Improved spread spectrum image watermarking in contourlet domain

A novel watermarking algorithm operating within the contourlet domain is presented. We have used an improved spread spectrum technique for additive watermark embedding. As usual, the watermark embedded is retrieved by computing the correlation between the watermarked coefficients and the pseudorandom sequences. The detection threshold is optimized via Neyman-Pearson criterion, hence the watermark detection can be performed without knowing the value of the watermark strength, which can thus be accurately adapted to each image. In order to prove the validity of our approach, we present experimental results, our algorithm is robust to common signal processing operations, which could be both intentional or unintentional. Proposed algorithm also shows improvement in visual and invisibility of the hidden data.