摘要
对宿主图像频域进行二维DCT变换,采用灰度图像JND门限值和多数字基底对宿主图像的DCT频域构造了选块分组技术,对中频分量的DCT交流系数分组嵌入水印比特,对高低频分量的DCT系数随机嵌入水印比特,嵌入了多个灰度数字盲水印。该算法预估了水印嵌入量,提取水印无需宿主图像和水印图像,仿真实验表明能有效抵抗滤波、加噪、剪切、JPEG有损压缩,具有较好的鲁棒性。
This approach is chiefly concerning a series of techniques : the substitution of 2D DCT in frequency domain of the owner image, the use of selection-and-group tactics conducted by means of JND threshold of gray-image and the multiple-based work underlying DCT frequency domain of owner image, the classification of embedded watermarking with AC of intermediate frequency, the performance at random on AC of high-low frequency that embedded, and the realization of digital blind watermarking of multiple gray watermarking. This algorithm evaluates the capacity of the digital watermarking and extracts the digital watermarking without the original watermark and the owner image. The experiment proves that such arithmetic can resist the procedure, such as filtering, adding noise, clipping,JPEG compression, and is also characteristic of good robustness.
出处
《计算机应用与软件》
CSCD
北大核心
2007年第8期186-188,共3页
Computer Applications and Software