Types Of Digital Watermarking Techniques Computer Science Essay

Digital watermarking is a procedure of implanting information, i.e. , water line, into a multimedia component to guarantee the cogency of the signal. The multimedia component may be an image, sound, picture or text. The water line which has been embedded into the multimedia component can subsequently be extracted for security intents. In general, digital watermarking can be subdivided into assorted types such as seeable and unseeable watermarking, robust and delicate watermarking, blind and non-blind watermarking.

Besides that, there are two wide classs of watermarking techniques, viz. spacial sphere watermarking and frequence sphere watermarking. Compared to spacial sphere techniques, frequence sphere watermarking techniques has been proved to be more effectual and at the same clip accomplish the imperceptibility and hardiness demands of digital watermarking algorithms ( Al-Haj, 2007 ) . This undertaking focuses on frequence sphere techniques, i.e. , DCT-based digital watermarking.

Digital watermarking techniques have been used in assorted applications, for illustration right of first publication protection, fingerprint protection, content hallmark, broadcast monitoring, transcript control, and biometric protection. What we will concentrate on this undertaking will be how to use watermarking techniques for right of first publication protection.

Alternatively of utilizing a water line image, a indiscriminately generated binary spot twine is used as the water line.

The public presentation of the DCT-based techniques is tested against a gray-scale Lena image of 512 ten 512 pels as the host. The imperceptibility standards of DCT-based are evaluated utilizing both nonsubjective and subjective measuring.

Objective measurings of host image are peak signal-to-noise ratio ( PSNR ) , and intend squared mistake ( MSE ) . Subjective measuring will be human ocular system but in this undertaking, we will merely utilize human review to measure quality of host image.

Undertaking Objective

The aim of this undertaking is:

To analyze different types of digital image watermarking techniques specifically DCT-based image watermarking technique.

To analyse the public presentation of DCT-based techniques.

To better the hardiness and imperceptibility demand of DCT-based techniques.

Information concealment is a general term embracing assorted sub subjects ( mention Figure 1 ) . The term “ concealment ” means to do information invincible or maintain that peculiar information undetectable in a host such as images, paperss. Two of the most of import bomber subjects are steganography and watermarking. ( Al-Mualla and Al-Ahmad, 2010 )

Figure 1: Categorizations of Information Hiding Techniques ( Adopted from Petitcolas, 1999 )

Cryptography refer to the processing of any information into an encrypted signifier to forestall it from being read by other people who should non read it or to be able to direct it through a channel without being eavesdropped. Although the message is encrypted, hacker or any user who are able to acquire the message while it is reassigning might be able to decode it. ( Chris, 2002 )

Steganography is different from cryptanalysis where cryptography hides the fact that a communicating is even occurred. Figure 2 shows the procedure flow to a steganographic system. In cryptography, original message which is defined as m is embedded into a harmless message degree Celsius which act as a cover-object. When the message m is embedded into screen object degree Celsiuss, a key K which acts as stego-key will be used.

Then the ensuing message will be embedded into the cover-object degree Celsius, which the consequence can be defined as stego-object s. A cover-object is merely used to bring forth stego-object and will be discarded after stego-object generated. In ideal status the stego-object should be near to the original screen object so that the presence of embedded information inside a message will be undetected.

Figure 2: Procedure of a Steganography system

Watermarking is a method of concealing information which is really similar with Steganography. Both methods embed information in a cover message with about no debasement of the cover-object. The difference of watermarking and cryptography is that watermarking adds in the demand of hardiness.

An ideal steganographic system is to implant a batch of information but with no seeable debasement to the screen object. However, an ideal watermarking system will implant a batch of information that could non be deleted or modified without doing the screen object to be wholly unserviceable. As a negative consequence of the demand of high hardiness, a digital host which has undergone a watermarking procedure will devour more infinite comparison to those which procedure by steganographic system.

Figure 3: Water line system

Requirement of digital watermarking is as following:

Imperceptibility

A watermarking system is considered to be high imperceptibility when the watermarked host image looks identical from the original even on the highest quality equipments. Due to the nature of an image where it is useless to anyone if it has been destroyed or extremely distorted by the screen image, a watermarking system which can bring forth higher unperceivable image is considered a better system.

Robustness

In the 2nd demand, a water line must be extremely robust. It should hold high resistant against deformation either during normal usage, or knowing or malicious onslaughts such as effort to disenable or take the water line. Unintentional onslaughts include common image transmutations such as cropping, contrast and brightness sweetening, etc.

Capacity

Another less of import demand for ideal watermarking system is capacity. A watermarking system is a system which will execute embedding of a utile sum of information onto an original image. The information embedded can run from a individual spot until multiple paragraph of text.

Statistical imperceptibility

Statistical imperceptibility may be the last possible demand of an ideal watermarking system. Watermarking algorithm must modify the spots of the screen in such a manner that the statistics of the image are non modified in any other manner which may bespeak that a water line is presence. This demand might non be of import in cryptography, but it do play an of import function as one of the demand in watermarking system.

After we perform watermarking procedure on an image, there are few methods that we can utilize to measure the peculiar watermarking techniques. Capacity can easy be evaluate by utilizing the figure of spots per screen size. Then statistical imperceptibility is calculated by correlativity between the original images and watermarked portion.

Besides that, to measure a watermarking technique, we can besides utilize Peak Signal-to-Noise Ratio ( PSNR ) . PSNR is the ration between the maximal power of a signal and the power of noise which affects the fidelity of its representation. PSNR can merely used as a unsmooth estimate of the quality of the water line.

Least Significant Bit ( LSB ) Alteration

Least Significant Bit Modification ( LSB ) is a method where water line will be embedded into the least important spot of the screen object ( C.Katzenbeisser, 1999 ) . LSB permutation is easy to implement but it besides brings a batch of drawbacks. LSB permutation may last image use such as cropping, but any add-on of noise or lossy compaction will get the better of the water line. A better onslaught is to put the LSB spots of each pel to one. Once the algorithm is detected, the embedded water line will be easy modified by other party.

Basic LSB permutation can be improved by utilizing a pseudo-random figure generator to find the location of the pels to be used for implanting based on a key. Securities of the water line will be improved because the chosen pels are indiscriminately generated. LSB permutation has been proved to be simple, but this method lacks the basic hardiness which is a of import demand for watermarking mechanism.

Frequency Domain Techniques

Discrete-Cosine-Transform ( DCT ) is the authoritative and most popular sphere for image processing. DCT allows images to be broken up into different frequence sets, doing it easier to implant watermarking information into the in-between frequence sets of the image.

Figure 4: Definition of DCT Regions

In low frequence part, information interior is really important and the energy of the part is really high. In contrast, high frequence part has insignificant information and low energy. In low frequence, if information is embedded indoors, the original image will acquire distorted easy.

If user embed watermark information into the high frequence part of a host image, when image undergo a compaction procedure, some of the content in the part will be discarded without important quality debasement ( Khayam, 2003 ) . So, the in-between frequence set is chosen so that the watermarking procedure would non impact the most of import ocular portion of host image and acquire the water line removed during compaction and noise onslaught.

In conformity to Lin and Chen ( 2010 ) , watermark information should be embedded into the lower frequence of the host image. They proposed to implant lesser water line spots into each block of host image so that the influence to the host image can be reduced. A pseudo-random sequence technique of watermark spot watercourse is used to bring forth the place of the place to implant water line. Then, binary spots of water line will be straight embedded into the least important spot of DCT coefficient. Duplicate transcripts of water line spot will be embedded to increase the hardiness of the image.

To pull out the water line, least important spot of DCT coefficient is replaced by watermark spot watercourse. So, the proposed method is self-extractable.

The proposed method has been tested against three different host image which have 512 ten 512 pels each and has been proved to be robust to assorted onslaught, i.e. , JPEG compaction and image use, where noise is added into an image.

Harmonizing to Tang and Aoki ( 1997 ) , they have proposed a DCT-based method to implant water line into a host image. In the method proposed, watermark spot twine will be embedded into the middle-band of the host image. The procedure of implanting start by naming an original host image D and watermark image W, the implanting procedure can be derived as: Dw = D ( + ) W, where ( + ) is the procedure of implanting and Dw is the watermarked images.

The demand of this procedure is both host images and watermarked image must be in the gray-scale manner. The watermark image size should be half of the original host image, so that when the water line image is embedded, it would non do the watermarked image become perceptible. Besides that, residuary image should be same size with that watermarked image so that water line embedding can be performed easy.

Watermark implanting procedure

First, the original host image is broken up into 8 ten 8 blocks and a DCT-based algorithm is used to take coefficient of the original host image in Zigzag order. Since the declaration of the watermarked image is half of the original host image, the implanting procedure is simple.

Then, a Differential Pulse Code Modulation ( DPCM ) is used to implant the water line into original host image. The information of the modified water line will be probably to concentrate on the country [ -1,1 ] of the image ( mention figure 5 ) , so after implanting procedure, the residuary form is obtained. Then for each pronounced pels of the permuted water line, DCT coefficient is modified harmonizing to the residuary mask, so that the matching residuary value is reversed. Last, DCT coefficients is inversed to obtain the embedded image.

Figure 5: The water line implanting frequence sphere

Watermark extraction method

The procedure of water line extraction requires both the original image and embedded image. DCT algorithm is used against the original image D and embedded image Dw to acquire the DCT coefficient. Then the middle-band DCT-coefficients is used to bring forth the residuary forms. Following, exclusive-or ( XOR ) operation is used on these two remainder forms to obtain a permuted information. Last, the permuted information is decoded to acquire the water line W.

 

Figure 7: Watermark extraction procedure

This method has been tested and is proved that it is valid in implanting a water line into an image, and it besides has been proved to be effectual in diminishing mosquito noise.

Hubali and Kanyakumari ( 2009 ) have proposed a water line coevals strategy based on histogram of the image and use to the original image on the transform DCT sphere.

The proposed algorithm is a frequence sphere watermarking strategy and operate by modifying the DCT coefficients. The water line is content based which means that the water line is generated from the host image itself and no external water line is used. Binary sequence of water line is generated with spacial sphere information but the embedding procedure is done in transform sphere.

Watermark spot twine is generated by calculate the mean of histogram of host image. Hm is denoted as the mean of histogram. Then the grey graduated table threshold of the image is calculated utilizing Otsu ‘s method. Gt will be denoted as the grey degree threshold of the image where the figure will be between 0 and 1. Next the average histogram Hm is downscaled by multiplying it with Gt. the value will be denoted as Th. The original image is so divided into block utilizing expression below ( 1 ) where N1 and N2 are the figure of rows and columns of image.

Then the mean of the blocks is calculated utilizing following expression ( 2 ) .

Th and Mb ( K ; 1 ) will so be compared and bring forth a twine of binary sequence. If Mb ( K ; 1 ) is larger than Th, computed matrix W will be 0, else if Mb ( K ; 1 ) is smaller than Th, W will be 1.

The water line W generated from the original image is embedded into the DCT sphere of the image itself. Original host image is divided into blocks of 8 ten 8 and DCT transmutation is apply to each block utilizing equation below.

The mid frequence coefficients which is generated by DCT procedure above will be altered utilizing expression below.

Watermark generated will be embedded into the in-between frequence coefficient so that these coefficients would non be altered significantly when the image undergo some image use procedure.

To pull out a water line, original image is required. First, the consequence of water line is undone on the leery image. Then the water line form is calculated from the image by reiterating the measure used to bring forth water line form.

This method has been tested utilizing matlab image processing tool chest and the end product shows that there is no great ocular deformation on the image after watermark implanting procedure. Besides that, it besides shows that this method is robust against assorted image processing i.e. , JPEG compaction.

In conformity to Taheriniaiand Jamzad, two degree DCT based digital watermarking techniques is proved to be extremely immune to compaction and linear noise, at the same clip continuing high PSNR for watermarked images.

“ In spread spectrum communications, a narrowband signal is transmitted over a much larger bandwidth such that, the signal energy nowadays in any individual frequence is unperceivable ” ( A.H. Taherinia, unknown ) .On the other manus, in dispersed spectrum watermarking strategies, host image is viewed as a communicating channel, and the water line is viewed as a signal to be transmitted. So, the water line is distributing many samples of the host signal by adding a low pseudo-random noise sequence to them. The embedded water line sequence is detected by correlating a specific pseudo-random noise sequence with the watermarked signal.

Blocked based DCT is applied on a host image which is of size 512 ten 512. Then, for each 8 x 8 image block, merely DC coefficient is selected out of the 64 DCT coefficients to implant the water line. Selected coefficient will so be mapped into a reduced image which called low-resolution estimate image ( LRAI ) ; which is formed from the DC coefficients of all transformed blocks of host image ( see Figure 8 ) .

Figure 8: Low-resolution Approximation Image

Then harmonizing to the value of water line spot which is traveling to be embedded in each block, a imposter random noise sequences added to the high frequences of DCT transform of each 8 x 8 block of LRAI utilizing formula shown below.

In above expression, FH denotes the high set frequences, K refer to the addition factor, ( x, y ) is the location of an 8 ten 8 DCT block of LRAI L, and Wi is the pseudo random noise sequence harmonizing to the value of I of watermark spot pattern. Two separate imposters random noise sequences are used to stand for the spot values of 0 and 1. Furthermore, by taking two pseudo random noise sequences to be every bit un-related as possible, the opportunity of false sensing can be reduced significantly.

Then, each block is inverse-transformed to give us watermarked LRAI. Last, DC coefficient of LRAI is replaced with their corresponding watermarked 1s, IDCT transform of each 8 x 8 block of LRAI is computed and watermarked image will be constructed.

The same pseudo-random noise generator algorithm is seeded with the same key to observe the embedded water line. Blocked based DCT is applied on the watermarked image. Then for each 8 x 8 image block, merely the DC coefficient is selected out of the 64 DCT coefficients. The coefficients will so be mapped into LRAI. The transformed LRAI of watermarked image is now constructed. Then the correlativity between both noise forms and blocks of transformed LRAI is computed. Model with higher correlativity will be chosen as extracted water line spot. Presence of water line is detected by comparing the mean correlativity coefficient of detected water line with a predefined threshold. If the mean correlativity is greater than the threshold defined, water line is concluded as presence.

The proposed method has been tested on several trials and has been proved to hold highest opposition to JPEG compaction compared to other good known techniques. Watermark is able to be extracted even the watermarked image has been compressed with choice factor of 1 % . Besides that, this method besides shows that it is robust with regard to additive Gaussian noise. Furthermore, in this method, water line can be extracted without even holding the implanting parametric quantities. The last advantage of this method is it does non do any important alterations on watermarked image.