Robust Watermarking in Wavelet Domain Using Rank Order and Genetic Algorithm for Image Authorization

Abstract

A robust watermarking (ROW) technique is proposed to protect image authorization. The ROW technique explores features which are the distances among rank orders of the coefficients in subbands of a wavelet-domain image. The features are then utilized to divide wavelet blocks of the image into three categories. A set of thresholds, which corresponds with the three categories, is exploited to develop the embedding and the extraction algorithms. The ROW technique retrieves watermarks without original images. The GROW technique, which employs a genetic algorithm to optimize the ROW technique, to enhance the robustness of the ROW technique. Experimental results illustrate that the performances of the ROW and the GROW techniques are acceptable.