Preprocessing method for improving the robustness of blind watermarking technology

摘要

With the development of information technology, many image, video and audio resources are rapidly spreading on the Internet, which raises great privacy concerns. As a result, copyright protection has attracted more and more attention. Blind watermarking which is commonly known as a useful technique for copyright protection aims to embed the information into the host imperceptibly and extract the information from the host after unknown attacks during the transmission. Therefore, the embedded watermark’s invisibility and robustness to attacks are two important metrics to evaluate a watermarking algorithm’s performance. This paper studies the influence of the watermark properties and the host image on the robustness and invisibility of the watermarking method and brings up general rule-of-thumb to improve the performance. Based on the above, a novel watermark preprocessing method which can enhance blind watermarking’s robustness is proposed and applied to the robust embedding algorithm based on DWT-DCT (Discrete wavelet transform – discrete cosine transform). Experiments show that the proposed watermark preprocessing method can significantly enhance the watermarking algorithms’ robustness and generalization ability.