Abstract: A blind image watermarking algorithm in contourlet domain based on improved modified spread spectrum (IMSS) technique is proposed. The original binary watermark is firstly scrambled by two-dimensional Arnold transform and mapped from unipolarity to bipolarity, and then embedded into the richest textured region of the subband, which has the maximum energy among the second biggest directional subbands of the host image decomposed by the contourlet transform. Meanwhile, the IMSS technique is adopted in the process of watermark embedding. The watermark embedded is retrieved by computing the correlation between the watermarked coefficients and the pseudo random sequences. The optimal threshold is calculated to determine the existence of the watermark in the correlation detector. Experimental results demonstrate that the proposed algorithm not only has good imperceptibility but also has higher robustness against common image processing and attacks than conventional SS (spread spectrum) watermarking; furthermore, it overcomes the major disadvantage of conventional SS watermarking which completely fails if the tested image is pure.

Key words: computer application, blind watermarking, Contourlet transform, improved modified spread spectrum technique