A Blind Watermarking Algorithm Based on ICA

A blind watermarking algorithm based on independent component analysis (ICA) is proposed. According to the human visual system (HVS) in the domain of DWT and DCT, watermark is embedded to the host image adaptively. The advantages of the multi-resolution and multi-scale of the DWT, combining with the ability of energy compression and ridding correlation of DCT make sure the robustess of the algorithm. We utilize ICA to extract the watermark without the original image. Experiment results show that the proposed algorithm enhances the watermark robustness and invisibility.