Fast digital watermarking algorithm based on cellular automata and multi-level texture synthesis

In this paper, we present a novel technique for quickly generating large-scale digital watermarks. Generally, digital watermark can be embedded in any copyright image whose size is not larger than it. But nowadays, more and more high-resolution pictures call for larger and larger watermark templates, while most traditional watermarking algorithms work slowly on templates of large scale. To solve this problem, we firstly produce a pseudo-random number sequence, using the Logistic Function and then convert it to a binary 2D image. Afterwards, cellular automata are used to turn this image into an elementary watermark template. We regard this chaotic featured watermark as an input texture, and apply a multi-level texture synthesis method to it. In this way, large-scale watermarks can be obtained at acceptable cost of time.