Analysis of a wavelet-based watermarking algorithm

摘要

The analysis of a wavelet-based watermarking algorithm is presented, in which the discrete wavelet transform (DWT) is used for watermark embedding and detection. In the image compression field, the relationship between DWT coefficients and the human visual system (HVS) is researched intensively to get better compression performance. Recently, this knowledge has been applied to watermarking algorithms based in the DWT domain. The watermarking algorithm analyzed is adaptive to image characteristics, such as texture and luminance content, and it is realized using pixel by pixel masking in the DWT domain to embed the watermark with maximum energy. The detection process consists of cross-correlation between the marked image and pseudorandom sequences, and, in this process, the original image is not required. Computer simulation results show the robustness and perceptual invisibility of the algorithm.