一种改进的奇异值奇偶区间量化图像隐藏算法  

An Improved Image Hiding Algorithm of Singular Value Parity Interval Quantization

作　　者：冯丹青 陈亮 FENG Dan-qing;CHEN Liang(The Army Engineering University of PLA,Nanjing 210042,China)

机构地区：[1]陆军工程大学研究生院

出　　处：《科学技术与工程》2019年第26期283-287,共5页Science Technology and Engineering

基　　金：国家自然科学基金(61072042)资助

摘　　要：针对空域中载密图像不可感知性差、鲁棒性劣的问题,提出一种改进的奇异值奇偶区间量化的图像信息隐藏方法。该算法采用图像子块矩阵的奇异值分解技术,根据奇异值的最大值划分其奇偶区间,按照信息隐写规则跨区移动,改进了传统方法的量化修改过程,达成隐藏数据的嵌入和盲提取。实验结果表明:算法图像视觉质量优,JEPG压缩方面鲁棒性良好,能抵御一定噪声和压缩,同时提高了不可感知性,比传统方法高了20%以上。In order to solve the weakness of the robustness and imperceptibility in image spatial domain,an improved image information hiding algorithm of singular value parity interval quantization is put forward. This algorithm used singular value decomposition technique for image subblock,which marked out its parity interval in accordance with the maximum singular value. It improved the quantitative modification process of the traditional method by using the information implicit rule,and achieved the embedding and extraction blindly of the secret information. The results of experimentation demonstrate that the algorithm has great image quality,and it can sustain certain noise and have good robustness under the JPEG compression. More than up to 20% improvement of imperceptibility,respectively,is shown over other traditional algorithm.

关 键 词：奇异值分解 信息隐藏 奇偶区间量化 

分 类 号：TP309.7[自动化与计算机技术—计算机系统结构]