A Robust Watermarking Method for Color Images using Naive-Bayes Classifier

F. Yaghmaie and M. Jamzad (Iran)

Keywords

Watermark, Color image, rotation, Com pression, Naive-Bayes classifier.

Abstract

By watermarking an image, we hide a pattern in it in such a way that the pattern is not visible but can be extracted using a decomposition algorithm and a key in the receiver side. The watermark pattern can be a character string or any small image (pattern). One of the main applications of watermarking is its application in proving digital image ownership in widely used Internet. In this paper, we present a watermarking method for color images which is robust with respect to usual attacks such as noise addition, smoothing, compression and also rotation. The validity of correctness of retrieved water mark is based on the result of a Naive-Bayes classifier. This classier was trained on a set of 2000 images with low, medium and high level of details. Our algorithm was tested on a large number of images. It proved to be robust and in all cases the watermark pattern could correctly be retrieved after the watermarked images have gone through the above mentioned attacks.

Important Links:



Go Back