Edge Sensitive Frequency Domain Watermarking

Abstract

Digital watermarking technique has been recognized as a helpful way in dealing with the copyright protection problem for high quality images and videos. Besides some practical challenges, it is adhered to provide security and currently no other technique is ready to substitute it. Watermarking an image is essentially the process of adhering an imperceptible mark within the original image in a manner that ensures that the watermark is not obtrusive and there is only indistinguishable difference between the original and the watermarked image. Many efficient techniques have been proposed but demanding security issues require more robust and complex techniques.

A lot of work has been done in the area of gray-scale image watermarking, but relatively less earlier work exists for colored image watermarking. We propose a colored image watermarking technique in which the location and the strength of the watermark to be embedded is varied in accordance with the underlying content of the image to be watermarked. The algorithm classifies the image blocks based on their sensitivity to noise. A highly textured block represents a highly important part of the image and hence would be put into a higher class in comparison to blocks with flat and featureless regions. Hence, for our perception, edge information of an image is the most important factor. The embedding process is automated and the bits are embedded in the frequency domain using discrete cosine transformation.

Keywords: Color Image, Digital Watermarking,Discrete Cosine Transformation (DCT), Edge detection, Intellectual Property Right (IPR), Imperceptibility