**1.5 Watermarking requirements**

520 Advances in Wavelet Theory and Their Applications in Engineering, Physics and Technology

LL1, LH1, HL1 and HH1. The process can then be repeated to computes multiple "scale"

One of the many advantages over the wavelet transform is that it is believed to more accurately model aspects of the HVS as compared to the FFT or DCT. This allows us to use higher energy watermarks in regions that the HVS is known to be less sensitive, such as the middle frequency bands (LH, HL) and high resolution band (HH). But watermark embedded in high resolution band can be easily be distorted by geometric transformation,

Embedding watermarks in middle frequency regions allow us to increase the robustness of

**LH1 HH1** 

Although the main motivation behind the digital watermarking is the copyright protection, its applications are not that restricted. There is a wide application area of digital watermarking, including broadcast monitoring, fingerprinting, authentication and covet

By embedding watermarks into commercial advertisements, the advertisements can be monitored whether the advertisements are broadcasted at the correct instants by means of an automated system [7, 8]. The system receives the broadcast and searches these watermarks identifying where and when the advertisement is broadcasted. The same process can also be used for video and sound clips. Musicians and actors may request to

Fingerprinting is a novel approach to trace the source of illegal copies [7, 8]. The owner of the digital data may embed different watermarks in the copies of digital content customized for each recipient. In this manner, the owner can identify the customer by extracting the watermark in the case the data is supplied to third parties. The digital watermarking can also be used for authentication [7, 8]. The authentication is the detection of whether the content of the digital content has changed. As a solution, a fragile watermark embedded to the digital content indicates whether the data has been altered. If any tampering has occurred in the content, the same change will also occur on the watermark. It can also

Covert communication is another possible application of digital watermarking [7,8]. The watermark, secret message, can be embedded imperceptibly to the digital image or video to communicate information from the sender to the intended receiver while maintaining low

ensure that they receive accurate royalties for broadcasts of their performances.

provide information about the part of the content that has been altered.

probability of intercept by other unintended receivers.

**HL1** 

wavelet decomposition, as in the 2 scale wavelet transform shown in Fig. 6.

compression and various signal processing operations.

**LL2 HL2**

**LH2 HH2**

Fig. 6. Scale 2 Dimensional DWT

**1.4 Watermarking applications** 

communication [7, 8, 9, 10].

our watermark, at little to no additional impact on image quality [6].

The efficiency of a digital watermarking process is evaluated according to the properties of perceptual transparency, robustness, computational cost, bit rate of data embedding process, false positive rate, recovery of data with or without access to the original signal, the speed of embedding and retrieval process, the ability of the embedding and retrieval module to integrate into standard encoding and decoding process etc. 7, 8, 9, 12, 13].

Depending on the application, the properties, which are used mainly in the evaluation process, varies.

The main requirements for copyright protection are imperceptibility and robustness to intended or non-intended any signal operations and capacity.

The owner of the original data wants to prove his/her ownership in case the original data is copied, edited and used without permission of the owner. In the watermarking research world, this problem has been analyzed in a more detailed manner [13, 14, 15, 16, 17, 18].

The imperceptibility refers to the perceptual similarity between the original and watermarked data. The owner of the original data mostly does not tolerate any kind of degradations in his/her original data. Therefore, the original and watermarked data should be perceptually the same. Robustness to a signal processing operation refers to the ability to detect the watermark, after the watermarked data has passed through that signal processing operation.

The robustness of a watermarking scheme can vary from one operation to another. Although it is possible for a watermarking scheme to be robust to any signal compression operations, it may not be robust to geometric distortions such as cropping, rotation, translation etc. The signal processing operations, for which the watermarking scheme should be robust, changes from application to application as well. While, for the broadcast monitoring application, only the robustness to the transmission of the data in a channel is sufficient, this is not the case for copyright protection application of digital watermarking. For such a case, it is totally unknown through which signal processing operations the watermarked data will pass. Hence, the watermarking scheme should be robust to any possible signal processing operations, as long as the quality of the watermarked data preserved.

The capacity requirement of the watermarking scheme refers to be able to verify and distinguish between different watermarks with a low probability of error as the number of differently watermarked versions of an image increases [17]. While the robustness of the watermarking method increases, the capacity also increases where the imperceptibility decreases. There is a trade off between these requirements and this trade off should be taken into account while the watermarking method is being proposed.

A DFT-DWT Domain Invisible Blind Watermarking

I(m,n)=I(m,n) + K\*seq\_zero(m,n)

Where 1≤ i ≤ M, 1≤ j ≤ N, and 1≤ m,n ≤ 4 Here I denotes to 4x4 DFT blocks.

middle frequency components (LH2, HL2).

Detect the watermark according the following rule:

If corr\_zero(i) > corr\_one(i) then watermark\_detected(i)=0;

 Reshape the recovered message. Display recovered message.

as the watermark image, as shown in Fig. 8.

copyright protection scheme for a specific attack.

watermark\_detected(i)=1

If template(m,n)==1 then

if template(m,n)==1 then I(m,n)=I(m,n)+K\*seq\_one(m,n)

Display watermarked image.

these blocks.

sequence.

Else

End

(4).

**3. Experimental results** 

End Else

End End

Techniques for Copyright Protection of Digital Images 523

 Apply IFFT to each image block and use the result as the middle frequency component of DWT to recover the component which has been embedded watermarking messages.

Implement Wavelet transform on Host image using wavelet function 'Haar' and Extract

Divide the HL2, LH2 components in several blocks of size 4x4 and DFT is applied to

 Use same highly uncorrelated PN sequences (key1) and the template matrix of 4x4 (key2) to select elements that are embedded watermarking message to make up

 Calculate the correlation separately between sequence and seq\_zero and between sequence and seq\_one. The result is stored in corr\_zero and corr\_one respectively.

Calculate the quality of recovered image by using PSNR function according to the

Calculate the Accuracy rate of recovered image by using AR function as per the (6).

In this section, we show some experimental results to demonstrate the effectiveness and success of our digital watermarking techniques. The standard 512 × 512 grayscale image "pepper" is used as host image, as shown in Fig. 7. The 32 × 32-pixels binary image is used

We applied the peak-signal to noise rate (PSNR) given in (4) to measure the image quality of an attacked image and accuracy rate AR given in (5) to evaluate the robustness of a

Replace the component of the host image by the watermarked component.

The watermark extraction steps of this technique are as follows:
