Robust encoded spread spectrum image watermarking in contourlet domain

This paper presents a new robust watermarking system for digital images. The main challenge in every watermarking algorithm is to create a high capable, robust and imperceptible scheme. Unlike other watermarking methods, our system encodes the watermark using a convolutional encoder and embeds the encoded watermark in the transform domain of an image using a direct sequence spread spectrum technique to accomplish a high robustness. The embedded watermark bits are retrieved computing the correlation between the watermarked coefficients and the pseudorandom sequences used in the watermark embedding. In addition, the retrieved bits are decoded using the Viterbi algorithm to recover the hidden watermark. A low perceptibility of the watermark on the host image is preferred, therefore, some features of the human visual system are considered for achieving a high perceptual quality of the watermarked image. For assessing such quality, a structural similarity index measure is adopted. Our watermarking algorithm is robust to common image processes in the spatial domain.