Robust Wavelet based Watermarking Using Particle Swarm Optimization

Abstract

From the 13th century we are using watermarking techniques and first time watermarking used by paper industry because the benefit of watermarking in this industry to unique identify of papers. Thus we can use this technique for authentication and copyright purpose. Watermarking is a pattern which is used to insert in a entity. When we insert any watermark then it will not change any functionality or structure of an entity. We can say digital entities like audio, video, computer programs, software, hardware like chips. Main purpose of watermark is to provide authenticity and illegal distribution of work attacks.

Introduction

I. INTRODUCTION

From the 13th century we are using watermarking techniques and first time watermarking used by paper industry because the benefit of watermarking in this industry to unique identify of papers. Thus we can use this technique for authentication and copyright purpose. Watermarking is a pattern which is used to insert in a entity. When we insert any watermark then it will not change any functionality or structure of an entity. We can say digital entities like audio, video, computer programs, software, hardware like chips. Main purpose of watermark is to provide authenticity and illegal distribution of work attacks. The purpose is to detect a strong method that find such type of illegal work. Digital watermarking can be defined as the way to store data we can say watermark in to digital multimedia files like images such that we can extract this information later. We can explain the process of watermark with the help of fig.1.3. Here a signal is embedded with the mark with function value ..The watermark is robust against various attacks.

A. Objective And Scope

1. To create imperceptibility and high PSNR in HSV color space in frequency domain.
2. To improve the robustness and remove the false positveness using DWT and SVD methods.
3. To optimize the results by using embedding and extraction process which will improve the quality and robustness of watermark.

In this paper three technique are used for watermarking. All these techniques belongs to transform domain’. The watermark information embedded in frequency domain coefficients of HSV color space representation of the images. This helps to preserve the chromatically information resulting in good imperceptibility and high PSNR value. The frequency domain transform uses the discrete wavelet domain(DWT) method and singular wavelet transform(SWT) method. (SVD)singular value decomposition gives high robustness against compression and noice.The main problem with SVD(singular value decomposition) is false positive problem and the methods which we are using here to remove the limitations of SVD based on watermarking techniques. Here we are using the embedding method and extraction method. PSNR value used as the criteria for optimization. The new robust algorithm have improved the quality and robustness of watermark. This is the main purpose in this paper. And it also improve the authentication. In paper here use new color space, new algorithm’ and new transform methods.

II. DESCRIPTION OF WORK

In Spatial domain watermarking method: This is one of the good method for watermarking. And it takes less time and computational complexity. This technique is not robust against outside attacks.

1. LSB: (Least Significant bit insertion technique)- In this technique random pixels are selected from the host image and then the watermark is embedded on the least significant position.
2. Predictive Coding Scheme: This technique is more robust than previous technique. In this first set of pixels need to be embedded with watermark which we taken and the difference between adjacent pixels are used to replace alternate pixels in the image.
3. Correlation Based Technique: In this pseudo random noise is added and during decoding a correlation between them is found. Here is the equation:

Iwm2(m,n) = Im(m,n) + K*R(m,n) Iwm is the watermark image. R(m,n) is the pseudo random noise is added to host image Im(m,n).K is   gain factor , if we increase the k value then quality of the image decrease but robustness of image increases. In this method robustness also depends upon the gain factor.

4. Patchwork Technique: In this method the image is divide in to two parts. Some operations are applied in these parts in opposite direction. If one part is decreased by y factor then another part is also increased by same amount . This technique is more robust against many type of attacks.

III. TRANSFORM DOMAIN WATERMARKING TECHNIQUE

Spatial domain watermarking techniques are very easy to understand but they are less robust and we can’t do any further processing in this technique. Other than transform domain watermarking technique gives more robustness. In this technique host image first change in to transform domain and then watermark is embedded.

A. Discrete Cosine Transform Method

By using this method we can divide image in to low, medium and high frequency coefficients. Fig 1.1 shows the coefficients after applying discrete cosine transform of an image.

1. Initialization: Generate a population of particles with random positions and velocities in the search space. Each particle's position represents a potential solution, and its velocity determines how it moves through the space.
2. Evaluation: Evaluate the fitness or objective function value for each particle's position. This function quantifies how good the solution is for the optimization problem.
3. Update Particle's Best Position (PBest): Each particle records its best position (solution) found so far based on its fitness value.
4. Update Global Best (GBest): Determine the best-performing particle among all particles in the population (global best) based on their fitness values.
5. Update Velocity and Position: Update the velocity and position of each particle using the following equations:

New Velocity(i) = (Inertia Weight * Current Velocity) + (Cognitive Coefficient * Random Number * (PBest - Current Position)) + (Social Coefficient * Random Number * (GBest - Current Position))

New Position(i) = Current Position + New Velocity

The inertia weight controls the impact of the particle's current velocity. The cognitive coefficient and social coefficient are constants that regulate the influence of personal experience (PBest) and the experience of the best particle in the population (GBest), respectively. Random numbers are introduced to add exploration to the algorithm.

6. Termination: The algorithm continues iterating through steps 2 to 5 until a termination condition is met. Common termination conditions include reaching a maximum number of iterations or achieving a satisfactory solution.

Particle Swarm Optimization is a simple and efficient optimization algorithm that can be applied to a wide range of problems, such as function optimization, engineering design, data clustering, and neural network training, among others. Its ability to explore and exploit the search space efficiently makes it popular in many optimization tasks.

A. Optimized Watermark Embedding Algorithm using Particle Swarm Optimization

1. Step 1 - Read a image as color image and decompose it into different channels that are R1 G1 B1.
2. Step 2 -Read watermark image and decompose it into R1 G1 B1 channels.
3. Step 3 - Convert host image into different blocks and find their entropy value. Here size of each block is (8*8).
4. Step 4 - Sort each of the block according to entropy value in descending order.
5. Step 5 – Then apply 2D DCT on image and select DCT coefficients to embed the watermarked blocks.
6. Step 6 –Then embed the calculated watermark blocks into chosen main image’s blocks using scale factor alpha1, which is

B. Optimized Watermark Extraction Algorithm using Particle Swarm Optimization

1. Step 1 - Read the host image and convert it into R1G1B1 channels .
2. Step 2 - Read the watermarked image and find its R1G1B1 channels used to calculate using Particle swarm optimization technique to optimize the final result so that balance between robustness and imperceptibility can be finally reached. Rather than using one scaling factors called alpha1 multiple values have been used for embedding then the best alpha1 value returned from the Particle swarm optimization algorithm has been used for embed the watermark. Embedding is performed using following equation (1)

Watermarked_block1= host_block1+            alpha     *watermark_block

………………….(1)

3. Step 7 – Then apply inverse transform to get final watermark image.

4. Step 8 - Combine all channels of watermark image to get final watermarked image as output.

5. Step 3 -Read the alpha1 value calculated by Particle swarm optimization algorithm.

6. Step 4 – Change the host image into (8*8) block and then calculate their entropy value.

7. Step 5-Then  arrange  all           the          blocks according to the values in descending order .

8. Step 6 – Change the signed image into (8*8) sized blocks and then apply sorting according to entropy values in descending order.

9. Step 7 – Then finally extract the watermark on the sorted block datasets in first (1028) elements using equation (2) and the alpha1 value calculated by the Particle swarm optimization algorithm. Watermarked_block_extractedi=(signed_blo ck- host_block)/alpha1…          (2)

10. Step 7 – Then finally extract the watermark on the sorted block datasets in first (1028) elements using equation (2) and the alpha1

11. Step 8 - Combine all the R1G1B1 planes to get color watermarked image.

References

[1] Provos, Niels, and Peter Honeyman, \"Hide and seek: An introduction to steganography.\" Security & Privacy, IEEE, Vol. 1, No. 3, pp. 32-44, 2003. [2] A. Cheddad, J. Condell, K. Curran and P. McKevitt, “Digital Image Steganography: Survey and Analyses of Current Methods”. Signal Processing, Elsevier, Vol. 90, No. 3, pp. 727- 752, 2010. [3] P. Singh and R S Chadha, “A Survey of Digital Watermarking Techniques, Applications and Attacks”. International Journal of Engineering and Innovative Technology (IJEIT), Vol. 2, No. 9, pp. 165- 175, 2013. [4] Amit Kumar Singh, Mayank Dave, and Anand Mohan, “Multilevel Encrypted Text Watermarking on Medical Images Using Spread-Spectrum in DWT Domain”. Wireless Personal Communications, pp.1- 18, 2015. [5] W. Zhicheng, L. Hao, Dai Jufeng and W. Sashuang, “Image watermarking based on Genetic algorithm”. IEEE International Conference on Multimedia and Expo, ICME, pp. 1117-1120, 2006. [6] V. Aslantas, A. L. Dogan and Serkan Ozturk, “DWT-SVD based image watermarking using Particle Swarm Optimizer”. IEEE International Conference on Multimedia and Expo, ICME, pp. 241-244, 2008. [7] V. Aslantas, S. Ozer and S. Ozturk, “Improving the performance of DCT-based fragile watermarking using intelligent optimization algorithms”. Optics Communications, Elsevier, Vol. 282, No. 14, pp. 2806–2817, 2009. [8] N. Nilchi Ahmad R and T. Ayoub, “A new robust digital image watermarking technique based on the discrete cosine transformation and neural network”. IEEE International Symposium on Biometrics and Security Technologies, ISBAST, pp. 1- [9] C-T Yen and Y-J Huang, “Frequency domain digital watermark recognition using image code sequences with a back-propagation neural network”. Multimedia Tools and Applications, Springer, pp. 1-11, 2015. [10] T. S Tagare and S. Kuri, “Digital Image Watermarking -An Overview”. International Journal of Computer Science Information Engineering Technologies, Vol. 1, No. 3, pp. 2277-4408. 2015. [11] R. Singh, “Digital Image Watermarking: An Overview”. International Journal of Research (IJR), Vol. 2, Issue 05, pp. 1087- 1094, 2015. [12] V. M. Potdar, S. Han and E. Chang,\"A survey of digital image watermarking techniques.\" 3rd IEEE International Conference on Industrial Informatics (INDIN), pp. 709-716, 2005. [13] P. Parashar and R. K. Singh, “A Survey: Digital Image Watermarking Techniques”. International Journal of Signal Processing, Image Processing and Pattern Recognition Vol. 7, No. 6, pp. 111-124, 2014. [14] N. Ahmidi. and Reza Safabakhsh, “A Novel DCT-based Approach for Secure Color Image Watermarking”. International Conference on Information Technology: Coding and Computing, IEEE, pp. 709-713, 2004. [15] Han Baoru, and Jingbing Li. \"Medical Image Watermarking in Sub-block Three- dimensional Discrete Cosine Transform Domain.\" International Journal Bioautomation, Vol. 20, No. 1, pp. 69-78 2016. [16] Kumar Arvind, Pragya Agarwal, and Ankur Choudhary. \"A Digital Image Watermarking Technique Using Cascading of DCT and Biorthogonal Wavelet Transform.\" International Conference on Recent Cognizance in Wireless Communication & Image Processing, Springer, pp. 21-29, 2016. [17] M. Barni and F Bartolini, “Improved Wavelet-Based Watermarking Through Pixel-Wise Masking”. IEEE Transactions On Image Processing, Vol. 10, No. 5, pp. 783- 791, 2001.

Copyright

Copyright © 2023 Anu Jayalwal, Dr. Latika Singh. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.