Toward a new wavelet-based watermarking approach for color images using bio-inspired optimization principles

This paper presents a novel watermarking approach for copyright protection of color images based on the wavelet transformation. We consider the problem of logo watermarking and employ the genetic algorithm optimization principles to obtain performance improvement with respect to the existing algorithms. In the proposed method, the strength of the embedded watermark is controlled locally and according to the visual properties of the host signal. These parameters are varied to find the most suitable ones for images with different characteristics. The experimental results show that the proposed algorithm yields a watermark which is invisible to human eyes and robust to a wide variety of common attacks.