基于块分类的文档图像水印自适应算法  

An Adaptive Document Watermarking Technique Using Block Classification

作　　者：夏明一[1] 王朔中[1] 王威娜[1] 

机构地区：[1]上海大学通信与信息工程学院,上海200072

出　　处：《上海大学学报（自然科学版）》2004年第5期463-466,共4页Journal of Shanghai University:Natural Science Edition

摘　　要：该文提出一种自适应文档图像水印算法.该算法考虑到文档图像的特点,利用了图像的纹理、亮度特性,将原始图像分解成块,然后根据纹理特性,把图像块归为平滑块和纹理块两类.对于不同类别的图像块分别采用不同的嵌入方法,并根据亮度特性在不同的亮度图像块中使用不同的嵌入强度,以同时满足水印的不可见性和鲁棒性要求.An adaptive watermarking algorithm suitable for document images is proposed. Characteristics of a document image are analyzed, and the texture and luminance properties are considered. The original image is first segmented into blocks. Based on the texture property, the blocks are classified into two different categories: smooth blocks and texture blocks. Different embedding techniques are adopted in these two types of blocks. In order to meet both invisibility and robustness requirements, different embedding intensities are used in blocks of different luminance according to the luminance property.

关 键 词：文档图像 数字水印 视觉特性 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]