基于灰度值统计特征的数字水印方法  

Image Watermarking Based on Statistical Characteristics of Gary Value

作　　者：王庆梅[1] 汪春生[1] 刘凤玉[1] 

机构地区：[1]南京理工大学计算机科学系,南京210094

出　　处：《计算机工程》2005年第2期138-140,共3页Computer Engineering

基　　金：国防基础科研基金资助项目;江苏省自然科学基金资助项目(BK2002080)

摘　　要：提出了一种基于图像分块灰度值统计特征的数字水印方法,算法利用图像分块的灰度差异值判定水印信息的嵌入位置,对选定的图像分块进行DCT变换,然后将水印信息嵌入到中频系数当中。由于水印信息嵌入到了纹理复杂度因子比较大的块中,因此就增强了水印信息的不可见性。实验表明,本算法较好地保持了原始图像的品质,对非几何攻击和有损压缩具有较好的稳健性。An image watermarking algorithm based on statistical characteristics of the image segmentation is proposed in this file. The embedding location is decided by calculating the different value of the gray values of image segmentations. And then watermarking information is embedded into the meddle coefficients of the DCT transformation of the selected host image segmentations. Watermarking information is embedded into image segmentation that has larger texture complication, and the watermarking information is invisible. Experimental results show that the algorithm maintains the quality of the original image and is robust to non-geometric attack and loss compression.

关 键 词：数字水印 图像分块 灰度值 不可见性 基于图像 几何攻击 嵌入 有损压缩 DCT变换 灰度差 

分 类 号：TP391[自动化与计算机技术—计算机应用技术] TP278[自动化与计算机技术—计算机科学与技术]