摘要
基于8×8DCT分块变换,根据Watson模型的灵敏度表将水印嵌入到DCT域的低频和中频部分．为了提高安全性,根据扩频技术对原始水印进行预处理．描述了所提出的算法实例,并进行了仿真实验,对嵌入水印的图像进行了一系列恶意攻击实验．结果表明。该算法能较好地抵抗剪切、模糊、噪声、锐化等图像处理操作,具有良好的鲁棒性和不可见性．
Digital watermarking is an important branch of information hiding. This paper proposes an algorithm based on 8 ×8DCT transform. According to the Watson model＇ s sensitivity table, the watermarking embeds to the low and middle frequencies of the DCT domain. In order to improve the security, it firstly transforms the original watermarking by Spread - Spectrum. The method is introduced detailedly and experiments are done. Based on a series of attacking experiments on the embedded_ watermarking image, the proposed technique could actually survive several kinds of image processing such as image cropping, blurring, noise, sharpening and so on. It also has good robustness and invisibility.
出处
《哈尔滨理工大学学报》
CAS
2007年第2期19-22,26,共5页
Journal of Harbin University of Science and Technology
基金
黑龙江省教育厅科学技术研究项目(10051054)
黑龙江省研究生创新科研项目(YJSCX2005-249HLJ).
