提出了一种基于Baker变换加密技术的多小波域自适应图像水印算法,充分利用了多小波变换、混沌序列的优点和人眼的视觉特性,使嵌入的水印具有很好的鲁棒性和很强的隐蔽性。实验证明该算法具有很好的抗剪切、滤波、旋转和JPEG压缩等优良性能。较好地解决了水印透明性与鲁棒性之间的矛盾。
In this paper, an adaptive watermarking algorithm based on the discrete multiwavelet transform and the human visual perception characteristics is proposed.In this algorithm,the Bakers' transformation is used to encrypt watermark image and utilizes the visual masking in multiwavelet domain to embed the encrypted watermark in different image parts with different intensity.Experimental results show that the algorithm can improve the robust of watermarking and keep it invisible and stand against several attacks,such as filtering,scaling,cropping image,rotation and JPEG compression,also can be helpful to protect the copyright of digital images.