Project topics and materials
Get Complete Project Material File(s) Now! »
Digital image watermarking attacks and robustness
The robustness is the strength of the watermarking approach against the general attacks. The robust watermarking preserves the watermark through the transmission process. If intentional or unintentional attacks are happening, the authorized user can retrieve the watermark with high quality. This aspect is necessary in copyright application, the ownership issues and the medical application. The common and image processing attacks will be discussed. Moreover, some of the watermarking approaches that focus on robustness will be presented.
Digital image watermarking attacks
The attack is this process aiming at destroying, manipulating and impairing the detection of the embedding watermark. Watermark attacks include unauthorized embedding, unauthorized detection, unauthorized removal and system attack. Known Original Attack (KOA) happens to watermarked and original content. Known Message Attack (KMA) happens to watermark contents and hidden messages. Watermark Only Attack (WOA) happens to watermarked contents [Kavadi 12] [Voloshynovskiy 01], [Malshe 12], [Song 10] classified the watermark attacks into four categories explicitly Removal Attacks, Geometric attacks, Cryptographic attacks and Protocol attacks. Removal attacks are those removing the watermark from the watermarked image, but they do break the security of the watermarking algorithm. Also, they preserve the content so the cover image is still available after the end of the attacks. After a removal attack are cannot rebuild the watermark image from the attacked watermarked image. Examples of these kinds of attacks are noisy (Gaussian, Salt and Pepper, uniform), compression (JPEG, JPEG 200), filtering (Mean, Median, Gaussian), histogram and sharpness attacks. Geometric attacks are different from the removal attacks. Geometric attack aims to destroy the watermark image rather than to remove it from the watermarked image using the geometric distortion. It is still possible to rebuild the watermark after the geometric distortion if the detail of the attack can be built. The method to correct this kind of attacks is called synchronization. Examples of these kind of attacks are cropping, rotation, translation, skewing.
Cryptographic attacks aim to destroy the security feature in the watermarking system, then to find how to remove the watermark embedded and try to embed a deceptive watermark. The example of this kind is the brute-force search method which tries to break the watermarking security by using a large amount of the possible meaningful security information. Another example is an Oracle attack which is trying to create a non-watermarked image at the decoder side if the decoder policies are accessible.
Robust image watermarking approaches
In [Subramanyam 12], Subramanyam et al. proposed a robust watermarking of compressed media for tamper detection, ownership declaration, or copyright management purposes. The proposed technique is dividing the input image into non-overlapping rectangular tiles. Then, it is applying the DWT technique. In result, multiple levels of DWT give a multi-resolution image, the lowest resolution contains the low-pass image, higher resolution contains the high pass, the resolution is divided into smaller blocks known as code-blocks, and each code encodes independently. The extraction of the watermark was done after decryption of the image. They analyzed payload capacity and quality of the image for different resolutions. The experimental results showed that the higher resolution carries a higher payload capacity without affecting too much the quality, and middle resolutions carry lesser capacity and leads to more degradation in the quality. This approach is weak against noise and filtering attacks.
In [Moghaddam 13], the authors applied new robust blind watermarking techniques based on an imperialist competitive algorithm (ICA) in the spatial domain. Firstly, the approach determines the location in the host image to be selected based on modified ICA. Secondly, 5 by 5 neighbors of the modified block are chosen and the least significant color in the neighborhood of each pixel is selected for embedding, then the best neighborhood which will increase the quality of the watermarked image is selected. The proposed approach is applied in the spatial domain in order to find the optimal position of embedding. It is robust against some kinds of attacks, but it is still weak against compression and sharpening attacks.
Fragile image watermarking approaches
A fragile watermarking is a watermarking which cannot accept the modification or the changing of the image, and for which the watermark is destroyed in case of a modification. It is used for content authentication, integrity and tamper detection applications. We can classify the fragile watermarking into: hard and soft authentication. The hard authentication doesn’t allow the changing even if it is occurring from the authorized user and the watermark will be broken. While the soft-authentication accepts slight modification from the authorized user, but the watermark is broken and damaged against intentional and dangerous attacks. Semi-fragile watermarking is also called a soft [Mousavi 14] and [Malshe 12]. In [Walia 13], proposed a watermarking technique in the spatial domain is presented. The work is based on Weber law and its two properties: differential excitation and orientation. The authors divided the image into 3 × 3 blocks. Then, they computed the differential excitation of the center pixel with its 8 neighbors. The positive excitation value of the neighbor’s pixel is selected. Finally, the watermark is inserted and the selected pixel intensity is modified. The proposed technique has a high imperceptibility; it is sensitive to noise attacks like a Gaussian. It can tolerate compression attacks (JPEG) which are less than 15% compression rate. The localization of the altered region by cut and paste attack was satisfying. It is useful in the hard authentication application.
Symmetric cryptography
Symmetric cryptography is also called single key because it uses one key in both the encryption process and the decryption process [Kahate 13]. Figure 2.3 illustrates the general process of encryption and of decryption based on symmetric technique. The sender encrypts the plain text using the encryption algorithm and the secret key to provide the ciphered text, while the receiver decrypts the ciphered text based on the decryption algorithm and the same secret key to get the original text. In the symmetric technique, the secret key that is used for the encryption and the decryption processes is the same. The security in the symmetric cryptography is based on the algorithm and the shared key [Mel 01]. The drawback in the symmetric techniques is that it is difficult to share the key in a secure way. The most known examples of the symmetric techniques are DES [Standard 99], AES [Chown 02], IDEA [Leong 00], RC4 [Gupta 14]. The general diagram of the symmetric encryption and decryption process can be summarized in the following equations.
Table of contents :
Table Of Contents
Table Of Contents
List Of Figures
List Of Tables
General introduction
FIRST PART
Chapter 1 Digital Image Watermarking
1.1 Introduction
1.2 Properties of digital image watermarking
1.2.1 Robustness
1.2.2 Imperceptibility
1.2.3 Capacity and payload
1.2.4 Security
1.2.5 Computational complexity
1.2.6 Blindness
1.3 Application of digital image watermarking
1.3.1 Copyright protection
1.3.2 Transaction tracking or fingerprinting
1.3.3 Authentication
1.3.4 Integrity
1.3.5 Tamper detection
1.4 Domains of image watermarking
1.4.1 Spatial domain
1.4.2 Frequency domain
1.5 Visible image watermarking
1.5.1 Visible image watermarking techniques
1.5.2 Visible image watermarking applications
1.6 Digital image watermarking attacks and robustness
1.6.1 Digital image watermarking attacks
1.6.2 Robust image watermarking approaches
1.6.3 Fragile image watermarking approaches
1.7 Image quality measurement
1.7.1 Mean square error (MSE)
1.7.2 Peak signal to noise ratio (PSNR)
1.7.3 Structural similarity index measurement (SSIM)
1.7.4 Bit error rate (BER)
1.8 Conclusion
Chapter 2 Digital Image Cryptography
2.1 Introduction
2.2 Cryptography: definition and principle
2.3 Cryptography classification
2.3.1 Symmetric cryptography
2.3.2 Asymmetric cryptography
2.3.3 Hash function cryptography
2.4 Cryptographic techniques
2.4.1 Symmetric encryption algorithms
2.4.2 Asymmetric encryption algorithms
2.5 Cryptanalysis
2.5.1 Brute-force attack
2.5.2 Birthday attack
2.5.3 Meet in the middle attack
2.6 Conclusion
Chapter 3 Hybrid Watermarking and Cryptography Techniques
3.1 Introduction
3.2 Combined watermarking and encryption approaches classifications
3.2.1 Watermarking followed by encryption (WFE)
3.2.2 Encryption followed by watermarking (EFW)
3.2.3 Joint watermarking / decryption (JWD)
3.2.4 Joint watermarking / encryption (JWE)
3.3 Conclusion
Chapter 4 Digital Medical Image and Telemedicine Security.
4.1 Introduction
4.1.1 Digital image
4.1.2 Digital medical image
4.2 Picture archiving and communication system (PACS)
4.3 Telemedicine
4.3.1 Telemedicine and teleradiology
4.4 Digital imaging and communication in medicine (DICOM)
4.5 DICOM security profiles
4.5.1 Secure use profiles
4.5.2 Secure transport connection profiles
4.5.3 Digital signature profiles
4.5.4 Media security profiles
4.6 Medical image security requirements
4.6.1 Medical image confidentiality
4.6.2 Medical image integrity
4.6.3 Medical image authentication
4.7 Medical image security applications
4.7.1 Pure watermarking applications in the medical image
4.7.2 Cryptography and watermarking applications in the medical image
4.8 Conclusion
Chapter 5 Watermarking approaches
5.1 Introduction
5.2 Formal Concept Analysis to Improve Robustness on Medical Image Watermarking Scheme in the Spatial Domain.
5.2.1 Formal concept analysis (FCA)
5.2.2 FCA domains
5.2.3 The proposed approach
5.2.4 Experimental results
5.2.5 FCA conclusion
5.3 Generating Optimal Informed and Adaptive Watermark Image Based on Zero Suppresed Binary Decision Diagram
5.3.1 Binary decision diagram (BDD)
5.3.2 Proposed approach to hide data in a medical image
5.3.3 Watermarked image quality, watermark robustness and evaluation
5.3.4 Comparative study
5.3.5 ZBDD conclusion and discussion
5.4 A Novel Robust Informed Watermarking approach and Tamper Detection Based on Weber Law
5.4.1 Weber law
5.4.2 Robustness based proposed approach
5.4.3 Experimental results
5.4.4 Tamper detection based proposed approach
5.4.5 Comparative study
5.4.6 Weber conclusion
5.5 A Novel CT Scan Images Watermarking Scheme in DWT Transform Coefficient
5.5.1 Wavelet decomposition principle
5.5.2 Proposed watermarking approach
5.5.3 Experimental results
5.5.4 Comparative study
5.5.5 DWT conclusion
5.6 Conclusion
Chapter 6 Informed Symmetric Encryption Algorithm for DICOM Medical Image Based on N-grams .
6.1 Introduction
6.2 N-Gram
6.3 Proposed algorithm based on N-grams
6.4 Experimental results
6.5 Conclusion
Chapter 7 Efficient and Robust Encryption and Watermarking Technique Based on a New Chaotic Map Approach
7.1 Introduction
7.2 One time pad principle and chaotic maps
7.3 Proposed approach
7.3.1 The encryption and watermarking process
7.3.2 The decryption and extraction process
7.4 Tests and outcomes
7.4.1 Performance under watermark attacks
7.4.2 Encryption attacks
7.5 Comparative study
7.6 Conclusion
Conclusion and Future Works
Conclusion
Future works
References
