can incorporate the fuzzy and nonlinear aspect of human vision. The inference results are subsequently computed by means of the centroid defuzzification method, where the inferred value of a specific block k of an image is calculated as in Equation (10), where η c is the aggregated resultant membership function of the output

‘Correlation ' is a measure of linear dependency of intensity values in an image. For an image with large areas of similar intensities, correlation is much higher than for an image with noisier, uncorrelated intensities. •

fuzzy sets and

i

N

ik (8)

j

And ‘entropy' is an indication of the complexity within an image. A complex image produces a high entropy value.

2.3 Mamdani Fuzzy Inference System (FIS) Fuzzy Inference Systems (FIS) are popular computing frameworks based on the concepts of fuzzy set theory, which have been applied with success in many fields like control, system identifcation, etc. [8]. Their success is mainly due to their closeness to human perception and reasonig, as well as their intuitive handling and simplicity, which are important factors for acceptance and usability of the systems. Three main modules are of particular interest: a fuzzifier, a rule base and a defuzzifier. While the fuzzifier and the defuzzifier have the role of converting external information in fuzzy quantities and vice versa, the core of a FIS is its knowledge base, which is expressed in terms of fuzzy rules and allows for approximate reasoning. In this work, we focus on the Mamdani type, which is characterized by the following fuzzy rule schema: If

x is A then y is B

(9)

Where A and B are fuzzy sets defined on the input and output domains respectively. The main feature of such type of FIS is that both the antecedents and the consequents of the rules are expressed as linguistic constraints. As a consequence, a Mamdani FIS can provide a highly intuitive knowledge base that is easy to understand and maintain, though its rule formalization requires a time

is the universe of discourse correspond-

ing to the centroid of η c .

The entropy

ENT = −∑∑ Co(i, j ) log Co(i, j )

in

∑ = ∑

n =1 N

µ c(in )in

(10)

µ (i ) n =1 C n

In order to compute the adaptive watermark strength, the inferred value ik is multiplied by the frequency sensitivity as it is shown in the following formula:

α x , y ,k = Fx , y .ik Where

α x, y , k

(11)

the corresponds to the adaptive strength th

of a watermark at index ( x, y ) of the k block of an image. Also, Fx , y corresponds to the frequency sensitivity at index ( x ,

y) .

3 EMBEDDING AND DETECTION OF WATERMARK 3.1 The Watermark Embedding Process Among the work proposed in watermarking the medical domain, the algorithm, which is encoded on a pair of frequency values {0, 1}. Use of frequency domain DCT can fulfill not only the invisibility through the study of optimizing the insertion gain used, but also security by providing a blind algorithm or use the original image No is not essential and the extraction of the mark is through a secret key [5], [12]. In addition we will target a specific robustness against distortions such as JPEG lossy compression, the approach we propose, results in different marks binary coding in DCT coefficients selected on each image block. In what follows we present the algorithm of embedding in which different stages are well detailed.

JOURNAL OF COMPUTER SCIENCE AND ENGINEERING, VOLUME 3, ISSUE 2, OCTOBER 2010 13

Read an image and decompose it into blocks of 8x8

Read the mark and turn it into vector

M (i)

Apply the DCT

M (i ) = 0

Read the watermarked image

Apply the DCT

Take the same coefficients chosen when inserting

Ci (1,5)〈Ci (3,4) Ci (1,5)〈Ci (3,4) Swap values

Ci (1,5) ≥ Ci (3, 4 ) M (i ) = 1

M (i ) = 0

Ci (1,5) ≥ Ci (3,4) A binary vector

M (i )

dimension

equal to the number of block in the image watermarked

Diff = Ci (1,5) − Ci (3, 4) Watermark extracted M ' (i )

Abs ( Diff ) ≤ gain

Ci (1,5)〉 Ci (3,4)

Ci (1,5) = Ci (1,5) + ( gain / 2) Ci (3,4) = Ci (3,4) − ( gain / 2)

Calcul of correlation between M (i ) and M ' (i )

Ci(1,5) = Ci(1,5) − (gain/ 2) Ci(3,4) = Ci(3,4) + (gain/ 2)

Apply the IDCT

Original Watermark M (i)

Fig.2. The watermark detection process.

Watermarked image

Fig.

1.

The

watermark

embedding

3.2 The Watermark Detection Process In our approach we have two key insertion markings to secure the site where the mark was introduced. The first key tells us the positions of the two coefficients selected with the same value of quantization [18], [20]. While the second key position Relates blocks which bear the marks among all the components total blocks transformed image. Phase extraction is as follows: Compare the values of DCT coefficients to determine if the respective bit of the message was a "0" or a "1". We present in what follows the extraction algorithm in the frequency domain.

4 VALIDATION OF THE NEW APPROACH We evaluated the performance of the proposed method by conducting several simulations that based on the algorithm described in Section3, using standard medical images. Simulation conditions, results and interpretation are presented in this section.

Fig.3. Originals medicals images used in experiments

JOURNAL OF COMPUTER SCIENCE AND ENGINEERING, VOLUME 3, ISSUE 2, OCTOBER 2010 14

4.1 Robustness Measure The pick signal to noise ratio (PSNR) is used to evaluate the image quality by calculating the Mean Square Error (MSE) between the images to compare and NCC (Normalized Cross-Correlation) to evaluate the similarity between the original watermark and the extracted watermark.

MSE =

1 MN

M −1 N −1

∑∑ ( f ( x, y) − f ' ( x, y ))

2

(12)

x =0 y =0

Where M, N is the size of the image and contains M x N pixels, f(x, y) is the host image and f’(x, y) is the watermarked image. This measure gives an indication of how much degradation values near to zero indicate less degradation.

PSNR = 10l log10 ( xmax

2 X max 255 2 ) = 10 log10 ( ) (13) MSE MSE

4.2 Robustness towards JPEG Compression We always need to apply JPEG compression to the medical images for archive or transmission [21]. It is therefore crucial to examine whether the proposed watermarking scheme can survive JPEG compression attacks. In the medical field, the compression rate is constantly a discussion subject because of the importance of the medical information. In order to perform this experiment, the watermarked image was compressed using different quality factors 90%, 70%, 50%, 30% and 10%. We obtain the watermark detection results for the JPEG compression of watermarked medical image with different quality factor; where lower quality factor value corresponds to higher compression ratio. Obviously, when the JPEG compression quality factor decreases, the detector response also decreases. It is noted that the compressed image is significantly degraded at quality factor = 20 (see Figure 6 (a)), yet the watermark is still easily detected (see Figure 6 (b)).

is the luminance max.

Corrélation entre la signature et le dictionnaire 0.6 0.5 0.4

By using the proposed scheme, the watermark is almost imperceptibility to the human eyes, as shown in figure 4 and figure 5.

Corrélation

0.3 0.2 0.1 0 -0.1

100

(a)

90 80

(b)

-0.2 0

100

200

300 400 500 Indexe de la signature

600

700

800

Fig.6. (a) JPEG compression (with quality factor 20%) of the watermarked image (b) Watermark detector response of attacked by JPEG compression (with quality factor 20).

70

PSNR dB

60 50 40 30 20 10 0

0

5

10

15 Images test

20

25

30

Fig.4. Mean values of PSNR for different medical image test.

(a)

(b)

Fig.5. (a) Original medical image (b) Medical image watermarked PSNR: 42.12 dB

In addition, we compare our experimental results to Cox et al.’s scheme [27], Huang and Shi’s scheme [26], as well as Lou and Yin’s scheme [25]. Watermark schemes proposed by these authors have a weakness that the HVS model has not been taken into account. For the sake of the imperceptibility constraint of a watermark, Authors include low strength watermark to avoid degrading the image quality. Unfortunately, it reduces the robustness of the watermark. In this paper, using the HVS model and the FIS, the watermark can be adjusted for each different image that provides a maximum and suitable watermark subject to the imperceptibility constraint. The results are shown in figure 7. Our method clearly outperforms the other three aforementioned schemes.

JOURNAL OF COMPUTER SCIENCE AND ENGINEERING, VOLUME 3, ISSUE 2, OCTOBER 2010 15

Corrélation entre la signature et le dictionnaire 0.7 0.6 0.5

Corrélation

0.4 0.3 0.2 0.1 0 -0.1 -0.2 0

100

200

300 400 500 Indexe de la signature

600

700

800

Fig.9. (a) Watermarked image attacked by Gaussian noise (b) Watermark detector response of attacked by Gaussian noise

Fig.7. Comparison with different algorithm

4.3 Robustness towards Noise It seems to be interesting to evaluate the robustness of the proposed method towards noise. In fact, in the medical field, the used instruments add different noise types to the medical images [14], [19]. We have tested the proposed method using different noise generations by modifying either its type or its variance. These values reach their maximum for the speckle noise. For the images attacked by a gaussian noise, the values of PSNR are minimal as shown in figure 8. Whereas, these values are always above 30dB which allows considering that the image quality is good and these watermarked images remain useful even with the presence of these noises. The watermark detector response when the watermarked image is introduced to additive Gaussian noise is shown in Figure 9(b). The results demonstrate that this scheme is significantly robust against additive Gaussian noise.

Fig.8. Means values of PSNR for several test images watermarked and attacked by different noise types.

4.4 Robustness towards Filtering We have tested the robustness of our proposed method towards 3 filter types: Gaussian, unsharp and average. These are the most used filers to eliminate noise from images. We have changed the filter size until bluffing the image and we always succeeded to extract the embedded signature. Figure 10 displays good values of PSNR.

Fig.10. Mean values of PSNR for several tests images Watermarked and attacked by different filters.

5 Conclusion Proposed new watermarking scheme shows superior performance. The advantages of the proposed technique include: (1) It provides the advantage of fuzzy inference system that extracts the image characteristics using HVS model. . (2) By different embedding strengths decided by fuzzy inference system according to different textural features and the mean luminance of each block, the resulting watermarked image is extremely robust to a wide range of image JPEG compression. (3) It does not require the original image for watermark detection. Experimental results show that the Fuzzy inference system can satisfactorily maximizing the watermark strength with properly trainings, which is adaptive based on the knowledge of the local block features.

JOURNAL OF COMPUTER SCIENCE AND ENGINEERING, VOLUME 3, ISSUE 2, OCTOBER 2010 16

ACKNOWLEDGMENT The authors express gratitude to Dr. REZGUI Haythem clinic from the MANAR II of TUNISIA, for helping and being in assistance.

REFERENCES [1]

[2]

[3]

[4]

[5]

[6]

[7]

[8]

[9]

[10]

[11]

[12]

[13]

[14]

[15]

[16]

S.Srdjan, O.Irena, “An application of multidimensional timefrequency analysis as base for the unified watermarking approach,” IEEE Transactions on image processing, vol. 19, no. 3, pp. 736–745, March. 2010. R. M. Haralick, K. Shanmugam and I. Dinstein, “Textural features for image classification”, IEEE Transactions on Systems, Man, and Cybernetics, 3(6), pp.610-621, 1973. P. Maji, M. K. Kundu, B. Chanda, “Segmentation of Brain MR Images Using Fuzzy Sets and Modified Co-Occurrence Matrix”, Proceedings of The IET International Conference on Visual Information Engineering, 327–332, 2006. M.Barni, F.Bartolini, and T.Furon. “A general framework for robust watermarking security”. Signal Processing, 83(10), pp. 2069-2084, 2003. J. M. Zain, L. P. Baldwin, and M. Clarke, “Reversible watermarking for authentication of DICOM images,” in Proc. 26th Annu. Int. Conf. Eng. Med. Biol. Soc. (EMBC 2004), vol. 2, pp. 3237–3240, 2009. A. Giakoumaki, S. Pavlopoulos, and D. Koutsouris, “Secure and efficient health data management through multiple watermarking on medical images,” Med. Biol. Eng. Comput., vol. 44, pp. 619–631, 2006. A.O.Zaid, “Compression et tatouage d’images à des fins d’archivage et de transmission : application aux images médicales,” HDR at the university, Tunis El Manar, 2009. N. Sakr, N. D. Georganas, and J. Zhao, “Copyright Protection of Image Learning Objects using Wavelet-based Watermarking and Fuzzy Logic”, in proceedings of Intelligent Interactive Learning Object Repositories (I2LOR 2006) Conference, Montreal, Canada, November 2006. D. Anand and U. C. Niranjan. “Watermarking Medical Images with Patient Information”. In Proceedings IEEE/EMBS Conference, Hong Kong, China, pp. 703-706, October 1998. P. Foriš, and D. Levický, “Implementations of HVS Models In Digital Image Watermarking”, Journal of Radio engineering, vol. 16, no. 1, pp. 45-50, April 2007. F. Cao, H. K. Huang, and X. Q. Zhou, “Medical image security in a HIPAA mandated PACS environment,” Comput. Med. Imag. Graph., vol. 27, pp. 185–196, 2003. G.S.El-Taweel, H.M. Onsi, M.Samy, M.G. Darwish, "Secure and Non-Blind Watermarking Scheme for Color Images Based on DWT", GVIP Special Issue on Watermarking, 2007. C. S. Hsu and S. F Tu, "An Imperceptible Watermarking Scheme Using Variation and Modular Operations", International Journal of Hybrid Information Technology, vol. 1, no. 4, October 2008. S. C. Pei and Y. C. Zeng, “A novel image recovery algorithm for visible watermarked images,” IEEE Trans. Inf. Forens. Security, vol. 1, no. 4, pp. 543–550, Dec. 2006. N. Khademi, S. M. Ahadi, “A Robust image watermarking in the ridgelet domain using universally optimum decoder,” IEEE Trans. Circuits Syst. Video Technol., vol. 20, no. 3, pp. 396–406, March. 2010. S. A. Shearer and R. G. Holmes, "Plant identification using color

[17]

[18]

[19]

[20]

[21]

[22]

[23]

[24]

[25]

[26] [27]

co-occurrence matrices» Transactions of the ASAE, 33(6), 20372044, 1990. Y. Yang, X. Sun, H. Yang, and C.-T. Li, “Removable visible image watermarking algorithm in the discrete cosine transform domain,” J.Electron. Imaging, vol. 17, no. 3, pp. 033008-1–03300811 Jul.–Sep. 2008. Y. J. Hu, S. Kwong, and J. Huang, “An algorithm for removable visible watermarking,” IEEE Trans. Circuits Syst. Video Technol., vol. 16, no.1, pp. 129–133, Jan. 2006. H. M. Tsai and L. W. Chang, “A high secure reversible visible watermarking scheme,” in Proc. IEEE Int. Conf. Multimedia Expo, Beijing, China, pp. 2106–2109, 2007. J. Cox, M. L. Miller, J. A. Bloom, and J. Fridrich, Digital Watermarking and Steganography, 2nd ed. San Mateo, CA: Morgan Kaufmann, 2007. U. R. Acharya, P. S. Bhat, S. Kumar, and L. C. Min, “Transmission and storage of medical images with patient information,” Comput. Biol. Med., vol. 33, pp. 303–210, 2003. Y. Zhang, “Blind Watermark Algorithm Based on HVS and RBF Neural Network in DWT Domain”, WSEAS Transactions on Computers, vol. 8, no. 1, January 2009. S. G. Miaou et al. “A Secure Data Hiding Technique with Heterogeneous Data-Combining Capability for Electronic Patient Record”. In Proceedings of the World Congress on Medical Physics and Biomedical Engineering, Session Electronic Healthcare Records, USA, July 2000. B. Macq and F. Dewey, “Trusted Headers for Medical Images”, In DFG VIII-DII Watermarking Workshop, Erlangen, Germany, October 1999. D.C. Lou and T.L. Yin, “Adaptive digital watermarking using fuzzy logic techniques”, Optical Engineering, Volume 41, Issue 10, pp.2675-2687, 2002. J. Huang and Y. Q. Shi, “Adaptive image watermarking scheme based on visual masking”, IEE Electron. Lett., 748-750, 1998. I.J. Cox, J. Kilian, T. Leighton, and T. Shamoon, “Secure spread spectrum watermarking for multimedia”, IEEE Transactions on Image Processing, volume 6, no. 12, pp. 1673-1687, 1997.

S. Oueslati is a researcher at the Image and Information Processing Department Higher National School of Telecommunications of Bretagne she is also in signal processing laboratory at the University of Sciences of Tunis - Tunisia (FST). Degree in electronics and she received a Masters degree in 2006 from the University of Sciences of Tunis. She is currently a PhD student at the Faculty of Sciences of Tunis of where she is a contractual assistant. His research interests include information hiding and image processing, digital watermarking, database security.

A. Cherif obtained his engineering diploma in 1988 from the Engineering Faculty of Tunis and his Ph.D. in electrical engineering and electronics in 1997. Actually he is a professor at the Science Faculty of Tunis, responsible for the Signal Processing Laboratory. He participated in several research and cooperation projects, and is the author of more than 60 international communications and publications.

B. Solaiman, Ing. Ph.D., Professor, is Telecom Engineer, holds a Ph.D. and HDR in Information Processing, University of Rennes I, is currently Professor and Head of Image and Information Processing from the Higher National School of Telecommunication of Bretagne in Brest, France. His research interests include, among others, on different approaches to treatment and Information Fusion and have been the subject of numerous publications.