Data hidding in color images using perceptual models

Abstract

One of the problems arising from the use of digital media is the ease of identical copies of digital images or audio files, allowing manipulation and unauthorized use. Copyright is an effective tool for preserving intellectual property of those documents but authors and publishers need effective techniques that prevent from copyright modification, due to the straightforward access to multimedia applications and the wider use of digital publications through the www. These techniques are generally called watermarking and allow the introduction of side information (i.e. author identification, copyrights, dates, etc.). This work concentrates on the problem embedding and optimum blind detection of data in color images through the use of spread spectrum techniques, both in space (Direct Sequence Spread Spectrum or DSSS) and frequency (Frequency Hopping). It is applied to RGB and opponent color component representations. Perceptive information is considered in both color systems. Some tests are performed in order to ensure imperceptibility and to assess detection quality of the optimum color detectors.

Peer Reviewed

Postprint (published version)