自适应可逆数字图像水印算法研究

发布时间：2018-05-03 03:37

本文选题：可逆图像水印 + 预测误差　；参考：《江西理工大学》2017年硕士论文

【摘要】：随着科学技术的不断发展,数字多媒体信息普及化给人们的生活带来了极大的便利,但与此同时也带来了一些挑战。对于有些不法份子可以很容易就能获得这些数字多媒体信息进行窜改后发布,给信息产业带来了一系列的信息安全问题。数字水印是一种能够解决该领域出现的相关安全问题,通过把含有用信息的水印嵌入相关数字产品中,合法使用的用户可以提取出该有用的水印信息用来证明使用者的相关权益。传统数字水印技术提取嵌入的水印后会对原始多媒体信息会造成不可逆的失真,不适合用于对原始多媒体信息有较高质量要求的领域,为了克服传统数字水印缺点,可逆水印就随之出现,主要的思想是指在提取嵌入水印后,并且可以完全无失真地恢复出原始多媒体信息。本文主要研究的是把图像作为原始载体的可逆数字图像水印技术,当采用不同的嵌入水印容量时,改进相关算法使得嵌入水印后图像的质量能得到提升。论文主要内容和创新点为:(1)对可逆数字水印技术发展和应用做了系统的介绍,并对当前主流的可逆数字水印算法展开研究,分析了不同算法的基本原理和实现过程。利用MATLAB搭建实验平台,在实验平台上对算法进行仿真实验。(2)为了解决预测误差在预测的准确度和图像不可感知性上的不足,提出了一种基于像素排序预测的自适应选块可逆图像水印算法。算法首先通过对原始图像分块;然后利用多个像素与平均值之间的梯度计算图像子块噪声级,再结合像素排序预测算子得到预测误差直方图;最后根据得到的预测误差直方图,可以自适应地选择合适的子块进行嵌入水印过程。算法能提高预测器的预测精度并能够自适应选择子块嵌入水印,使得嵌入水印后图像失真会有所降低。(3)为了解决图像隐蔽性和水印嵌入容量之间相互矛盾的问题,提出一种基于预测失真自适应两遍嵌入可逆图像水印。算法首先对原始图像进行分块,利用方差衡量图像子块纹理性,根据纹理性设子块嵌入强度初始值;然后对子块图像预测失真再动态调整嵌入强度;最后通过最终确定的嵌入强度自适应地进行两遍嵌入水印。算法采用预测失真自适应两遍嵌入的方法,可以在高容量嵌入时有效减少图像失真,实现了高容量且有较高质量的可逆图像水印算法。
[Abstract]:With the development of science and technology, the popularization of digital multimedia information brings great convenience to people's life, but also brings some challenges. Some illegal elements can easily obtain these digital multimedia information for tampering and publishing, which has brought a series of information security problems to the information industry. Digital watermarking is a kind of security problem which can solve the related problems in this field by embedding the watermark with information into the relevant digital products. The legitimate user can extract the useful watermark information to prove the user's rights and interests. The traditional digital watermarking technology will cause irreversible distortion to the original multimedia information after extracting the embedded watermark, which is not suitable for the field of high quality of the original multimedia information, in order to overcome the shortcomings of the traditional digital watermarking. The main idea is that after extracting the embedded watermark, the original multimedia information can be recovered completely without distortion. This paper mainly studies the reversible digital image watermarking technology which takes the image as the original carrier. When different embedding watermark capacity is used, the improved correlation algorithm can improve the image quality after embedding the watermark. The main content and innovation of this paper are as follows: (1) the development and application of reversible digital watermarking technology are introduced systematically, and the current mainstream reversible digital watermarking algorithms are studied, and the basic principle and implementation process of different algorithms are analyzed. In order to solve the deficiency of prediction error in prediction accuracy and imperceptibility of image, we use MATLAB to build the experimental platform and simulate the algorithm on the experimental platform. An adaptive block selection reversible image watermarking algorithm based on pixel ranking prediction is proposed. The algorithm firstly divides the original image into blocks, then calculates the noise level of the image subblocks by using the gradient between multiple pixels and the average value, and then obtains the prediction error histogram by combining the pixel sort prediction operator; finally, according to the obtained prediction error histogram, The appropriate subblocks can be adaptively selected to embed the watermark. The algorithm can improve the prediction accuracy of the predictor and adaptively select sub-blocks to embed the watermark, which will reduce the image distortion after embedding the watermark.) in order to solve the problem of the contradiction between the image concealment and the watermark embedding capacity, the algorithm can reduce the image distortion after embedding the watermark. A reversible image watermarking based on predictive distortion adaptive two-pass embedding is proposed. The algorithm first divides the original image into blocks, then uses variance to measure the texture of the image subblock, and then sets the initial value of the embedding intensity according to the texture, and then dynamically adjusts the embedding intensity for the prediction distortion of the sub-block image. Finally, the watermark is adaptively embedded twice through the final embedding strength. The algorithm adopts predictive distortion adaptive two-pass embedding method, which can effectively reduce image distortion in high capacity embedding, and realize a reversible image watermarking algorithm with high capacity and high quality.
【学位授予单位】：江西理工大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP309.7

【参考文献】