Digital Image Watermarking Using Discrete Cosine Transform Biology Essay

Abstract-The purpose of this undertaking is to implant a digital water line in an image to turn out its ownership, genuineness and to forestall illegal reproduction of right of first publication images. A Binary Image is embedded in the image utilizing DCT. The selected DCT coefficients of the image are replaced by the DCT coefficients of the binary image and so the IDCT of the image is taken to acquire the watermarked image. The watermarked image hence obtained is rather robust and unperceivable to assail.

Keywords-Discrete Cosine Transform ( DCT ) , Inverse Discrete Cosine Transform ( IDCT )


We Will Write a Custom Essay Specifically
For You For Only $13.90/page!

order now

Digital watermarking is a procedure of implanting unnoticeable Markss or labels into digital content. These embedded Markss are typically unperceivable ( unseeable ) that can subsequently be detected or extracted. There are many watermarking techniques available for images, sound and picture. A digital H2O grade has been defined by research workers as designation codification which carries certain information about the right of first publication proprietor, the Godhead of the work, the authorised consumer, and so on. It is for good embedded into digital informations for right of first publication protection and may be used for look intoing whether the information has been modified. The water line can merely be detected if the contents of the water line are known. This belongings of the water line determines the manner it is used in practical applications.

Water lines are loosely divided into two classs – Visible and Invisible. Visible watermarking is a seeable semitransparent image, which is overlaid on the image. It could be your name, right of first publication, remark, website reference, your logo, text or graphical objects. Image filters, day of the months, exposure inside informations and other EXIF information, which holds the rights to the primary image can besides be used for image watermarking. Watermarking processing allows the primary image to be viewed, but still marks it clearly as the belongings of the having organisation.

Invisible watermarking is the digital information that is added to audio, images or picture. But it can non be perceived as such. Because of its different applications, there are two really different types of unseeable watermarking.

Invisible watermarking, which is destroyed when the image is manipulated digitally in any manner may be utile in turn outing genuineness of an image. If the water line is still integral, so the image has non been “ doctored. ” If the water line has been destroyed, so the image has been tampered with. Such a engineering might be of import, for illustration, in acknowledging digital images as grounds in tribunal.

Invisible watermarking, which is really immune to destruction under any image use might be utile in verifying ownership of an image suspected of embezzlement. Digital sensing of the water line would bespeak the beginning of the image.

The chief part of this paper is to show a new watermarking strategy that is based on distinct cosine transform

( DCT ) and Joint Photographic Experts Group ( JPEG ) theoretical account in the characteristic sphere. It besides gives an up-to-date overview of the field of watermarking. Section 2 presents the chief applications and belongingss of watermarking. Section 3 elaborates on the techniques of digital watermarking. The proposed watermarking strategy is presented in Section 4. This is followed by the experiments and consequences in Section 5. Finally, the decisions are presented in Section 6.


One of the chief applications of watermarking is the copyright protection of digital media. In the past duplicating art work was rather complicated and required a high degree of expertness for the forgery to look like the original. However, in the digital universe this is non true. Now it is possible for about anyone to double or pull strings digital informations and non lose informations quality. One manner to undertake this job would be to implant a water line in the image which permits designation of the proprietor of the work. In the field of informations security, water lines may be used for enfranchisement, hallmark, and conditional entree. Certification is an of import issue for official paperss, such as individuality cards or passports. Another application is the hallmark of image content. The end of this application is to observe any changes and alterations in an image. The three images below illustrate this application. The image on the left shows an original exposure of a auto that has been protected with a watermarking engineering. In the Centre, the same image is shown but with a little alteration: the Numberss on the licence home base have been changed. The image on the right shows the exposure after running the digital water line sensing plan on the tampered exposure. The tampered countries are indicated in white. We can clearly see that the detected country corresponds to the alterations applied to the original exposure.

hypertext transfer protocol: //

Using digital water lines for unity confirmation: the protected image is the image ( a ) above ; a modified image is obtained by trading the Numberss 9 and 4 of the figure home base ( B ) ; digital watermarking engineering allows observing and high spots the modified countries, as shown on ( degree Celsius ) .

Fig 1 Embedding and Detecting systems of Digital Watermarking

There are three functional constituents that are required in order to implant a water line in an image. These are a water line bearer, a water line generator and a bearer qualifier. A water line bearer is a list of informations elements from the original image used for encoding the water line. The water line is a binary image. The bearer modifier adds the generated noise signals to the selected bearer. Implanting the water line and observing the water line are the operations in the watermarking of digital media, which enable the proprietor to be identified. The watermarking strategy can be represented symbolically by

Iw = E ( Io, W )

Where Io, W and Iw denote the original multimedia signal ( audio, image or picture ) , the water line incorporating the information that the proprietor wishes to implant, and the watermarked signal, severally. The implanting map E modifies Io harmonizing to W. Fig. 1 ( a ) shows a general watermarking strategy.

For watermark sensing, a observing map D is used. This operation is represented by

W ‘ = D ( R, Io )

Where R is the signal to be tested, whether it is watermarked or non and R could be a deformed version of Iw. The extracted water line sequence ( W ‘ ) is compared with W and a Yes/No determination is made. The determination is based on a correlativity step Z, as follows:

Z ( W ‘ , W ) = { 1, degree Celsius & gt ; = y

{ 0, otherwise

Where degree Celsius is the value of the correlativity and Y is a positive threshold. The sensing procedure is shown in Fig. 1 ( B ) . Watermarking techniques that are intended to be widely used must fulfill several demands. The type of application decides which watermarking technique to be used. However, three demands have been found to be common to most practical applications and the treatment below dressed ores on these.

A. Watermark Imperceptibility

The water line should be hidden in the media signal in such a manner that it can non be seen. However, watermark invisibleness can conflict with other demands such as hardiness. Sometimes it is necessary to work the features of the human ocular system ( HVS ) or the human auditory system ( HAS ) in the watermarking embedding procedure. The

Fig 2 Zigzag telling for the JPEG theoretical account

water line should besides be statistically unseeable. An unauthorised individual should non be able to observe the water line utilizing statistical methods.

B. Robustness

The water line should be noticeable even if knowing or unwilled onslaughts are made on the watermarked image. If this is the instance, so the water line is robust. To accomplish a high grade of water line hardiness, the water line must be placed in important parts of the media signal. In the instance of image watermarking, opposition to geometric uses, such as interlingual rendition, resizing, rotary motion, and cropping is still an unfastened issue.

C. Detecting the Watermark

The chance of neglecting to observe the embedded water line and observing a water line when, in fact, one does non be, must be really little even after the media has been subjected to onslaughts or signal deformation. As a consequence, sensing of the embedded water line proves the ownership of the media signal.

It must be understood that the above demands compete with each other. Different watermarking applications consequence in the corresponding design demands. In any instance, a watermarking technique should be widely accepted and used on a big, commercial graduated table, so that it might so stand up in a tribunal of jurisprudence.


Watermarking techniques are loosely divided into three classs – Spatial sphere, Frequency Domain and Wavelet Domain.

Spatial Sphere

To plan a digital water line in the spatial or temporal spheres, these attacks need to modify the least important spots ( LSB ) of the host informations. These lowest order spots are visually undistinguished, so the water line will be unseeable. After implanting it, the water line is recovered utilizing cognition of the PN sequence and water line location.

Frequency Domain

Frequency sphere watermarking technique is besides called transform sphere. Valuess of certain frequences are altered from their original. Typically, these frequence changes are done in the in-between frequence degrees, since alternations at the higher frequences are lost during compaction and alternations in the lower frequence will do a seeable alteration. The water line is applied to the whole image so as non to be removed during a cropping operation. However, there is a trade-off with the frequence sphere technique. Confirmation can be hard since this water line is applied randomly across the whole image.

Wavelet Domain

Another possible sphere for water line embedding is that of the ripple sphere. The DWT ( Discrete Wavelet Transform ) separates an image into a lower declaration estimate image ( LL ) every bit good as horizontal ( HL ) , perpendicular ( LH ) and diagonal ( HH ) item constituents. The procedure can so be repeated to computes multiple “ graduated table ” ripple decomposition, as shown in the figure.

Fig 3 2 Dimensional Discrete Wavelet Transform

One of the many advantages over the ripple transform is that that it is believed to more accurately exemplary facets of the HVS as compared to the FFT or DCT. This allows us to utilize higher energy water lines in parts that the HVS is known to be less sensitive to, such as the high declaration item sets { LH, HL, HH ) . Implanting water lines in these parts allow us to increase the hardiness of our water line, at small to no extra impact on image quality.


The authoritative and still most popular sphere for image processing is that of the Discrete-Cosine-Transform, or DCT. The DCT allows an image to be broken up into different frequence sets, doing it much easier to implant watermarking information into the in-between frequence sets of an image. The in-between frequence sets are chosen such that they have minimize they avoid the most ocular of import parts of the image ( low frequences ) without over-exposing themselves to removal through compaction and noise onslaughts ( high frequences ) .

The watermarking strategy we ‘re following is based on the JPEG compaction theoretical account. Initially the image is segmented and subdivided into blocks of size 8×8. After spliting the image, we take the DCT of each block. After transforming the image into frequence sphere, the pels are ordered utilizing the zigzag scan which is used in JPEG compaction.

Fig 4 Proposed watermarking strategy

The general equation for a 2D ( N by M image ) DCT is defined by the undermentioned equation:

egin { displaymath } F ( u, V ) = left ( frac { 2 } { N }
ight ) ^ { frac { 1 } { 2 } } left ( frac { … … } ( 2i+1 )
ight ] cosleft [ frac { pi.v } { 2.M } ( 2j+1 )
ight ] .f ( I, J ) end { displaymath }


egin { displaymath } Lambda ( xi ) = left { egin { array } { ll } frac { 1 } { sqrt { 2 } } & amp ; {
m for } xi = 0 1 & A ; {
m otherwise } end { array }
ight.end { displaymath }

Here, a water line consists of a binary image. The image consists of lone black and white pels with magnitude 0 and 1 severally.

Fig 5.Binary Image Used

Once the DCT of the blocks is computed, we embed the water line in the selected spots.

The water line is added by replacing the DCT coefficients of the selected block utilizing the equation

The hardiness is determined by the DCT coefficients of the image which are being replaced.

Greater the hardiness invariable, greater will be the hardiness of water line. But as the magnitude of the hardiness changeless additions, the visibleness of the water line increases excessively.

Now, we need to change by reversal the above process to acquire the watermarked image. So the modified

DCT coefficients are reinserted into the zigzag scan.

The blocks which are obtained as the end product of the zigzag scan are so subjected to Inverse Discrete Cosine Transform.

Following this the different blocks are merged to acquire the watermarked image.

In our proposed watermarking strategy, we do non H2O grade the full image but alternatively every image section of 8×8 pels. Within each section, merely the coefficients of the in-between frequences are replaced. The in-between frequence set, FM, is chosen as the implanting part as to supply extra opposition to lossy compaction techniques, while avoiding important alteration of the screen image.

Fig 5 Definition of DCT Regions


The proposed algorithm with an original image ( Fig 6.1 ) was executed many different figure of times with different implanting conditions, such as pel location where the water line has to be embedded.

When the algorithm is executed with water line embedded in the in-between frequence, the watermarked image obtained is similar to the original image, without any seeable alterations. The water line therefore embedded is robust, as it can non be detected and besides it is non lost in instance the image is compresses. ( Fig 6.2 )

When the algorithm is executed with water line embedded in the low frequence set, the watermarked image obtained is distorted with extremely seeable alterations in the original image. The water line therefore embedded is caught by the bare oculus. ( Fig 6.3 )

When the algorithm is executed with water line embedded in the high frequence set, the watermarked image does non exhibit any seeable alterations with regard to the original image. The hardiness of this watermarking is really weak, as the water line from the full image can be lost if the image is compressed, therefore rendering the image without any cogent evidence of its genuineness or ownership. ( Fig 6.4 )

Fig 6.1

Fig 6.2

Fig 6.3

Fig 6.4

The watermarked extracted in all the three instances, nevertheless is same and matches the embedded water line.

In instance of fiddling with the image, for illustration if the image is cropped out and so tested for water line, the water line obtained is defective and does non fit the embedded water line. ( Fig 7 )

Cropped Watermarked Image

Extracted Watermark