Rubik’s cube watermark technology for grayscale images

A new chaotic mixing algorithm based on the concept of Rubik’s cube has been proposed in this paper for watermarking images. We use the discrete wavelet transform to decompose the primary image and perform the low frequency sub-band as robustness on embedding watermark. Before embedding the watermark, we obtain noise signals by our Rubik’s cube algorithm to protect original watermark. In the embedding phase, the eigenvalues of both reference watermark and primary images are extracted via SVD. We combine the eigenvalues which strongly represent the image and then perform inverse DCT and inverse DWT to result a watermarked image.