基于HVS的小波域信息隐藏算法的研究-智能与分布计算实验室

基于HVS的小波域信息隐藏算法的研究

姓名	胡珊珊
论文答辩日期	2007.05.31
论文提交日期	2007.06.07
论文级别	硕士
中文题名	基于HVS的小波域信息隐藏算法的研究
英文题名	Research on HVS-based Steganography Schemes in wavelet domain
导师1	卢正鼎
导师2
中文关键词	信息隐藏;容量;小波变换;置乱处理;人类视觉系统
英文关键词	steganograph;capacity;wavelet-transform;scrambing image by Arnold transform;HVS
中文文摘	在进行数字作品版权保护的某些时候或者是电子政务、军事及法律金融等领域内，有时需要嵌入很大容量的隐密信息。特别的是信息隐藏技术将是未来信息化战争对抗的焦点，作为未来情报战的重要组成部分，大容量的信息信息隐藏技术必将对战争的进程和胜败产生重大影响。从实际应用需求出发，研究如何在以图像为载体的通信系统中嵌入尽可能多的隐密信息，也就是一种大容量的图像信息隐藏算法。围绕提高信息隐藏算法的容量，对现有典型算法进行比较，总结各自的特点，分析影响信息隐藏容量的因素以及提高信息隐藏容量的方法。进而有侧重的对人类视觉特征和数字图像质量评价、数字图像置乱处理以及数字图像的小波变换进行相关的研究和分析。基于人类视觉特征和小波域置乱图像能量分布特点，提出一种基于HVS的小波域信息隐藏算法。该算法分析比较载体图像进行置乱处理前后，在小波变换域中其系数能量分布的特点。通过对载体图像进行置乱处理，改变其空域中象素之间的相关性，达到均匀化其系数能量分布的目的，从而兼顾信息隐藏系统不可见性和鲁棒性的特点，选择载体图像置乱处理后的小波低频的边缘区域作为嵌入区域。结合人眼视觉特征，采用非线性的嵌入规则，根据噪声可见函数NVF值自适应的调整嵌入强度，在容噪能力强的系数上嵌入多位隐密信息。从而结合了载体图像空域和小波变换域的特点，在一定程度上兼顾了容量和鲁棒性之间的均衡性，有效的提高了信息隐藏系统的容量。提取的过程，需要使用载体的原始图像，对原始图像和载体图像同时做相等程度的置乱变换，通过比较两者相应嵌入区域小波系数之间的协方差得到提取出的隐密信息。实验结果显示，基于HVS的小波域信息隐藏算法能够有效的提高信息隐藏容量，同时能够抵抗高斯噪声等一般的图像处理攻击手段。在一定程度上兼顾了容量和鲁棒性之间的均衡性，具有较好的实用性。
英文文摘	In conducting digital copyright protection at certain times or in the fields of e-government, military and legal and financial, it sometimes need to be embedded great capacity hidden information. Specially ,information hiden technology will be the focus of the future information warfare.As the key component of future information warfare ,large amounts of hiding information technology is bound to have a significant impact on the process of war and the outcome. From a practical application we needs to consider how to vector images for the communication systems embedded as many hidden messages. This is a large-capacity image Steganography algorithm. Focusing on improving Steganography algorithm capacity of the existing typical algorithm, summing up their own characteristics analyze Steganography capacity factors and increase the capacity of Steganography method. Turn the focus on human visual identity and image quality evaluation, Digital Image Scrambling digital image processing and wavelet transform related research and analysis. Based on the characteristics of human vision and wavelet domain scrambling images energy distribution characteristics, we may conclude a HVS-based steganography schemes in wavelet domain. The algorithm analysis vector images before and after scrambling in the wavelet transform domain coefficients of its energy distribution characteristics. Through the vector image scrambling it is changed its air space between the pixels relevance, achieving uniformity coefficient of its energy distribution, so as to balance the information hidden system which is not visible and robust feature. Vector images choice scrambling to deal with the low frequency of the wavelet edge as a regional embedded region. In conjunction with the visual characteristics ,we may adopt nonlinear embedding rules. Noise can be seen under the NVF function adaptive value adjustments embedded strength, Yung-noise capability of the coefficient on the number of hidden information embedded. Thereby combining the vector image airspace and wavelet transform domain characteristics, to some extent taken into account the balance between the capacity and robustness, effectively raising the hidden information system capacity. Extraction process, requiring the use of the original vector images the original image and vector images while doing the same degree of scrambling transformation. By comparing the two embedded in the corresponding regional wavelet coefficients between the covariance , hidden information may be extracted. The experimental results show that the HVS-based steganography schemes in wavelet domain can effectively improve the capacity of Steganography and resist Gaussian noise to the general image processing means of attack. To some extent,it takes into account the balance of the capacity and robustness, it is practical.