基于视觉特性的半色调图像水印算法

Watermarking algorithm for halftone images based on human visual system

针对半色调(二值)图像提出一种新的数字水印算法。利用人眼的视觉掩盖特性和亮度适应特性,将水印隐藏在半色调图的高亮度纹理区域;采用加密和置乱技术提高二值水印图像的安全性,定义位卷积并设计r-位滤波器来指导水印的嵌入和提取。仿真结果表明:在嵌入相同数据量时,该算法较常规方法有更好的水印不可见性,当图像遭受裁剪、加噪、JPEG压缩、随机涂改攻击时,仍能正确提取其中嵌入的水印。

A digital watermarking algorithm for halftone （binary） images was proposed. Using visual masking and luminance adaptation characteristics in human eyes, the watermark was embedded in high-light textured areas, and watermark security was improved with encrypt and scrambling techniques. Bit convolution was defined and an r-bit filter was constructed in order to help watermark embedding and extraction. Experiments show that the proposed method has a better watermark imperceptibility compared to traditional halftone watermarking algorithms. Furthermore, when watermarked halftone is attacked by cropping, noise, JPEG compression, or random marking, watermarking can still be correctly extracted.