基于遗传算法的数字水印嵌入位置的优化算法  

An Optimization Algorithm of Digital Watermarking Embedding Positions Based on the Genetic Algorithm

作　　者：崔攀[1] 王慧琴[1] 张小红[2,3] 

机构地区：[1]西安建筑科技大学信息与控制工程学院,陕西西安710055 [2]西安建筑科技大学管理学院,陕西西安710055 [3]西安科技大学通信与信息工程学院,陕西西安710054

出　　处：《电脑知识与技术》2016年第3X期207-211,共5页Computer Knowledge and Technology

基　　金：教育部高等学校博士学科点专项科研基金(20126120110008);教育厅提供专项科研项目(2013JK1144);陕西省自然科学基础研究计划项目(2012JQ8021);陕西省教育厅产业化项目资助(2011JG12);西安建筑科技大学校青年基金(QN1429);西安科技大学培育基金项目(201256)

摘　　要：在数字图像水印的水印嵌入与提取过程中,常选用嵌入位置等作为密钥,目前在选取嵌入位置时大多数方法都是直接通过肉眼与经验去人为选取,难以达到准确且具有不稳定性,而嵌入位置的选取将直接影响到嵌入水印后图像的质量。针对该问题,本文提出了一种基于遗传算法的数字水印嵌入位置的优化算法,智能而又准确,这里将衡量图像质量的峰值信噪比作为适应度函数,峰值信噪比越大,图像的质量越好,即嵌入水印后对图像质量的破坏越小。实验结果表明,利用遗传算法得到的水印嵌入位置嵌入水印后,图像的峰值信噪比有明显提升,即图像的视觉质量大大提高了,且水印信息安全和鲁棒性又能得到很好保证,实验效果较好。In the process of embedding and extracting of the digital image watermarking, mainly choosing as key parameters,such as embedding position. At present when choosing embedding position, most of methods are choosing artificially by human eyes and experience, which is difficult to achieve accuracy and has instability. And selection of embedding position will directly affect the quality of the image with watermark. In order to solve this problem, this paper presents an optimization algorithm of digital watermarking embedding positions based on the genetic algorithm, which is intelligent and accurate. Using peak signalto-noise ratio which is measuring the image quality as a fitness function. The greater the peak signal-to-noise ratio, the better the quality of the corresponding image, namely the smaller damage to the quality of the image after embedding the watermark. The experimental results show: With the use of genetic algorithm to optimize watermark embedding position, peak signal-to-noise ratio of the image is promoted obviously, which improves the visual quality of the image. And the information security of the watermark and robustness is very good. The experimental effect is better.

关 键 词：数字水印 鲁棒性 不可见性 遗传算法 峰值信噪比 离散余弦变 

分 类 号：TP309.7[自动化与计算机技术—计算机系统结构] TP18[自动化与计算机技术—计算机科学与技术]