摘要
对基于混沌序列定位的跳频盲水印算法进行了研究.采用跳频技术,利用描述图像自身特征的分形维数对图像块进行分类.根据图像块的种类决定水印嵌入的频带,将图像块的分类信息和一维混沌映射的初始条件作为密钥,由产生的混沌序列确定水印在频带中嵌入的位置,再通过系数关系比较的方法嵌入水印信息.实验结果表明,该算法在满足水印透明性的同时,水印信息在噪声干扰、图像处理、图像压缩及各种人为恶意攻击条件下具有很好的鲁棒性.采用跳频技术和混沌序列定位增强了该算法的安全性.
A frequency-hopping blind watermarking algorithm based on chaotic sequence locating is studied. The fractal dimension (FD) to describe the image's character was used to classify the image' s blocks. Different frequency-hopping bands to embed watermark were chosen according to the very FD classifications. The exact position to embed watermarks in the chosen band were determined by the classification and the chaotic sequence, as the output of one-dimensional chaotic map. The classification and the initial values of the chaotic sequence were saved as the secret key. The watermark information was embedded by the relationship of the coefficients. With favorable transparency, simulation showed that the algorithm had a good robustness against noise, image processing, image compression, and malicious attacks. Adoption of chaotic sequences enhanced the security capability.
出处
《北京理工大学学报》
EI
CAS
CSCD
北大核心
2007年第8期709-712,717,共5页
Transactions of Beijing Institute of Technology
基金
国家部委基金资助项目(10405020103)
关键词
盲水印
跳频
混沌序列
离散余弦变换
blind watermarking
frequency-hopping
chaotic sequence
discrete cosine transform