Image Watermarking
1. Introduction:
With the increasing importance and widespread
distribution of digital media and imagery, the protection of the
intellectual property rights of the owner for thei media has become
increasingly significant. One of the types of media is digital
imagery, which can be copied and widely distributed wihtout any
significant loss of quality.
Image Watermarking is the process of embedding into an image
some information which helps establish the ownership of the
image.
2. Fundamental definitions :
- Visible Watermark : An image is visibly embedded into
another. This is really the closest to the classical approach used to put
watermark on important paper documents.
- Invisible Watermark : Some information which may be
another image or some string of numbers or letters are embedded into the
image in a manner which does not alter the perceived quality of the image i.e. the
watermarking process does not 'visibly' alter the image.
- Frequency domain watermarking : In this method the
image is broken down into its frequency components using some image transform,
and the watermark in embedded therein.
- Spatial domain watermarking : In this method the pixel
intensities are manipulated directly to embed the watermark.
- Blind & Non-blind Watermarking algorithms : If
the watermarking algorithm needs the original image to help it extract
the watermark from the watermarked image then it is a non-blind watermarking algorithm, else
it is a blind one.
3. Desirable qualities :
All embedding schemes must meet several requirements
regarding the quality of the watermark. These may be enumerated as :
- The watermarked image should retain as closely
as possible the quality of the original image.
- The watermark recovery process must be able to
prove to a high degree of certainity that any image is actually a
watermarked image, i.e. it should be resistant to showing 'false
positives' when searching for watermarks in the image. Also if the
image is watermarked then the extracted wateramrk should match only
with the original watermark and not with any other randomly generated
watermark.
- The watermark should be robust. The
watermark should be able to retain a significant part of the watermark
when a deliberate or non-deliberate attempt is made through various
image processing techniques to remove or corrupt the watermark.
4. My Implementation :
Secure Spread Spectrum Watermarking of Images
References :
- Ingemar J. Cox, Joe Kilian, Tom Leighton, and Talal G. Shamoon.
Secure spread spectrum watermarking for multimedia.
In Proceedings of the IEEE ICIP '97, volume 6, pages 1673 - 1687,
Santa Barbara, California, USA, 1997.
- S. Craver
Can Invisible Watermarks Resolve Rightful Ownership?
IBM Research Report, 1996, p. 5.
- Peter Meerwald's watermarking sources at www.cosy.sbg.ac.at/~pmeerw/
Page last updated on 28 January, 2004.
|
AT cse.iitd.ac.in
|
© Parag Chaudhuri , 2009
|
|