A HYBRID ROBUST COLOR IMAGE WATERMARKING ALGORITHM USING CONTOURLET AND QUARTERNION HADAMARD TRANSFORM

A hybrid color image watermarking algorithm that use two different transforms such as Contourlet and Quaternion Hadamard transform is introduced in this work. The proposed algorithm also uses QR decomposition after the two transforms and the data is embedded on the decomposed Q-coefficients. At the embedding phase the band pass directional sub-band coefficients are estimated for R, G and B channels of the host image using Contourlet transform. The band pass directional sub-band coefficients of R, G and B channel forms a Quaternion image. The Quaternion image thus formed, is transformed using Quaternion Hadamard transform. The real coefficients obtained through Quaternion Hadamard transform is decomposed using QR decomposition after dividing the real coefficients into 4×4 sub blocks. One bit watermark information can be embedded on each sub-block. The watermark extraction phase is similar to watermark embedding phase where it extracts watermark information after transforming using contourlet, Quaternion Hadamard transform and decomposing it using QR. The performance of the proposed watermark embedding and extraction algorithm is evaluated using the parameters such as Peak Signal to Noise Ratio (PSNR) and Normalized cross correlation (Ncc). From the experimental results, it is verified that the proposed hybrid color image watermarking algorithm is robust against different types of attacks when compared to state-of-theart watermarking methods.