Sunday, October 6, 2013

A watermarked image is exposed to any form of distortion like lossy image compression, re-sampling, cropping or some deliberate attacks on the embedded data. Depending on the need to extract the embedded data, the extraction process may or may not require knowledge of the original host image to estimate the hidden signature from the distorted image that is received. The extraction
where the presence of the original image is required is known as non-blind (nonoblivious, public) watermarking extraction and the one which do not use the original image for comparison is called blind (oblivious, private) watermarking extraction. In principle the difference between the extracted and the original signature is as low as possible. 

0 comments :

Post a Comment

|