Ternary Picture as Watermark for Audio Files

摘要

Watermark inserted in an audio file may be damaged after attacking this file. The goal of the paper is the development of a class of watermarks which can be recognized by human being even if the watermark saved only a part of the original information. A picture is leveraged as a watermark, and effective ternary coding of such a picture is suggested. The insertion of the watermark is based on the modulation of the container. The resistance of the watermark to various attacks is investigated. The original container is used while watermark extracted.