提出了一种新的自适应盲检测水印算法。首先混沌序列经过细胞自动机和平滑处理,然后生成水印比特,最后根据噪声敏感度系数自适应地嵌入原始图像的中频系数中,并且提供了一套盲检测方法。实验证明,算法在保证图像具有良好感知透明性的同时,还对一些常见攻击方式(如高斯噪声、中值滤波、JPEG压缩和剪切等)具有很好的鲁棒性。
By integrating existing technology used in chaotic sequence and noise sensitivity, a novel adaptive blind watermark algorithm was proposed. In order to enhance robustness of the embedded watermarking, a chaotic sequence was changed into watermark bits by cellular automata processing and smoothing. The embedding procedure operates adaptively in the domain of intermediate coefficients, according to the noise sensitivity. Robustness of it on selected attacks like Gaussian noise, median filtering, JPEG compression and cropping was tested. In addition, the perceptual transparency had been improved as well. Furthermore, the watermark can be detected without the original image.