Improved watermark robustness via spectrum equalization

We describe and analyze an effective solution that can improve robustness for a large class of transform-domain watermarking algorithms. We present the concept of spectrum equalization for image watermarking and discuss its rationale and advantages. We demonstrate that if the watermark signal is embedded in a host image having a flat or equalized spectrum, the overall performance of the underlying algorithm can be significantly enhanced. The proposed idea is exemplified using a well-known spread-spectrum approach, and the experimental results demonstrate its effectiveness