Abstract:
Digital watermarking involves embedding a watermark value within a digital object, such as image, audio, video, text and software, to prove authenticity in case of intellectual property infringement. Headed to this direction, in this paper we survey our previous algorithmic techniques for software and image watermarking and present a new developing idea based on them for audio watermarking. Our watermarking techniques take as an input a watermark that is an integer $w$ which can be efficiently encoded as a self-inverting permutation $pi^*$. We demonstrate multiple representations of self-inverting permutations, namely reducible permutation graphs, two-dimensional and one-dimensional matrices. We propose efficient algorithmic techniques for watermarking software, image and audio that exploit self-inverting permutation representations in order to embed the watermark $w$ by making imperceptible modifications and producing equivalent watermarked objects of high fidelity.
