Robust Image Watermarking Technique to Increase Security and Capacity of Watermark Data

Jaison John, Jibin Thomas Varghese, Jobin Raj, Smina P Mathew Department of Electronics and Communication Engineering TKM Institute of Technology, Karuvelil, Kollam, Kerala,India

Abstract
The recent progress in the digital multimedia technologies has offered many facilities in the transmission, reproduction and manipulation of data. However, this advancement has also brought the challenge such as copyright protection for content providers. Digital watermarking is one of the proposed solutions for copyright protection of multimedia data. This technique is better than Digital Signatures and other methods because it does not increase overhead. Digital watermarking is not a very old field. Most of research is going on in this field. Researchers try to invent techniques that increase the security, capacity, and imperceptibility of watermarked images. Robust image watermarking technique is a new image watermarking technique to increase security and capacity of watermark data. In this project a new image watermarking technique is proposed that can embed more number of watermark bits in the cover image without affecting the imperceptibility and increases the security of watermarks. To increase the embedding capacity the concept of watermark in watermark is used. Means it embeds an extra watermark into the main watermark and then embeds the main watermark into the cover image. To increase security we embed encrypted watermarks in the image. This provides an additional level of security for watermarks.

Digital watermark is a pattern of bits inserted into a digital image, audio, video or text file that identifies the files copyright information (author, rights, etc.). The name comes from the faintly visible watermarks imprinted on stationary that identify the manufacturer of the stationery. The purpose of digital watermarks is to provide copyright protection for intellectual property thats in digital format [2]. Digital watermarking is not a very old field. Most of research is going on in this field. Researchers try to invent techniques that increase the security, capacity, and imperceptibility of watermarked images. In this paper we proposed a new image watermarking technique that increases the security and capacity of watermark data. Here we used the concept of watermark nesting. Only one technique (discussed in [3]) exists that embed watermark in watermark, which is based on vector quantization. So, till date no technique exists for watermark nesting in frequency or wavelet domain. The difference from [3] is that our proposed method encrypts both the watermarks before embedding. This provides an additional level of security for watermarks.

2. Purpose of the project

Today there are many watermarking techniques which embed one watermark in cover image. Only one technique is there that embed watermark in watermark, which is based on vector quantization. So, till date no technique exists for watermark nesting in frequency or wavelet domain. The difference from [5] is that our proposed method encrypts both the watermarks before embedding. This provides an additional level of security for watermarks. So the problem we are trying to solve is not previously solved. Watermarking is not a fully mature technology; lot of research is going on in this field, spatially to increase security and capacity of watermark data. Most of researchers try to increase the watermark capacity by compromising image quality, because there is a tradeoff among data rate, security and imperceptibility (figure 1). But with our scheme we will be able to embed more number of watermark bits without affecting the imperceptibility of the cover image.

Key words: discrete wavelet transforms, encryption, spatial domain. 1. Introduction

The recent progress in the digital multimedia technologies has offered many facilities in the transmission, reproduction and manipulation of data. However, this advancement has also brought the challenge such as copyright protection for content providers. Digital watermarking is one of the proposed solutions for copyright protection of multimedia data. This technique is better than Digital Signatures and other methods because it does not increase overhead.

Imperceptibility

3.1 Watermark Embedding Algorithm

Input Watermark1 a binary image act as a watermark that we embed in the main watermark. Watermark2 a binary image act as main watermark. Cover Image gray scale image to be watermarked. E1 key used for encrypting the Watermark1 E2 key used to encrypt watermarked watermark. Algorithm 1. We take Watermark1 and encrypt it by performing XOR operation with the key E1.The output of this step is called Encrypted1. Apply procedure proposed in section 5.1 to embed Encrypted1 in the second binary watermark image (Watermark2). Let output image is First watermarked. Again encrypt First watermarked using XOR with key E2 to give the output image Encrypted2. Apply procedure given in section 6.1 to embed Encrypted2 in the gray-scale Cover Image. Output image is final watermarked image (Final watermarked).

Security
imperceptibility

Capacity

Fig.1 Tradeoff among watermark data rate, security and

Also our watermarking technique will use cryptography. So, it will provide an additional level of security. For instance if watermarking key is hacked still the attacker will not be able to identify the watermark because it is encrypted. So, it is worth to solve this problem, because by solving it we will get a watermarking technique that will increase the security of watermarks and will be capable of embedding more number of watermark bits in the cover image.

2.

3. 4.

3. Proposed Watermarking Method

To increase watermark data capacity the concept of nesting is used. Means we embed one watermark in other. And to increase security of watermark cryptography is used. It is a blind watermarking method. Means original image is not required at the time of watermark recovery. For embedding 1st watermark in second we use spatial domain technique, because it is less time consuming as compare to wavelet or frequency domain techniques. Spatial domain techniques are less robust. But robustness is much more important issue to be considered for second watermark, because both unintentional and malicious attacks alter the final watermarked image, which directly affect the second watermark. So for embedding second watermark we used technique based on DWT, which is very robust against attacks [1]. Before embedding watermarks at both levels we encrypt them with XOR operation. XOR operation has one important property if we XOR the data twice with same key we get original data again. This property of XOR is used for encryption and decryption. For encryption we XOR the binary image with some key. For decryption, we have to XOR the encrypted image with the same key. It gives the original image.

Output Final watermarked the final watermarked image

3.2 Block diagram- Embedding

Watermark2
Encrypted1

Watermark1

Encryption E1

Watermark Embedding (According to 4.1)

First watermarked Watermark Embedding (According to 5.1)

Encrypted2

Encryption
E2

Final watermarked

Cover Image

Fig.2 block diagram of embedding procedure

3.3 Watermark extraction algorithm

Inputs Final watermarked ' it is the received watermarked image. E2 key used to decrypt Recovered watermark from cover Image. E1 key used for decrypting Recovered Watermark from main watermark. Algorithm 1) Apply procedure proposed in section 6.2 to extract encrypted watermark2 from the Final watermarked, say the recovered image is Encrypted2. 2) Decrypt Encrypted2 using XOR with key E2.output of this step is called Recovered2. 3) Apply procedure proposed in section 5.2 to extract encrypted watermark1 from the Recovered2, recovered image is called Encrypted1. 4) Decrypt Encrypted1 using XOR, with key E1. Output of this step is called Recovered1 Output Recovered2 main watermark recovered from the received watermarked image. Recovered1 watermark recovered from the main watermark.

4. Spatial Domain Embedding Algorithm

This algorithm is based on dividing the cover image into blocks, each of size 55. In each block, 4 bits from the secret message can be embedded with changing at most two bits in the block. [4].

4.1 Embedding Algorithm

Input: Cover image and secret message Output: Stego-image. Step1: Divide the cover image into blocks (F) each of size 55. Step2: For each block, proceed as follows: 1. for every row in the first four rows of the block, exclusive-or all the bits of that row to get r r r r
1 2 3 4.

2. for every column in the first four columns of the block, exclusive-or all the bits of that column to get cccc
1 2 3 4.

3. Exclusive-or the results in 1 and 2 to get s s s s where s =r

1 1

1 2 3 4

c , s =r
1 2

c , and so on.
2

4. Compare the result obtained from 3 with the four embedded bits b b b b .
1 2 3 4

3.4 Block diagram- Extraction

Encrypted1 Final watermarked

E1

Watermark Extraction (According to 5.2)

Decryption

Recovered2

Encrypted2 Recovered 1

Decryption
E2

Watermark Extraction (According to 4.2)

Fig. 3 block diagram of extraction procedure

If there is no difference, no change of bits in F is needed, otherwise, consider the following cases: If the difference in one bit , the bit or should be changed. y Else if difference in two bits and then the bit or should be changed. Else if difference in three bits  and , then the bits (( or ) and  or )) Or (( or ) and ( or )) Or (( or ) and or )) should be changed Else difference in four bits  ,  and , then the bits (( or )and( or )) Or (( or ) and ( or )) Or (( or ) and ( or )) should be changed The selection of the bit depends on the number of adjacent bits with the same value. The bit that has the least number of adjacent bits is selected. This is because it has a minimum effect on the cover image when it is changed. 4.2 Extracting Algorithm Input: Stego-image Output: Secret message

The algorithm used for extracting is similar to that used for embedding. It performs the following steps to give the embedded data. Step1: Divide the cover image into blocks (F) each of size 55. Step2: For each block, proceed as follows: 1. for every row in the first four rows of the block, exclusive-or all the bits of that row to get r r r r
1 2 3 4.

2. 3.

4.

Store the dwt coefficients (A1, H1, V1 and D1) in a matrix, say Ey. Calculate N from the equation N= (Ey-Y). /C where N is the embedded data Compute for each value of N, Set N=1 for N>0.5 Set N=0 for N<0.5

2. for every column in the first four columns of the block, exclusive-or all the bits of that column to get cccc
1 2 3 4.

Output: watermark embedded in the cover image

3. Exclusive-or the results in 1 and 2 to get the embedded bits s s s s where s =r c , s =r c ,

1 2 3 4 1 1 1 2 2 2

6. Experimental Results
In our project, we have been successful in embedding one watermark over another as well as this nested watermark in the cover image and also successfully extracted back the embedded data. First we see the effect of embedding watermark in another watermark to obtain nested watermark. Figure 3.1shows the effect of embedding a watermark in Lenas image in spatial domain. Figure 3.2 shows the embedding of nested watermark in the cover image in discrete wavelet domain.

and so on.

5. DWT Algorithm
Most of the researchers focus on embedding watermark in wavelet domain because watermarks in this domain are very robust.

5.1 Embedding Algorithm

Input: cover image, watermark 1. Perform DWT decomposition of the Cover Image at level one. And store Approximation, horizontal, vertical and diagonal coefficients in A1, H1, V1, and D1 respectively. 2. Initialize the weight of Watermarking, say C. 3. Store the dwt coefficients (A1, H1, V1 and D1) in a matrix, say y. 4. Calculate a new matrix Y, using the above matrix in the equation, Y=y + C. *N where N is the watermark needed to be embedded. 5. Obtain the new wavelet coefficients from the new matrix. 6. Perform Inverse Discrete Wavelet Transform (IDWT) using the new wavelet coefficients, to create the watermarked image. Output Watermarked image it is a watermarked with binary image. gray-scale image (a) (b) Fig. 5 (a) original image (b) watermarked image Here, in these figures, we used Lena image as both the cover image and the second watermark, to show the difference between the spatial domain embedding and the discrete wavelet domain embedding. But in actual cases, the three images are different which may be having some relationship with each other.

(a) (b) Fig.4 (a) watermark (b) nested watermark

5.2 Extracting Algorithm

Input- watermarked image 1. Perform DWT decomposition of the received Image at level one. And store Approximation, horizontal, vertical and diagonal coefficients in A1, H1, V1, and D1 respectively.

6.1 Capacity Increase Results

In our Watermarking technique the embedding capacity is more than normal watermarking because here watermark nesting is used [2].

6.2 Security Increase Results

In our watermarking method we used encryption. So in any case if watermarking key is leaked and attacker extracts the watermark, still he will not be able to read the watermark because it is encrypted. In our watermarking method user need two keys for watermark extraction. If any of keys is invalid then user will not be able to extract both watermarks correctly. Either one or both the watermarks will be incorrect. It depends upon which key is invalid.

more number of bits in the cover image as compare to without watermark nesting. y Due to nesting feature we can embed some metadata about watermark also. y Because our technique uses encryption, so it increases the security of watermarks. y It is a blind watermarking technique. So, original image is not required at the time of watermark recovery. y Because we embed final watermark in DWT domain, so this technique is robust against many attacks.

References
[1]. Harpuneet Kaur, Robust Image Watermarking Technique to Increase Security and Capacity of Watermark Data [2]. Alper Koz, Digital Watermarking Based on Human Visual System, The Graduate School of Natural and Applied Sciences, The Middle East Technical University, pp 2 8, Sep 2002. [3]. Amara Graps, An Introduction to Wavelets, in IEEE Computer Science and Engineering, vol. 2, num. 2, pp. 50-59, 1995. [4]. Ahmed Al-Jaber and Khair Eddin Sabri,Data Hiding In A Binary Image [5]. Feng-Hsing Wang, Lakhmi C. Jain, Jeng-Shyang Pan, Hiding Watermark in Watermark, in IEEE International Symposium in Circuits and Systems (ISCAS), Vol. 4, pp. 4018 4021, May 2005.

7. Conclusion
This paper presents a blind watermarking technique that uses watermark nesting (at level 2) and encryption. Nesting means it embeds an extra watermark into the main watermark and then embeds the main watermark into the cover image. For encryption we used XOR operation. For embedding watermarked watermark in Cover Image we used DWT based technique. Proposed watermarking technique has following advantages: y By using watermark nesting we can embed