数字博物馆中的图像盲水印技术  

Blind Digital Watermark for Image in Digital Museum

作　　者：周季峰[1] 武港山[1] 唐杰[1] 袁春风[1] 

机构地区：[1]南京大学计算机科学与技术系软件新技术国家重点实验室,江苏南京210093

出　　处：《计算机应用研究》2006年第5期236-238,241,共4页Application Research of Computers

基　　金：国家"863"计划资助项目(2002AA117010-10);2005年教育部科技基础条件平台建设项目

摘　　要：针对数字博物馆中图像数量大、类型多的特点,提出一种适用于多种类型图像的盲水印算法。该算法将用户ID作为水印信息,从而实现对图像的操作跟踪。水印嵌入时,首先将水印信息转换成ASCII码形式,并对其进行交织和BCH编码以提高抗突发和随机错误的能力,然后根据图像类型取得嵌入水印的载体,再对该载体语义丰富的位置进行分块DCT变换,最后通过修改子块DCT系数间关系嵌入水印。水印检测时,以嵌入过程的逆过程抽取水印信息,并计算其与所有用户ID的相似度来获得初始嵌入的ID。实验表明,该算法能抵抗一般的信号处理操作和特定条件下的几何变换。Because there are large amounts of and varieties of images in digital museum, a blind watermark algorithm suitable for multiple types image is proposed. In order to track the use of image, user＇ s ID is used as watermark. While embedding watermark, first the watermark is translated into corresponding ASCII code, and in order to resist against burst and random error, the watermark is then interleaved and BCH coded. We then classify the image and get the carrier for embedding watermark. At last, we do DCT on center of the carrier and modify coefficients in DCT domain to embed watermark. On extracting, watermark is extracted in a contrary embedding procedure and we compute the similarity of the extracted watermark and each user ID to find the originally embedded ID. Preliminary experiment shows that our algorithm can resist against ordinary signal processing and special geometric translation.

关 键 词：交织器 BCH码 DCT变换 盲水印 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]