一种基于HVS的DCT域图像自适应数字水印新算法

An Image-adaptive Digital Watermarking Algorithm Based on HVS in DCT Domain

本文提出了一种基于人类视觉系统（HVS）的DCT域图像自适应数字水印算法，该算法可对原始宿主图像进行分块DC了变换，使用Robert算子对每一小块追行边缘检测提取，采用确定数值作为阀值对图像块进行分类，从而对嵌入的水印强度进行调整。添加水印时，对二值水印图像进行Arnold置乱、降维预处理，并将此序列嵌入到每块的DC了低频系数上。仿真实验结果表明，本文实现的水印具有不可见性，且对于常见的JPEG压缩、噪声、裁剪等攻击具有较好的鲁棒性。

This paper presents an adaptive digital image watermarking algorithm based on HVS in DCT domain. The original host image is divided into blocks and manipulated by DCT. The Robert operator is used to carry on the edge examination extraction for each block after piecemeal original image. The value determined is as the threshold value to classify the image blocks, and then to adjust the embedding intensity. Before being embedded the watermarking, the binary image is scrambled by Arnold, then the watermark image is pulled into a sequence, and the watermarking is hidden in the blocked DC components. The experimental results show that the watermarking is imperceptible and robust to common JPEG compression, noise, crop, etc.