Watermarking is a technique of hiding a message about a work of media within that work itself in· the purpose of protecting the digital information against illegal duplication and manipulation. The objectives of this study are to analyze the robustness and distortion performance of watermarking system and to explore watermarking schemes which balance the robustness-distortion tradeoff optimally.
In this thesis, We present a detector algorithm to adaptively extract spread spectrum watermark by filtering the watermarked images with Wiener filter. Two optimization algorithms for quantization watermarking are proposed. First one optimizes uniform quantization based look-up table embedding which minimizes watermarking distortion. Secondly, we analyze the robustness-distortion tradeoff and formulate the robustness-distortion tradeoff into a Lagrangian function. Hence optimal quantizers for watermarking subject to given robustness or fidelity constraint are achieved.
Adaptive multimedia data hiding and watermarking