Essay: Digital watermarks

Now a days, Internet become more user friendly as it is an excellent sharing system for digital media which is basically a form of electronic media where data are stored in some digital structure. Thus today, it is almost possible for anyone to make replica or manipulate the digital data, while not losing data quality. The content owners thus see a high risk of piracy and they are very much concern about the protection of their intellectual property from illegal copying of data. Hence, the embedded watermark will become a solution because it allows identification of the owner of the work. Watermarking belongs to the information hiding. It simply referred to as a pattern of bits inserted into a digital data (that is image, audio or video file) which identifies the file’s copyright information (such as rights, author, etc’). But the idea of embedding some information within a digital document in such a way that it seems the inserted data is intrinsically part of the document itself is the main issue. The challenge is successfully addressed by using the methods such as encryption, to make sure that the only intended recipients are able to access that copyrighted work from an author’s or legitimate distributor’s point of distribution. However, encryption cannot address the problem of how to prevent the data from an unauthorized re-distribution, because eventually the encrypted work can be decrypted so that the legitimate recipients can right to use it and once it decrypted, the copyrighted work is again at risk. Thus the concept of watermarking was introduced.
The word “digital watermark” was firstly coined in the year 1992 by Charles Osborne and Andrew Tirkel. The watermarks are identification marks which produced during the paper making procedure. The first watermark was appeared in Italy around the 13th century. They were used as a means to recognize the papermaker or the trade association that actually manufactured the paper. And watermarks continue to be used today also as manufacturer’s marks and to prevent from forgery. And now watermarking is one of the alternatives for copyright-protection problem.
Digital Watermarking describes methods and technologies which hides information, for example a number or text, in digital media, such as images, audio or video. A digital watermark is a signal permanently embedded into digital data (audio, images, video, and text) that can be detected or extracted further by means of computing operations in order to make assertions about the data. Digital watermarking is basically the act of hiding a message (information hiding) related to a digital signal within the signal itself. The watermark is hidden in the host data in such a way that it is inseparable from the data and so that it is resistant to many operations not degrading the host document. Typical watermarking system mainly consists of Insertion, Extraction and Detection processes. The embedding of data takes place by manipulating the content of the digital data, which means the information is not embed around the data frames. The hiding process is to be such that the modifications of the media should be unnoticeable means for the images the modification in the pixel values have to be invisible.
Digital watermarking techniques derive from steganography, which means covered writing (from the Greek words stegano means ‘covered’ and graphos means ‘to write’). Steganography is the science of communicating information while hiding the real existence of the message. The basic goal of steganography is hiding an information message inside harmless messages in such a way that it is not possible even to detect that there is a secret message present. Both steganography and watermarking belong to a category of information hiding. In watermarking, for example, the important information is the ‘external’ data (e.g., images, voices, etc.). The ‘internal’ data (e.g., watermark) are additional data for protecting the external data and to prove ownership. [1] In general steganography is a technique that allows a secret communication by hiding or embedding the secret data into another unsuspected data. The steganographic methods are generally do rely on assumptions, that survival of the covert communication is unidentified for the third parties. And watermarking as opposed to the steganographic technique has the extra notion of the robustness against all the attacks. Also if existence of hidden data is known it will be difficult or we say it is ideally not possible for an enemy (attacker) to destroy embedded watermark data also if algorithmic method of the watermarking is public.
Watermarks and watermarking techniques can be divided into various categories in different ways [2]. According to the type of document which has to be watermarked, the watermarking techniques can be divided into four categories as follows:
‘ Text Watermarking
‘ Image watermarking
‘ Audio Watermarking
‘ Video Watermarking
According to the division based on the human perception, watermarking technique can be divided as:
‘ Visible Watermark
Visible watermarks can clearly be seen by the viewer and also get identify the owner. Thus it is also called as perceptible.
‘ Invisible Watermark
A digital watermark is invisible or imperceptible if the original signal (cover signal) and marked signal are not perceptually distinguishable.
On the basis of Robustness digital watermarking techniques can be classified as:
‘ Fragile watermark
A digital watermark is fragile if it is undetectable after the small amount of modification. These are commonly used for (integrity proof).
‘ Semi-fragile Watermark
A digital watermark is semi-fragile if it can resist gentle transformations, but fails its detection after the malignant transformations. These are commonly used to detect the malignant transformations.
Robust watermarks may get used in copy protection applications to carry copy and no access control information.
The threats may come in various forms in this field and it may weaken detection of watermark or the information convey by watermark data. In the most applications of watermarking, the marked data will likely to be process in a few ways before it reaches to the watermark receiver. The processing can be lossy less compression, lossy compression, signal improvement etc. Well these threats may be result due to the various attacks which are generally classified as the Intentional Attacks and the Non-Intentional Attacks. The wide range of the existing watermark attacks contains four classes:
‘ Removal Attacks
The objective of this type of watermark attack is complete removal of watermark information from the watermarked data without breaking the security of watermarking algorithm (like without using the key which is used for the watermark embedding process). And the data gone due to this type of attack cannot be recovered. This attack does not result on complete watermark removal but they may damage the watermark information considerably.
‘ Geometric Attacks
This attack does not really remove the embedded watermark data, but it tries to disturb the watermark detector by the embedded data. Yet, whenever desired, detector use to recover the embedded watermark data if the perfect organization is regained, but the method may be too huge to become practical and with the use of this special organization technique watermark may survive with these attacks.
‘ Cryptographic Attacks
The cryptographic attack will attacks the watermark by determining a way to remove the embedded watermark data by breaking security processes in the watermarking scheme. For access the embedded secret data, brute force attack sounds to be perfect. Also an attack called the oracle attack which mainly creates a non watermarked signal if the watermark detector founds available. But their application is overwhelmed, due to its high complexity during computation.
‘ Protocol Attacks
The protocol attack threatens the watermark data by attacking the whole concept of the application of watermarking. This attack is based on the idea of the invertible watermarks through which the attacker easily claims as the owner of watermarked data. This attack straightforwardly affects the rightful ownership of data and thus for copyright protection watermarks need to be the noninvertible and also the possibility of extracting the watermark data from any non-watermarked file aids the protocol attack. It can be solved by creating the watermarks as signal dependent and with the one-way function. One other technique is copy attack where it does not destroy watermark but it copies it into the target data and this attack is appropriate if a valid watermark is copied without the watermarking algorithm.
‘ Estimation based Attacks
Here in the concept of these attacks, they mainly lie in assessment of the embedded watermark data used with some signal statistics knowledge. The estimation in this attack is free from any information of watermark key and the embedding rule. And therefore with it, the attack is more efficient and for the attack to be completed, the attacker should concentrate on different possibilities to utilize the obtained estimate to damage the embedded watermark data.
The watermark is additional data for protecting the external data and to prove its ownership. A digital watermark is a digital data which can be embedded with all forms of media content (including digital images, digital audio, video and even certain objects also). Digital watermarks have been broadly and effectively deployed in lots of media objects in a wide range of applications and they are:
‘ Content identification
‘ Content protection
‘ Forensics and piracy prevention
‘ Filtering of Contents (includes triggering and blocking of actions)
‘ Copyrights and Ownership protection
‘ Security of Document and images
‘ Verification of content and objects
‘ Broadcast monitoring
Image watermarking is a method of embedding the image with some secondary factor proposed for an authenticity of the ownership, without any loss in the quality of image and also to make legal affirmation based on the parameter, which can later be detected or extracted and this parameter is called as a watermark. Retrieval of watermark image unmistakably identifies the rightful owner [3] [4]. In general any watermarking system consists of:
‘ Watermark embedding method, which embeds the watermark into the data
‘ Watermark detector which verify the existence of watermark.
The inputs to the watermark embedding method are the watermark image, the cover image and the security key. The key used here is to increase the security level of the whole system. The watermark can be a sequence of numbers, a sequence of binary bits or may be any image. The output with the watermark embedding method is the watermarked data.
The image watermarking is method of embedding the image with some secondary parameter that is watermark, without failure in the quality of the image and to make permissible assertions based on the parameter, which can be detected (or extracted) by the owner.
An effective watermark should have a several properties ‘
‘ Robustness
The watermark should be reliably detectable after all the alterations to the marked document that means it should not easily be defeat.
‘ Imperceptibility
This means the quality of the marked document should not be degrade, the presence of watermark should not distort the original data.
‘ Security
This means that any unauthorized party should not read (or alter) the watermark data. The watermark should not be detectable. The security of the watermark is required for authentication and copyright protection.
‘ Fast (in embedding and retrieval)
The speed of a watermark embedding becomes necessary when the documents are distributed. Recovery algorithm of a watermark may give emphasis to reliable recognition over speed due to ownership disputes.
‘ No reference with original document
For some applications, it is necessary sometimes to recover the watermark image without requiring the original data, unmarked document (which would store in secure files).
‘ Multiple watermarks
It may also be desirable to embed multiple watermarks in the document. For instance, an image sometimes may be marked by a unique watermark each time it is downloaded.
‘ Unambiguous
A watermark must express unmistakable (unambiguous) information about the correct owner of a copyright data, point of sharing, etc.
Digital images are produced from the many resources, like everyday photographs taken, satellite viewed pictures, computer graphics etc… [5]. Watermarks for the natural images are typically modifying pixel intensities or we can say transform the coefficients, although it will be possible that a watermark could modify the other features like edges, textures. An image may be subject to great deal of manipulations like filtering, cropping of image, geometric transformations (such as rotation, scaling etc…), compressions, compositing with the other images and unreceptive (hostile) attacks. Thus we say that imperceptibility, security and robustness of image are mainly the most vital properties of the image watermarks also we say that speed, complexities are often secondary one.
Digital watermarking system is used in hiding of an undisclosed (secret) message or information inside an ordinary message or information and the extraction of it will takes place at its destination [6]. The secret message which embedded as a watermark will be anything, like a sequence of number, a plain text, or an image, etc… In general the digital watermarking system basically involves two main operations and they are:
‘ Watermark embedding
The watermark embedding algorithm is used to embed the watermark image with the main cover image.
‘ Watermark extraction
The watermark extraction algorithm is used to recover or extract the embedded watermark image from the main cover image.
And for both the operations to takes place, a secret key is required to secure the watermark data.
The keys in watermarking algorithms can be applied in the cryptographic mechanisms to provide more secure services to copy right protection.
Security mechanisms for the multimedia systems build on the basis of cryptographic algorithms and the digital watermarking methods or techniques [7]. The following security necessities, where cryptographic algorithms and digital watermarking methods or techniques are quite essential for the multimedia systems are:
‘ Confidentiality: The Cipher systems should keep the information in a secret manner from the unauthorized entities.
‘ Data integrity: The One way hash functions, digital signatures, message authentication codes, fragile digital image watermarking and robust digital watermarking should detect the data modification.
‘ Data-origin authenticity: The message authentication codes, fragile digital watermarks, digital signatures and robust digital image watermarks permit there proof of source (origin).
‘ Entity authenticity: The authentication protocols should ensure that an entity is the one which it claims to be.
Watermarking represents a resourceful technology for ensuring the integrity of data and authenticity of the main origin of data. With the use of watermarking one can embed the copyright or customer or integrity information within the data as the transparent patterns by using some secret key. Since we add the secured information within the data, we must ensure the privacy (confidentiality) of the secured information by using a secret key but remember not the privacy or confidentiality of the data itself. The concept of cryptographic algorithms are not been able to provide proper security to the watermarked data alone, so to make it more secure the new concept of using watermark embedding and cryptographic algorithm together is introduced. Many encryption algorithms or methods are broadly available and are used in the security of information [8]. Basically they can be broadly categorized into the Symmetric key (that is private key) and Asymmetric key (that is public key) cryptographic encryption algorithms. In the Symmetric key encryption method or we say in the secret key encryption algorithm, only one key can be used to encrypt the data as well as decrypt the data and the key would be distributed before get transmitted between the entities. Thus we can say that a Key plays a vital role. A key in cryptographic algorithm is a long sequence of bits used by encryption and decryption processes. Blowfish cryptographic algorithm has the better performance as compared to the other usually used encryption algorithms and since the Blowfish algorithm so far has not any identified security weak points, it is considered to be an excellent standard encryption cryptographic algorithm [9].
Blowfish cryptographic algorithm is a symmetric key block cipher algorithm which can be effectively used for the process encryption and decryption of data. Blowfish algorithm uses a key of variable length, starting from 32 bits up to 448 bits, which makes it an ideal for data security. The Blowfish algorithm was designed in the year 1993 by Bruce Schneier and defined as a fast and free alternative for the other existing encryption algorithms. The Blowfish algorithm is an unpatented and also license free, and it is available free of cost for all the users. Blowfish Algorithm is basically a Feistel network cipher which iterates a simple function of encryption around 16 times. The block size here is 64 bits and its key can be of any length up to the 448 bits. The Blowfish Algorithm can:
‘ Manipulates data in large blocks
‘ Has a 64-bit block size
‘ Has a scalable key, from 32 bits to at least 256 bits
‘ Uses only simple operations which are efficient for microprocessors
‘ Employs pre computable subkeys
‘ Consists of a inconsistent number of iterations
‘ Uses the subkeys that are one way hash of key
‘ Have no linear structures which reduces complexity of complete search
‘ Uses a design which is simple to understand
There are various watermarking hiding techniques have been evolved for embedding and extraction of watermark image in the recent years for the copyright protection [10] [11] [12]. One of the useful and simplest methods for hiding watermark image is LSB (Least Significant Bit). The Least Significant Bits or we say LSB hiding technique is the simplest and also straight forward approach for embedding or hiding a message within a cover main image. Here in the least significant bit method the eighth bit of the image is changed with a bit of the secret message. When we are using an image of 24 bit, we can easily store three bits into each pixel by altering a bit of each of the RGB colors (that is red, green and blue color) components, as they each are represented by a byte and each pixel of the image is consists of some variation of all the three base (primary) colors that is red, green and blue colors and requires 8 bits of each for all these three primary colors that is there will be 256 different variations possible, starting from 00000000 up to 11111111 in a pixel for each color. In the RGB model each pixel is made of the RGB values and each color requires 8 bit for their representation thus we can say each pixel is designated by 24 bits. Therefore total amount of color possible with this 24 bit RGB image is (28)3 equals 16,777,216 [13]. And, to signify a white color, code should look like this 11111111 11111111 11111111. The human eye are not been able to distinguish the dissimilarity between so many colors (for example the color 11111110 11111110 11111110 is look exactly same as the white color, which shows that, the last digit of every bit in the pixel would be changed without too much degradation (visual distortion) in quality [14] and this is the basis of Least Significant Bit Hiding technique.
The randomized least significant bit hiding technique is used here. This is the new concept for embedding and extraction of watermark image with the main cover image. As we know that the Least Significant Bit method hides the message in such a way that human eyes are not being able to distinguish it, but also it becomes possible for opponents to retrieve the message due to the simplicity of the LSB technique. Thus we can say that the malicious people easily try to retrieve the message if from the starting of the image they have doubt about that there exists some
secret data that was inserted within the image. Thus we say that there is a need to increase the LSB hiding technique. And thus the new concept of randomized least significant bit hiding technique is used here to recover from the above describe drawback of using direct least significant bit method for embedding the watermark image. In this technique the message is inserted within the images in some random manner into the pixels of the cover image.
The measurement of quality of an image is important for many applications of image processing. Image quality measurement is strongly related to the image similarity measurement where quality is totally based on various differences or we say similarity between the corrupted or distorted image and the original image or we say an unmodified image [15]. Digital images are focus on a large variety of damages during processing, its storage, its transmission any of it may result in a distortion or degradation of quality of image. Mean Square Error (MSE) and Peak Signal to Noise Ratio (PSNR) are commonly used to measure the degree or quality of image degradation or distortion [16] [17]. PSNR is commonly used to evaluate or measure the quality of the reconstruction of the lossy compression codecs (example for the image compression). Here in this case the main signal is original data and the noise will be the error introduce due to compression. While comparing the compression codecs PSNR will be estimation to the human observation of quality reconstruction. Basically PSNR and MSE are the two error measurement or metrics which are used to match (or compare) the quality of image compression. The MSE that is Mean Square Error represent the cumulative squared error among the compressed image and the original image whereas PSNR that is Peak Signal to Noise Ratio represents a measure of the peak error. Also higher the value of PSNR, the better will be the quality of the compressed image or reconstructed image and when the two images were identical, the value of MSE will become zero and the value of PSNR will be undefined that is infinity [18] [19].
PSNR is easily defined by the value of mean squared error (MSE). Given a noise free m x n image I and its noisy estimation K then MSE is given as:
And the PSNR is calculated or evaluated in decibels and also it is inversely proportional to the Mean Squared Error that is MSE value. It is given by the equation
Here, MAXI is the image’s maximum possible value of pixel and when these pixels are represented by 8 bits per sample then this is 255. For the color images having three RGB (that is Red, Green, Blue colors) values per pixel then the definition of the PSNR is same except that the value of MSE is sum over all the squared value differences divided by the image size and by three.
A lot of work has been done in the field of watermarking (digital image watermarking) for receiving the proper security from an illegal or unauthorized copying of the data. Digital watermarking has been considered as a method used for the purpose of copyright protection and also for authenticity. There are various research work has been done to focus completely on improving the security of watermarks. They mainly work on increasing the embedding capacity of the watermark with the cover image so as to enhancing the security thereby.
The purpose of digital watermarking is to provide copyright protection (security) for intellectual property of owner. Copyright protection of such data has been done by using cryptographic algorithms, but the cryptographic algorithms are not providing complete solution and thus digital watermarking arises with the solution related to the copyright protection.
In the year 2013, J S Bhalla and P Nagrath [20] had introduced the concept of Nested Digital Image Watermarking technique using the blowfish algorithm. The research work basically focused on increasing the embedding capacity and improving security of watermarks by introducing the concept of nested image watermarking for embedding and extraction. Here they proposed a technique of nested digital watermarking for watermark embedding and watermark extraction. The concept of ‘Nested’ here basically means a watermark gets embedded with another watermark. One watermark gets encrypted (using the Blowfish cryptographic algorithm) and then embedded into another watermark image and this nested watermark is again encrypted and finally gets embedded with main image. With this they provide an extra level of security for watermarks and also increase the embedding capacity of a watermark as compared to the others where single watermark use to embed an image for copyright protection.
In the year 2013, P Singh [21] introduced the research work on image encryption and decryption using blowfish algorithm in Matlab. The algorithm was basically developed to encrypt plaintext of 64 bits into cipher text of 64 bits, efficiently and securely. The operations table lookup, modulus, bitwise X-OR and addition were selected for the algorithm to minimize the time (required to encrypt and decrypt data) on the 32 bit processors. A special attempt was made in the designing of algorithm so as to keep its operations simple and also easy to code while not compromising with the security. Both the color and black & white images of any size saved in different formats like TIF, bmp, PNG, jpg etc…. be encrypted & decrypted using blowfish and will not be broken down until an attacker tries (28r+1 combinations where r is the number of rounds). Hence if no. of the rounds were getting increased then it shows that blowfish becomes a stronger method and as Blowfish does not have any weak points (in terms of security) so it is considered to be an excellent encryption algorithm.
In the year 2012, P Gupta [22] had introduced the concept of Digital image watermarking algorithm based on cryptographic algorithm for to enhancing the security of watermarked data by taking the concept of blind watermarking which uses basically nesting of the watermark and its encryption. Here nesting means it will embed an extra watermark image onto the main watermark image and then embed the main watermark image with the main cover image. Here for encryption, the exclusive-OR operation was used. For embedding the watermarked watermark into Cover Image, here DWT technique was used and thus original image will not required at the time of recovery.
In the year 2012, O Mohammad and A Al-Hazaimeh [23] had introduced the concept of hiding the data in Images Using a New Random Technique. In this method, the message gets inserted between the images in some randomized manner within the pixels of the cover image (main image) and least significant bits hides these messages in such a manner that it becomes difficult to determine it. In this paper the insertion of message bits into the image will not only within the least bit but also in other bits pixel in the random fashion. It can be done by compare the message bits to the pixel bit randomly chosen (from second to the last bit) and based upon this comparison, value 1 will be inserted in the LSB (if the message bit is identical to that of image) and 0 will be inserted if the message bit is not matched with the chosen bit.
In the year 2011, the research work was done by the authors S P Singh, R Maini [24] which shows a comparison (performance comparison) amongst most of the common encryption cryptographic algorithms (DES, 3DES, Blowfish and AES). The result obtained shows that blowfish algorithm has the better performance when compared with other encryption algorithms (that is here DES, 3DES and AES) and it has no any known security weak points so far and the comparison was done on the basis of the processing (different sizes of data blocks) by using the algorithms (to calculate their encryption/decryption speed).
In the year 2011, D Biswas, D Sarkar and A Pal et al. [25] had done their research work on comparison and analysis of watermarking algorithms in color images ‘ Image Security Paradigm. They suggests that by using the direct LSB hiding technique has some disadvantage of having the higher complexity and its dependency in the form of image while the Randomized Least Significant Bit hiding technique scores more in terms that it is having the lesser complexity and is more robust for variations in the type of the image by using PSNR that is Peak Signal to Noise Ratio into the consideration.
In the year 2012, Y Perwej, F Parwej and A Perwej [26] had introduced the concept of An Adaptive Watermarking Technique for the copyright of digital images and Digital Image Protection. Here they are proposed the edge detection from the Gabor Filter method. In this paper the simple LSB substitution method are used for data hiding. In this method, a set of pixels which constitute a block will jointly share the bits from the watermark and the values for mean square error (MSE) and peak signal to noise ratio (PSNR) are calculated. The result indicates the method which introduces a low noise and thus ensures the lesser visible distortions.
In the year 2007, K M Singh, S B Singh and L S S Singh [27] had introduced the concept of Hiding Encrypted Message in the Features of Images. Here the message to make secret in the image was firstly get encrypted using S-DES algorithm. And then the features like edges, thin straight lines, corners, end of lines etc…. get detected from the cover image by using Roberts edge detection method and random pixel locations were found in the cover image by PRNG. Now the message bits get embedded on the random edge pixel locations (using LSB insertion algorithm). Here the proposed method is named as the Random Edge LSB (that is RELSB) method. The proposed method is not requiring the original image for the process of extraction of secret message and also ensures that edge pixels were detected before embed the secret message are not unlike from the points detected after its embedding.
In the year 2012, M A Kumar and Dr. S Karthikeyan [28] had done their research work on Investigating the Efficiency of Blowfish and Rejindael (AES) Algorithms. Here the performance assessment of two common symmetric key cryptographic algorithms had done. These algorithms were tested upon various performance metrics. The simulation result here shows that Blowfish have the better performance as compare to AES. There were no major differences in the result while considering base64 encoding and the hexadecimal encoding methods and also it was found that blowfish results good for text based encryption process and AES have better result for image based encryption. It also had identified that there were change in the performance of AES when key size changes. Overall it shows that AES can be used in where high security is needed and Blowfish can be used in the case of performance aspects.
In the year 2011, Least Significant Bit (LSB) based Lossless Digital Image Watermarking using Polynomials in a Spatial Domain for DRM proposed by A. Siva Sankar, T. Jayachandra Prasad and M. N. Giri Prasad [29] proposes a new embedding method where they randomly hide the messages in Least Significant Bit (that is LSB) of any component of the selected (chosen) pixel using polynomial thereby probability of getting right coefficients suitably is difficult.
In the year 2013, G Kaur and K Kaur [30] had proposed Image Watermarking Using LSB (Least Significant Bit). There are lots of techniques were used in watermarking for security of the images like Frequency domain, Spread Spectrum and Spatial domain. In this paper spatial domain technique Least Significant Bit used for the security of images, which is easy and simple method and also more effective method. The process of Least Significant Bit is simple when it will be used in MATLAB and a different image in MATLAB tells different steps of process and their result.
Watermarking belong to a group of information hiding. Digital watermarking proves to be a good means for protecting the intellectual property from illegitimate copying or we can say that it provides a required copyright protection. Digital watermarking provides a way of embedding any message or information in a section of digital data without destroy its original value.
The work entitled ‘Nested digital image watermarking using randomized LSB hiding algorithm’ is based on the concept of digital image watermarking. Here a watermark is embedded into another watermark which is considered as the main watermark and then this main watermark is embedded with the cover image (main image). The main concept behind this work was taken from the various concepts.
The main problem observed that the use of the direct LSB (Least Significant Bit) hiding technique for the process of embedding of the watermarked image with another image has some drawback of having higher complexity and its dependency upon the type of the image whereas the Randomized Least Significant Bit scores more when compared with the direct LSB hiding technique as it is having lesser complexity as compared to LSB and also the approach of Randomized LSB hiding technique is more robust towards the variations in the type of image. The main focus emphasize while working on the concept of digital image watermarking is the security of the image from various attacks. There are various cryptographic algorithms, but the security is the main issue, and some research work done yet had also shown that the blowfish algorithm has no known security weak points so far as compare to the other cryptographic algorithms. Thus to enhancing the embedding capacity of watermark and also increasing the security of the image will become the great issue so far.
A solution to the above problem is to propose the technique which uses the cryptographic algorithm and the hiding technique in an effective manner that is they must provide maximum security. The solution for the above discussed problem is we will get by using the cryptographic algorithm for encryption process before the embedding process. Nesting of watermark image that is when one watermark embed into another and gets encrypted before embedding. This will definitely be enhancing the embedding capacity as well as increasing the security of the image.
The main focus wholly emphasizes onto improving or we can say rising the security of the watermark image by increasing its embedding capacity. The digital watermarking techniques are used to give copyright protection to the intellectual properties of the owners which is mainly in the digital form. So we can say that basically there were two challenges for what we have to take concern about are-
‘ How to maintain the illegal recipients (means those who do not have the right authority) to get the ‘copyrighted’ digital work.
‘ How to avoid the ‘illegal redistribution of data’ by the lawful recipients of the copyrighted digital data.
The hiding technique LSB (Least Significant Bit) for embedding as well as extraction of the watermark image was used for the digital image watermarking. But as the concept of the direct LSB hiding method has some drawbacks as it is having greater complexity as well as its dependency on type of image also whereas Randomized LSB hiding method scores more because it is having lesser complexity and is quite strong (we say robust) with some variations in image type with respect to their PSNR values (Peak Signal to Noise Ratio value) computation. And also the combination of cryptographic algorithms and hiding techniques together will definitely provide more security to the image. And if encryption cryptographic process can be applied only some of selected pixels then the security will definitely get increased as when get applied on the whole image.
This section provides the methodology which uses the concept of nested image watermarking for the process of embedding of watermark image as well as for the extraction of that watermark image. Nesting of the watermark image means when one watermark image will get embedded with another image (watermark image). The concept of embedding one watermark image with another watermark image will be used here to increase the security level by enhancing the embedding capacity of the watermark image. If we apply only embedding techniques for hiding the watermarking image into the cover image that is the main image, then it might not provide good security level to the cover main image and also if we use to apply only the cryptographic algorithm alone then they are not able to provide the complete solution for the problem related to copyright and hence the security of cover main image is not proper. For this reason the concept of nested digital image watermarking is used here.
The nested digital image watermarking using randomized LSB hiding method basically comprises of the following methods described below.
Cryptography is the art and science of achieving security by encoding messages to make them non-readable. Clear text or the plain text signifies a message that can be understood by sender, recipient and also by anyone else (who gets an access to that message). And when this plain text gets codified using any suitable scheme, then the resulting message is known as cipher text. And the process of encoding plain text messages into cipher text messages is called encryption process and the reverse process of transform the cipher text messages back to the plain text messages is called decryption [32].
Here we are using Blowfish cryptographic algorithm to perform the encryption and decryption process. Blowfish is a symmetric block cipher which can be effectively used for the encryption and safeguarding of the data. It takes a key of variable-length, starting from 32 bits up to 448 bits, which makes it an ideal for securing data. Blowfish was developed by Bruce Schneier and has the reputation of being a very strong symmetric key algorithm as it is a fast, free alternative to the other existing encryption algorithms. Blowfish algorithm is a symmetric key algorithm with following features:
‘ Fast ‘ Blowfish encryption algorithm rate on a 32-bit microprocessor is 26 clock cycles per byte.
‘ Compact ‘ Blowfish algorithm can execute in less than a 5 kb memory.
‘ Simple ‘ Blowfish algorithm uses only primitive operations that are addition, XOR (Exclusive OR), which makes its designing and its implementation simpler.
‘ Secure ‘ Blowfish algorithm is having a variable key length up to a maximum of 448 bits, which makes it both flexible as well as secure.
The Blowfish algorithm is a Feistel Cipher having 16 round and also it uses large key dependent S-boxes. In terms of structure it resembles to the CAST-128, which also uses fixed S-boxes. The fig 4.1 shows the action of the Blowfish algorithm or we can say Feistel structure of the blowfish algorithm. Here each line represents the 32 bits and the blowfish algorithm keeps two subkey arrays, which are 18 entry P-arrays and four 256 entry S-boxes. The S-boxes will accept the 8 bit input and produces 32 bit output. Here one entry of the P-array will be used in every round and after the final round; each half of the data block will be XORed with the one of the two remaining P-entries which is unused.
The Fig 4.2 shows Blowfish algorithm Feistel function (Round Function) which splits the 32 bit input in four 8 bit quarters, and then uses this quarters as an input to S-boxes. And the outputs are the added modulo 232 and the XORed to produce final 32 bit output.
The basic algorithm for Blowfish Encryption process is:
Blowfish has total 16 rounds and the input is a 64 bit data element termed x.
1. Divide x in two 32 bit halves xL and xR.
2. for i = 1 to 16
3. xL = xL XOR Pi
4. xR = F(xL) XOR xR
5. Swap xL and xR
6. After the sixteenth round, swap xL and xR again to undo the last swap.
7. xR = xR XOR P17 and xL = xL XOR P18.
8. Finally, recombine xL and xR to get the ciphertext.
The decryption of the blowfish algorithm is exactly same as the encryption process, excluding the fact that P1, P2… P18 get used in reverse order.
In general the process of encoding the plain text into the cipher text is called as encryption process. Here the original data image is a plaintext and both the plain text and encryption key are the two inputs of the blowfish encryption process. And also here in this case, the original image data bit stream will be divided into blocks length of the Blowfish algorithm. The byte elements of array will be stored in the row order that is from left to right and each row represent one scan line of image and rows of the image will encrypted from the top to the bottom.
The process of transforming the cipher text back into the plain text is called the decryption process. Here the encrypted image will be divided in the similar block length of the Blowfish algorithm from the top to bottom. The first block will enter to decryption function and also the similar encryption key is used always to perform the decryption process of the image but the applications of the sub keys are reversed and the process of decryption is continue with the other blocks of image from the top to bottom.
There are various watermarking hiding techniques have been evolved for embedding and extraction of watermark image in the recent years. One of the useful and simplest methods for hiding watermark image is LSB (Least Significant Bit) [31]. In this method, the least significant bits or we can say bits of a pixel get replaced with the bits of data which is to be hidden. Least Significant Bit will be extended up to the 4 least significant position from a byte that means we will be able to replace four bits of the hidden data with original values of a pixel (whose binary value consists of 8 bits). So, from 8 bits, maximum 4 bits will be used to hide the data.
Here we are using the concept of Randomized Least Significant Bit hiding technique for the embedding and extraction of the watermark image. As it has the complexity quite lesser and also it is stronger towards the variations in image types by considering PSNR values. In a RLSB, Randomized Least Significant Bit hiding technique, firstly we have to generate m random nos. which acts as a secret key. Now next we apply the LSB hiding technique onto some selected pixels whose values can be determined by the secret keys. Depending upon those values of m, the Least Significant Bit hiding technique will be applied either into the rows or into the columns (or we can say either horizontally or vertically). The RLSB hiding technique is powerful in terms that it performs the embedding process only in some of the selected pixels.
The Least Significant Bits (LSB) method is the simplest and the most straight forward method for embedding or hiding a message within a cover image (main image). As LSB method hides the messages in such a way that humans are not able to differentiate it, but still it is possible for any opponent to retrieve the entire message because of the simplicity of the LSB hiding technique. This problem would be rectified by using the technique also known as RLSB hiding technique. In this technique the message is inserted within the images in some random manner into the pixels of the cover image. Here we are using the RGB (Red, Green, and Blue) image for the cover image and watermark images and thus it is having 3 color channels so it is possible to store three hidden bits of information in each pixel. And this representation can be shown as follows:
Here we are using the Randomized LSB hiding technique for embedding and extraction of the watermark image. In the methodology of RLSB hiding we are using the cover image and nested watermark image (which we have to embed with the main cover image). Nested Watermark means when one watermark image gets embedded with the other watermark image. The main concepts behind the embedding and extraction process which are used here have some constraint. And as we know that value of 1 pixel equivalent to 8 bit thus we can say that for the process of embedding or extraction, last 2 bits will be used in making 1 pixel. And thus in other words we can also say that to encrypt 1 pixel we need 8 bits.
In RGB images (that is Red-Green-Blue images) every pixel is having 24 bit value in which (00000000-00000000-00000000) represents black color and (11111111-11111111-11111111) represents the white color. This 24 bit is divided into three divisions having 8 bits each where the 1st 8 bits were represented as the Red color, the 2nd 8 bits were represented as the Green color and finally the last 8 bits were represented as the Blue color. So we can say that a single bit is a combination of these 3 colors that is RGB. In the RGB model each pixel is basically composed of the RGB values and 8 bits required by each of these colors for their representation. And thus each pixel will be represented by the 24 bits. So we can say that total number of color possible is (28)3 equals 16,777,216 by 24 bit RGB image. In LSB method of the 24 bit color image, the least significant bit of each pixels of either specific color channel or we say all color channels were replaced by a bit from the secret image.
So here we are using the random chosen pixels of image also can be termed as randomized LSB hiding technique for embedding process and extraction process of the nested watermark image with the main cover image.
The watermark embedding algorithm is used here to embed the watermark image. Here in this methodology, firstly the encrypted watermark is embed with the other watermark image also called as nested watermark image and then this nested watermark is then embed into the main cover image. The Randomized Least Significant Bit hiding technique is used here for the process of embedding.
Basic abbreviations:
WM1: Watermark 1 (Primary Watermark Image)
WM2: Watermark 2 (Secondary Watermark Image)
Cover Image: Main Original Image
Key 1: Key for encrypting WM2
Key 2: Key for embedding encrypted WM2
Key 3: Key for encrypting Nested Watermark
Key 4: Key for embedding encrypted Nested Watermark
Algorithm:
1. Some selected pixels of Watermark 2 (WM2) are encrypted using Blowfish algorithm with Key 1. Encrypted Watermark 2 (EWM2) is generated as output.
2. Some selected pixels of Encrypted Watermark 2 (EWM2) are embedded into Watermark 1 (WM1) using RLSB technique with Key 2. Output image is Nested Watermark (NWM).
3. Again some selected pixels of Nested Watermark (NWM) are encrypted using Blowfish algorithm with Key 3. Encrypted Nested Watermark (ENWM) is generated as output.
4. Finally some selected pixels of the Encrypted Nested Watermark (ENWM) are embedded into the Cover Image (Main Original Image) using RLSB technique with Key 4. Output image is Watermarked Cover Image.
The watermark extraction algorithm is used here to recover or extract the embedded watermark image. Here in this methodology firstly the encrypted nested watermark image is extracted or retrieved from the cover main image and then the encrypted watermark 2 is being retrieved from the nested watermark the randomized least significant bit hiding technique is used for the process of extraction.
Basic abbreviations:
WM1: Watermark 1 (Primary Watermark Image)
WM2: Watermark 2 (Secondary Watermark Image)
Watermarked CI’: Already Watermarked Image
Key 1: Key for decrypting WM2 from the cover image
Key 2: Key for extracting encrypted WM2
Key 3: Key for decrypting Nested Watermark
Key 4: Key for extracting encrypted Nested Watermark
Algorithm:
1. Extract the encrypted nested watermark ENWM from the original cover image using RLSB technique with Key 4. Output is the ENWM’.
2. Decrypt ENWM’ using the blowfish algorithm with the Key 3. NWM’ image is generated as output.
3. Again Extract the EWM2 that is encrypted watermark 2 from the NWM’ image using RLSB technique with the Key 2. Output is the EWM2′.
4. Finally decrypt the EWM2′ using the blowfish algorithm with the Key 1. WM2 is generated as output.
The process of extraction or we can say the process of retrieving the watermark image from the cover main image is just opposite process of the embedding process. The same keys are used for the process of decryption and extraction which is used for encryption and embedding processes.
The main methodology used here is the technique of the nested digital image watermarking using randomized least significant bit hiding technique for embedding and extraction of the watermark image. Nested watermark image means when one watermark image is embedded into the other watermark image. In this methodology the encryption as well as embedding algorithm both is used. Here a new concept is used in which some selected pixels of a watermark image is first encrypted and then embedded with the other watermark image and then this combined watermark image also termed as the nested watermark image will be encrypted again (only some selected pixels) and then finally embedded into the main cover image. As one watermark image is embedded into the other watermark image, thus this technique termed as the nested digital image watermarking. Here by using the concept of nested digital image watermarking will increase the embedding capacity of the watermark image and also the concept of encryption before embedding will also increases the security of the watermark image. Here in this methodology, we use to perform the embedding as well as encryption only on some selected pixels or we can say randomly chosen pixels thus the security will increase and it becomes difficult for the attackers to determine the values of the random selected pixels.
The step by step process used in the methodology here is described below:
1. A watermark image that is Watermark 2 (WM2) is encrypted using the blowfish cryptographic algorithm. Here whole image is not encrypted; but encryption is done only on some of the selected pixels. The encryption is done by using a key termed Key 1. The encrypted watermark that is EWM2 is generated as an output.
2. Next, Encrypted Watermark Image that is EWM2 is embedded into a watermark 1 image that is WM1 using Randomized Least Significant Bit hiding algorithm, here also the hiding is done on some selected pixels using a Key 2. The nested watermark image NWM is generated as an output.
3. The Nested Watermark image that is NWM (one watermark image is embedded into another watermark image) is encrypted using blowfish algorithm (only some selected pixels are encrypt) using a Key 3.
4. Now, some selected pixels of the encrypted nested watermark image that is ENWM is embedded into the main cover image using the Randomized Least Significant Bit hiding method by using Key 4. And here comes the final embedding of the nested watermark image into the main cover image.
5. After completing the process of embedding of nested watermark image into the main cover image, the next process is to Retrieve or Extract the encrypted nested watermark that is ENWM from the original main cover image using Randomized Least Significant Bit method with Key 4. Output is the ENWM’ which is recovered image.
6. Now next is to Decrypt the recovered image in the previous step that is ENWM’ using the blowfish algorithm by using the Key 3. NWM’ image is generated as output after decryption.
7. Now again Extract the EWM2 that is encrypted watermark 2 from the NWM’ image using Randomized Least Significant Bit technique with the Key 2. Output generated is the EWM2′.
8. Finally decrypt the EWM2′ by performing the decryption process using the blowfish cryptographic algorithm with the Key 1 and WM2 is generated as output.
In this chapter we present the experiment performed to validate the proposed method. The work has been implemented and realized using MATLAB. A common need of watermarking of an image is to provide copyright protection or we say to provide security to the watermarked image. The cryptography is also one of the field which works on to increasing the security. As cryptography is the art of achieving the security by encoding the messages so to make them unreadable. But as the cryptographic algorithms are not been able to provide the complete solution therefore the concept of digital watermarking are used which being helpful in providing the proper solution for one’s intellectual property and also being used for identifying the rightful owner of a data. Thus our main focus is on to achieving the goal of improving the embedding capacity of the watermark image and increasing its security from the unwanted attacks for retrieving the watermark image from the main cover image. Moving in the same way for achieving this we focuses on to improving the embedding capacity of watermark in such a way that it being quite difficult to extract it easily.
In this experiment the result is observed on the basis of two measurements of errors and that are PSNR means Peak Signal to Noise Ratio calculation and MSE that means Mean Square Error calculation. These both that is MSE and PSNR are the error metrics which are used to compare the compression quality of the image. Lower the value of MSE and higher the value of PSNR means higher the quality of the image.
Basically there are four main processes which we have used during our experiment and they are as follows:
‘ Encryption of the image
‘ Embedding of watermark image
‘ Decryption of the encrypted image
‘ Extraction of the watermarked image
For performing the experiment we have three color images with us. The Cover Image, which is the main original image which we have to protect then two watermark images that is Watermark Image 1 and Watermark Image 2. We perform all the processes mentioned above and calculate the result in terms of MSE and PSNR by testing the resultant images at the different conditions and they are:
‘ When No Attack takes place
‘ When Rotation Attack takes place
‘ When Noise Attack takes place
‘ When Cropping of image takes place
We determined the values of MSE and PSNR for all the conditions mentioned above and on the basis of that result conclusion had been drawn.
We perform the experiment with different sizes of cover images, watermark1 images, and watermark2 images.
1) The work has been done by taking the cover image of size 1000 x 1000 pixels, watermark1 image of size 248 x 252 pixels and watermark2 image of size 60 x 60 pixels. And after performing various processes on original image, watermark1 image and watermark2 image, the various effects are:
In Fig 5.2 a) is the Encrypted Watermark2 Image b) is the Embedding of previous encrypted image into Watermark1 Image c) is the encrypted nested watermark image d) is the final embedding with the cover image e) is the extraction of watermarked image from cover image f) is the decrypted nested watermark image g) is the extraction of encrypted of watermark 2 image h) is the decryption of watermark2 image.
And the result obtained in terms of MSE and PSNR will be shown below:
2) The work has also been carried out for cover image of size 900 x 900 pixels, watermark1 image of size 224 x 226 pixels and watermark2 image of size 56 x 56 pixels. And after performing various processes on original image, watermark1 image and watermark2 image, the various effects are:
3) Similar work has also been carried out for cover image of size 800 x 800 pixels, watermark1 image of size 224 x 226 pixels and watermark2 image of size 56 x 56 pixels. And after performing various processes on original image, watermark1 image and watermark2 image, the various effects are:
Its been observed from the calculations or we say from the result obtained, the values of the MSE and the PSNR changes according to all the variations from the types of attacks applied on the Original Image which is Cover Image, Watermark1 Image (first watermark) and Watermark2 Image (second watermark) respectively. The result obtain shows the value of MSE is almost tends to zero and also the value of PSNR is high. And thus we can conclude that quality of the image we get will be higher. The PSNR calculation gives the measurement of degradation (distortion) of the carrier signal after hiding the information. Here the signal is the original data, and noise is error which introduced by the compression. Therefore, higher the value of PSNR means better is the quality of the compressed image (reconstructed image). And also when two images are identical then value of MSE will be zero and for this PSNR value should be infinity.
It also observed from the experiment that value of MSE and value of PSNR for the Original Image and the Watermark1 Image when go through various attacks will give us better result since their MSE values we obtained are very low (or we say approximately tends to zero) and PSNR values we obtained are high, which shows that the image quality remains unaffected after reconstruction. But also we have observed that when Watermark2 image when go through the various attacks, it does not give us the better result which we had saw for the watermark1 image and original cover image. And this happens because watermark2 image is going through the two embedding processes and two encryption processes and therefore its pixels gets disturbed (distorted) more while compare with the original image and watermark1 image.
The concept of digital image watermarking is a new concept for researchers who mainly focus onto improving the watermark techniques which already exist. The problem of copyright protection of one’s intellectual property is the main issue for which lots of research works are done now days. Also the main focus in the field of image watermarking techniques is how to prevent the image from the unwanted attacks. The image watermarking technique used here is contributed in this field of research work in which we tries to improve the existing image watermarking techniques.
The thesis work is carried out for cover images of different size and watermark images of different sizes (that is watermark1 image and watermark2 image). Each image results in different values for MSE and PSNR and after go through all the different type of attacks here it is (Rotation, Noise and Cropping), the retrieved extracted image is not degraded with its origin. The quality of image remains unaffected (if viewed by the human eyes). In the method used the main focus towards the security of the image without degrading or destroying quality of the original cover image. The basic advantage of the technique used here is that it uses concept of the nested digital image watermarking method and Randomized Least Significant Bit hiding method for embedding of the watermark image and extraction of the watermark image and the blowfish algorithm used for the process of encryption and for the process of decryption.
The main constraint here or we say main advantage here is that all the embedding and extraction technique as well as encryption and decryption technique is not applied within the whole image; but is applied only into some selected pixels. And therefore we say that it provide greater security to the images as it is quite difficult to find out the values of the selected pixels as they are chosen randomly. Also as the process of encryption and for decryption are done by using a blowfish algorithm so as to provide greater security, before the embedding of watermark image and extraction of watermark image respectively.
Here we had performed an experiment to improve the embedding capacity of watermark image and also for enhancing the security of the image. In this experiment we had applied the encryption process to the watermark image before it is embedded into the original cover image. And also the concept of nested watermark is used here which increases the embedding capacity of the watermark image. In this experiment the quality of image quite less damaged after go through all the processes and also through the various attacks when computed their MSE and PSNR values. In future some other cryptographic method might used and also some other watermarking technique should be used which results better in terms of embedding capacity, improving the security, not gets distorted with the attacks and results in more higher PSNR and also lesser MSE values.
The concept of digital image watermarking is a new concept for researchers who mainly focus onto improving the watermark techniques which already exist. The problem of copyright protection of one’s intellectual property is the main issue for which lots of research works are done now days. Also the main focus in the field of image watermarking techniques is how to prevent the image from the unwanted attacks. The image watermarking technique used here is contributed in this field of research work in which we tries to improve the existing image watermarking techniques.
The thesis work is carried out for cover images of different size and watermark images of different sizes (that is watermark1 image and watermark2 image). Each image results in different values for MSE and PSNR and after go through all the different type of attacks here it is (Rotation, Noise and Cropping), the retrieved extracted image is not degraded with its origin. The quality of image remains unaffected (if viewed by the human eyes). In the method used the main focus towards the security of the image without degrading or destroying quality of the original cover image. The basic advantage of the technique used here is that it uses concept of the nested digital image watermarking method and Randomized Least Significant Bit hiding method for embedding of the watermark image and extraction of the watermark image and the blowfish algorithm used for the process of encryption and for the process of decryption.
The main constraint here or we say main advantage here is that all the embedding and extraction technique as well as encryption and decryption technique is not applied within the whole image; but is applied only into some selected pixels. And therefore we say that it provide greater security to the images as it is quite difficult to find out the values of the selected pixels as they are chosen randomly. Also as the process of encryption and for decryption are done by using a blowfish algorithm so as to provide greater security, before the embedding of watermark image and extraction of watermark image respectively.
Here we had performed an experiment to improve the embedding capacity of watermark image and also for enhancing the security of the image. In this experiment we had applied the encryption process to the watermark image before it is embedded into the original cover image. And also the concept of nested watermark is used here which increases the embedding capacity of the watermark image. In this experiment the quality of image quite less damaged after go through all the processes and also through the various attacks when computed their MSE and PSNR values. In future some other cryptographic method might used and also some other watermarking technique should be used which results better in terms of embedding capacity, improving the security, not gets distorted with the attacks and results in more higher PSNR and also lesser MSE values.
Watermarking is an act of hiding the information. It simply referred to as a pattern of bits inserted into a digital data (image, audio or video file) which identifies the copyright information (such as rights, author, etc’) of files. A digital watermark is a signal permanently embedded into digital data (audio, images, video, and text) that can be detected or extracted further by means of computing operations in order to make assertions about the data. The hiding process is to be such that the modifications of the media should be unnoticeable means for the images the modification in the pixel values have to be invisible. And image watermarking is method of embedding the image with some secondary parameter that is watermark, without failure in the quality of the image and to make permissible assertions based on the parameter, which can be detected (or extracted) by the owner. Watermarks for the natural images are typically modifying pixel intensities or we can say transform the coefficients, although it will be possible that a watermark could modify the other features like edges, textures. An image may be subject to great deal of manipulations like filtering, cropping of image, geometric transformations (such as rotation, scaling etc…), compressions, compositing with the other images and unreceptive (hostile) attacks. With the use of watermarking one can embed the copyright or customer or integrity information within the data as the transparent patterns by using some secret key. Since we add the secured information within the data, we must ensure the privacy (confidentiality) of the secured information by using a secret key but remember not the privacy or confidentiality of the data itself. The concept of cryptographic algorithms are not been able to provide proper security to the watermarked data alone, so to make it more secure the new concept of using watermark embedding and cryptographic algorithm together is introduced. The survey part deals with the survey of various digital image watermarking techniques which uses cryptographic algorithms also. A lot of work was done in this field yet. All the work under this field greatly emphasizes on to increasing the security level of watermark. But to provide maximum security and also increasing the embedding capacity of watermark by using concept of nesting the watermark technique and cryptographic algorithm together will raised some great possibility. The lots of concepts were used previously. And the most important concept or we say hiding technique used was the direct LSB hiding technique for embedding and extraction of the watermark image. The main problem observed that the use of the direct LSB (Least Significant Bit) hiding technique for the process of embedding of the watermarked image with another image has some drawback of having higher complexity and its dependency upon the type of the image whereas the Randomized Least Significant Bit scores more when compared with the direct LSB hiding technique as it is having lesser complexity as compared to LSB and also the approach of Randomized LSB hiding technique is more robust towards the variations in the type of image. The main focus emphasize while working on the concept of digital image watermarking is the security of the image from various attacks. There are various cryptographic algorithms, but the security is the main issue, and some research work done yet had also shown that the blowfish algorithm has no known security weak points so far as compare to the other cryptographic algorithms. Thus to enhancing the embedding capacity of watermark and also increasing the security of the image will become the great issue so far.
Here in the thesis the new concept of nesting of image watermarking is used in which one watermark image is embedded within another watermark image and again this nested watermark image is embed into the cover original image (that is the main image). Here the RLSB technique that is Randomized Least Significant Bit Hiding algorithm is used for embedding one image with another and blowfish cryptographic algorithm is used for to encrypt the watermark image before embedding it with the cover original image (that is the main image). The encryption of the watermark image is done before it embeds into the main original image (that means cover image) so as to ensure higher security of the watermark image. And also as the concept of the direct LSB hiding method has some drawbacks as it is having greater complexity as well as its dependency on type of image also whereas Randomized LSB hiding method scores more because it is having lesser complexity and is quite strong (we say robust) with some variations in image type with respect to their PSNR values (Peak Signal to Noise Ratio value) computation. The main methodology used here is the technique of the nested digital image watermarking using randomized least significant bit hiding technique for embedding and extraction of the watermark image. Nested watermark image means when one watermark image is embedded into the other watermark image. In this methodology the encryption as well as embedding algorithm both is used. Here a new concept is used in which some selected pixels of a watermark image is first encrypted and then embedded with the other watermark image and then this combined watermark image also termed as the nested watermark image will be encrypted again (only some selected pixels) and then finally embedded into the main cover image. As one watermark image is embedded into the other watermark image, thus this technique termed as the nested digital image watermarking. Here by using the concept of nested digital image watermarking will increase the embedding capacity of the watermark image and also the concept of encryption before embedding will also increases the security of the watermark image. Here in this methodology, we use to perform the embedding as well as encryption only on some selected pixels or we can say randomly chosen pixels thus the security will increase and it becomes difficult for the attackers to determine the values of the random selected pixels. And the measurement of quality of an image is important for many applications of image processing. Image quality measurement is strongly related to the image similarity measurement where quality is totally based on various differences or we say similarity between the corrupted or distorted image and the original image or we say an unmodified image. Here the result is calculated on the basis of MSE (Mean Square Error) and PSNR (Peak Signal to Noise Ratio) which we obtained when the processes under go through various attacks like rotation, noise and cropping. Higher the value of PSNR, the better will be the quality of the compressed image or reconstructed image and when the two images were identical, the value of MSE will become zero and the value of PSNR will be undefined that is infinity.
Watermarking is a method of hiding digital data or information. Watermarking is an embedding method of original cover image (that is the main image) by some parameter which is known as watermark image. It is used to provide copyright protection means from the illegal copying without degrading the quality of the image. Copyright protection is used to protect one’s intellectual property. Here in the thesis the new concept of nesting of image watermarking is used in which one watermark image is embedded within another watermark image and again this nested watermark image is embed into the cover original image (that is the main image). Here the RLSB technique that is Randomized Least Significant Bit Hiding algorithm is used for embedding one image with another and blowfish cryptographic algorithm is used for to encrypt the watermark image before embedding it with the cover original image (that is the main image). The encryption of the watermark image is done before it embeds into the main original image (that means cover image) so as to ensure higher security of the watermark image. The research work is mainly focuses on to receiving the highly secured watermark image. Here we do not perform the encryption or decryption process to the whole image but we perform these operations to only on some of the selected pixels. The result is calculated on the basis of MSE (Mean Square Error) and PSNR (Peak Signal to Noise Ratio) which we obtained when the processes under go through various attacks like rotation, noise and cropping. And higher value of PSNR and lower value of MSE results in the less degraded image.