期刊文献+

基于分块和NSCT-SVD的彩色图像数字水印算法 被引量:6

Digital watermarking algorithm for color images based on blocks and NSCT-SVD
下载PDF
导出
摘要 针对目前彩色图像数字水印算法水印容量较小、水印攻击强度较大时鲁棒性弱的问题,提出了一种基于分块和NSCT-SVD的彩色数字水印算法.将彩色载体图像分解成R,G,B共3个通道,同时将每个通道分解成4个非重叠块,分别对这些非重叠块进行NSCT处理,可得到各非重叠块对应的低频子带信息,并对低频子带信息进行奇异值分解.同时将彩色水印图像的R,G,B通道信息进行奇异值分解,考虑水印鲁棒性和隐蔽性均能较好权衡的原则,优选水印嵌入因子,将水印图像R,G,B通道信息的奇异值加入到原图像各块的低频子带信息的奇异值中,实现彩色图像数字水印嵌入.试验结果表明:该算法嵌入水印的图像保真度较高、水印的隐蔽性较好,同时对包括噪声、滤波、图像剪切、JPEG压缩、旋转等攻击均具有较强的鲁棒性,尤其提高了攻击强度较大时水印的鲁棒性. To solve the problems of low watermark capacity and bad robustness of digital watermarking algorithm for color images,a method was proposed based on blocks and NSCT-SVD. The color image was divided into three channels of R,G and B,and the three channels were respectively broken down into four non-overlapping blocks. NSCT was applied to the non-overlapping blocks for achieving the lowfrequency part of each block,and SVD was applied to every low-frequency part. SVD was also applied to the three channels of R,G and B of color watermark images. Considering the principle that the robustness and invisibility of watermark could be well balanced,the optimal watermark embedding factor was selected. The singular value of R,G and B of watermark was added to the singular value of lowfrequency part of each block of original image to embed watermark information. The experimental results show that the proposed algorithm has good transparency and robustness for many attacks of noise,filtering,cropping,JPEG compression and rotation.
作者 张倩倩 詹永照 林涵阳 ZHANG Qianqian ZHAN Yongzhao LIN Hanyang(School of Computer Science and Communication Engineering, Jiaugsu University, Zhenjiang, Jiangsu 212013, China)
出处 《江苏大学学报(自然科学版)》 EI CAS CSCD 北大核心 2017年第5期582-589,共8页 Journal of Jiangsu University:Natural Science Edition
基金 国家自然科学基金资助项目(61170126)
关键词 彩色图像数字水印 NSCT SVD 非重叠块 鲁棒性 color image watermarking NSCT SVD non-overlapping block robustness
  • 相关文献

同被引文献42

引证文献6

二级引证文献14

相关作者

内容加载中请稍等...

相关机构

内容加载中请稍等...

相关主题

内容加载中请稍等...

浏览历史

内容加载中请稍等...
;
使用帮助 返回顶部