摘要
一直以来不可见性和鲁棒性是数字水印算法两个相互制衡的因素,结合人类视觉系统提出了一种建立在离散小波变换基础之上的图像数字水印算法,该算法依据人类视觉系统的光学特性在图像的蓝色通道实施2层离散小波变换,利用小波变换后的高频系数之和的近零性这一统计特性量化嵌入二值水印图像,水印实现了盲提取。同时水印在嵌入前实施了Arnold变换,在嵌入过程中利用密钥控制水印嵌入位置的随机选择,从两方面保证了水印算法的安全性。实验表明算法对压缩、添加噪声等攻击有抵抗力,同时确保算法的不可见性得以很好地维持,而且在针对图像的局部篡改可以定位篡改,能够实现版权保护和内容认证的功能。
For a long time,invisibility and robustness are two factors that restrict each other.This paper presents an image watermarking algorithm based on discrete wavelet transform combining human visual system.According to the optical characteristics of human visual system,the algorithm implements two-layer DWT in the blue channel of the image,the binary watermark image is embeded by using the statistical property of the sum of high frequency coefficients of the wavelet transform.The blind extraction of watermark is realized.At the same time,Arnold transform is implemented before the watermark is embedded.In the process of embedding,the key is used to control the random selection of the watermark embedding location,which ensures the security of the watermark algorithm.Experiments show that the algorithm is resistant to compression,noise and other attacks,while ensuring the invisibility of the algorithm can be well maintained,and the tampering can be located in the local tampering of the image,which can achieve the functions of copyright protection and content authentication.
作者
杨永锋
武文才
杨羽
YANG Yong-feng;WU Wen-cai;YANG Yu(College of Information Engineering,Longdong University,Qingyang 745000,Gansu)
出处
《陇东学院学报》
2022年第5期58-62,共5页
Journal of Longdong University
