兰州大学机构库 >物理科学与技术学院
信息隐藏算法研究
Alternative TitleStudy on Information Hiding Algorithms
郁军
Thesis Advisor赵书城 ; 张晓萍
2004-05-12
Degree Grantor兰州大学
Place of Conferral兰州
Degree Name硕士
Keyword信息隐藏 数字水印 Patchwork JND DCT嵌入 色彩空间 图像增强 图像锐化
Abstract本文在分析了当前信息隐藏及相关技术的基础上,提出了改进的Patchwork算法和基于JND技术的Patchwork与DCT域嵌入相结合的两种针对静止数字图像的信息隐藏算法。改进的Patchwork算法结合对HVS特性、嵌入数据频域特性的分析,并在 “灰度值修改最小”准则下给出了最优替换表,该方案在保持视觉不可见性、统计分析不可见性及算法简单性的同时,提供了常规空域算法所不具有的对JPEG压缩的抵抗能力,同时在大量实验的基础上,对置位方式、色彩空间选择等问题进行了探讨。基于JND技术的Patchwork与DCT域嵌入相结合的算法结合了一般DCT域信息隐藏算法和空域Patchwork算法的优点,解决了常规DCT域嵌入方案不能对抗剪切攻击的问题。文中澄清了许多相关文献中,在采用JND技术对RGB彩色图像实施嵌入时,所选的中介色彩空间与使用的静态JND矩阵之间存在的匹配问题。本文提出的两种信息隐藏算法在隐藏信息提取端均不需要原始图像的参与,并能判明任意一幅图像是否存在使用各自算法嵌入的数据。这两种方案已被转化为实用软件。 本文还首次提出可以在信息提取端采用图像增强技术来提高信息隐藏系统健壮性的观点,并提供了初步的实验验证。
Other AbstractTwo information hiding schemes for still images are proposed in this paper, “An improved Patchwork scheme” and “A JND-based method using both DCT and Patchwork”. Some HVS results and discussions on the relationship between the Patch shape and the data energy distribution pattern on the DCT spectrum as well as a look-up table which is built up under the principle of “Minimal pixel value modification” are included in the first scheme. While retaining visual and statistical imperceptibility as well as comparatively light computational load, it has the advantage in resisting the JPEG compression over most of the currently spatial embedding schemes. Discussions on color space selection and bit resetting skills are also presented on the ground of experimental results. Having combined the Patchwork and the DCT embedding methods, the second scheme settled the cropping attack which has defeated most of the currently DCT embedding schemes. A clarification on the usually confused matching relation between the color space and the static JND matrix for DCT basis function is also given. Both of the proposed schemes do not require the original host image during data extraction. Besides that, they could decide if an arbitrary image contains any information that is hidden by their own embedding algorithm. The two schemes have already been converted to application softwares. For the first time, image enhancement technique is proposed as pretreatment before data extraction so as to improve the robustness of the information hiding system. Some rudimentary supporting experimental results are given in the end.
URL查看原文
Language中文
Document Type学位论文
Identifierhttps://ir.lzu.edu.cn/handle/262010/229160
Collection物理科学与技术学院
Recommended Citation
GB/T 7714
郁军. 信息隐藏算法研究[D]. 兰州. 兰州大学,2004.
Files in This Item:
There are no files associated with this item.
Related Services
Recommend this item
Bookmark
Usage statistics
Export to Endnote
Altmetrics Score
Google Scholar
Similar articles in Google Scholar
[郁军]'s Articles
Baidu academic
Similar articles in Baidu academic
[郁军]'s Articles
Bing Scholar
Similar articles in Bing Scholar
[郁军]'s Articles
Terms of Use
No data!
Social Bookmark/Share
No comment.
Items in the repository are protected by copyright, with all rights reserved, unless otherwise indicated.