May 10, 2011
School of Engineering, Amritapuri
Computers and communication systems today handle vast amounts of digital data including multimedia, images and videos.
The storage and secure transfer of this data is no trivial matter. Innovative techniques for data authentication, security and compression are employed in modern communication systems towards this end.
Watermarking, or the embedding of a signature image within an image for identification purposes, is becoming widely used for authentication in image processing.
Now Amrita research scholars have discovered a novel method for digital image watermarking and digital image compression that has several advantages over methods currently in use.
Assistant Professor Rajesh Kannan Megalingam led this research initiative with his B.Tech. (ECE) students viz. Mithun Muralidharan Nair, Rahul Srikumar, Venkat Krishnan Balasubramanian and Vineeth Sarma.
The work of this Amrita team, already presented and published as research papers*, will now be available as part of a MATLAB book, ISBN # 979-953-307-124-3, published by InTech Open Access Publishers, University Campus, STeP Ri, Croatia.
MATLAB provides a powerful and user-friendly development environment for engineering and scientific applications. This fifth edition focuses on how new MATLAB tools can allow engineers to engage in comprehensive testing, modeling, simulation and analysis.
The Amrita chapter in this book is titled Novel Variance Based Spatial Domain Watermarking and Comparison with DIMA and DCT Based Watermarking Counterparts.
“In this chapter, we have discussed a novel method for digital watermarking in spatial domain called VB2 (Variance Based Variable Block) Algorithm,” shared Rajesh.
“Additionally, we have also introduced a compression technique which utilizes a variance-based DIMA (Diversified Intensity Matrix Algorithm) algorithm to facilitate multimedia data storage and management,” he added.
The new algorithm introduced by Amrita was shown to be more robust and perform better than existing algorithms. Its advantage lay in its ability to economize on the usage of memory when many images were to be watermarked and transmitted.
The research scholars watermarked various images with the Amrita logo using their DIMA algorithm. The image was divided into several small blocks. The variance and pixel value of each block and the number of bits required to display the image were determined.
If the variance was high, the values inside the block were quite different, the blocks needed to be all sent together. If the variance was low, most values in the block were the same, allowing the user to compress the image and send minimal number of essential details.
The receiver could recover the image without much distortion.
* Performance Comparison of Novel, Robust Spatial Domain Digital Image Watermarking with the Conventional Frequency Domain Watermarking Techniques, ICSAP Conference, Bangalore, February 2010.
A Comparative Study on Performance of Novel, Robust Spatial Domain Digital, Image Watermarking with DCT based Watermarking, International Journal on Scientific Computing, August 2010.