节点文献

面向内容的数字图像防伪和完整性保护技术研究

Research on Content-oriented Digital Image Anti-counterfeiting and Integrity Protection Techniques

【作者】 张艳

【导师】 管海兵; 宦飞;

【作者基本信息】 上海交通大学 , 计算机软件与理论, 2010, 硕士

【摘要】 随着多媒体技术和网络技术的迅速发展与广泛应用,对多媒体数字产品的版权保护已成为迫切需要解决的问题。数字水印认证技则是一种有效的解决方法。本文主要针对数字图像水印相关技术研究,完成以下工作:1.在研究灰度图像水印算法的基础上,本文提出一种基于人类视觉系统(HVS)敏感因素调整嵌入强度的自适应水印算法。该算法根据视觉系统对图像内容的敏感度自适应的调整水印嵌入强度,达到鲁棒性和透明性的平衡。实验结果证明了该水印算法的鲁棒性和透明性。2.本文进行了面向感兴趣内容的数字水印研究,并对其图像分区问题提出一种采用遗传算法的解决方法。该方法通过同时优化分区内相似度和分区间的差异度,实现图像内容的有效划分,为下一步水印设计算法的内容特征抽取提供了有利条件。

【Abstract】 With the rapid development and widely applications of the multimedia technology and network technology, the need for digital image integrity protecting and authenticity is becoming more and more prominent. Digital watermarking is one of the effective solutions. This paper focuses on the technology related to digital image watermarking, and finished the work following:1. This paper raises a self-adapted method based on the sensibility of human visual system (HVS) against image content. It adapts the embedding-strength factor to control the number of watermark bits embedded, reaching a balance between robustness and transparence. Experiments have showed that our algorithm has better robustness and transparence.2. This paper discusses content-oriented digital watermarking. After research on its first phase——image segmentation, we raised a new method which was based on genetic algorithm. By the simultaneous optimization of similarity in the region and difference between regions, it achieves a well segmentation, which prepared for the meaningful content detection in watermark design.

节点文献中: 

本文链接的文献网络图示:

本文的引文网络