摘要
针对目前信息隐藏算法抵抗隐写分析能力弱的问题,提出一种基于尺度不变(BRISK)局部特征的零低频信息隐藏算法。首先,对载体图像进行一阶CL多小波变换,在低频LL 2中提取BRISK特征点生成图像特征矩阵;其次,利用zig-zag和Logistic混沌置乱对秘密信息进行去相关性处理;再次,将图像特征与加密信息通过对比特征值形成关联序列;最后,将关联序列嵌入到高频HL 2、HH 2的低3位。算法将高能量区域的特征矩阵与两次加密信息所构建的关联信息隐藏于高频区域,有利于算法的鲁棒性和抗分析性。在高阶统计量对200幅图片的分析测试下,最大检出率低于7.516%,表明所提算法具有良好的抗分析性。
Aiming at inferior anti-analysis of current information hiding algorithm,this paper proposed a zero-low-frequency information hiding algorithm based on local BRISK(binary robust invariant scalable keypoints)feature.First,it carried out first-order CL multi-wavelet transform for carrier image.Then it extracted BRISK feature points in the low-frequency LL 2 to generate an image feature matrix.Second,it used zig-zag scrambling and logistic chaos scrambling for the secret image to decorrelate.Then,it associated the image feature with the encrypted information to form an association sequence by comparing feature values.Last,it would embed association sequence into lower three bits of high-frequency HL 2 and HH 2.The association information constructed by eigenmatrix of high energy region and encrypted information of two times was hidden in the high frequency region,which was beneficial to the robustness and anti-analysis of algorithm.Under the analysis of high-order statistics on 200 pictures,the maximum detection rate was less than 7.516%,which indicates that the proposed algorithm has good anti-analysis ability.
作者
任帅
贺媛
柳雨农
徐振超
张弢
王震
慕德俊
Ren Shuai;He Yuan;Liu Yunong;Xu Zhenchao;Zhang Tao;Wang Zhen;Mu Dejun(School of Information Engineering,Chang’an University,Xi’an 710064,China;School of Electronic&Control Engineering,Chang’an University,Xi’an 710064,China;College of Automation,Northwestern Polytechnical University,Xi’an 710072,China)
出处
《计算机应用研究》
CSCD
北大核心
2019年第11期3365-3368,共4页
Application Research of Computers
基金
国家自然科学基金资助项目(61702050,61402052)
国家级大学生创新创业训练计划资助项目(201610710036)
关键词
零低频信息隐藏
BRISK特征
CL多小波变换
抗分析性
zero-low-frequency information hiding
BRISK feature
CL multi-wavelet transform
anti-analysis ability
