摘要
提出了一种面向内容认证的文档图像半脆弱数字水印算法,该算法利用DCT变换域低频系数之间的关系生成水印信息,并将水印信息嵌入到矩阵的奇异值中,对文档图像进行认证时,不需要原始载体图像,也不需要额外存储水印信息,是一个完全的盲水印算法。实验表明,该算法具有较好的透明性,对JPEG有损压缩具有较好的鲁棒性;对剪切、替换等恶意篡改表现出脆弱性。该算法达到了半脆弱水印的基本要求,是一个安全且实用的半脆弱数字水印算法。
Abstract We proposed a semi - watermarking for content authentication of document image, In this algorithm, the watermark is generated based on the relationships of low frequency coefficient in DCT domain, which embedded watermarking signals into ma- trix singular value larger data, The watermarking system is Completely blind, and image authentication can be finished without exte- rior watermark, Experimental results indicated the effective of the algorithm. This watermarking is sightless, robust for JPEG loss compressing, and fragile for images modification or tampering. The algorithm reached the basic requirements of semi - fragile digital watermarking,which is safe and useful.
出处
《信息技术与信息化》
2009年第6期105-108,共4页
Information Technology and Informatization
基金
山东省自然科学基金(2005ZRB01001)
2008年度山东省中青年科学家科研奖励基金(2008BSB38001)
关键词
文档图像
DCT变换
SVD
内容认证
篡改定位
Document image DCT transform SVD Content authentication Tamper location