基于支持向量机和混沌序列的空域图像双水印

为解决图像同时具有版权保护和内容认证需求问题,提出了一种基于支持向量机的鲁棒水印和混沌序列与LSB相结合的脆弱水印的双重图像水印算法.利用图像邻域像素之间的相关性,通过训练回归型支持向量机模型实现鲁棒水印图像嵌入或提取操作.然后,再将已嵌入鲁棒水印的载体图像用最低有效位和混沌序列相结合的方法嵌入基于载体图像内容的脆弱水印.实验结果表明,该算法同时实现了图像的版权保护和内容篡改定位,提高了水印系统的安全性.

Spatial domain image dual watermarking based on SVM and chaotic-sequence-mapping

In order to solve the problem that some expensive images have the needs of not only protecting their copyrights but also assuring their integrity. A spatial domain image dual watermarking algorithm is proposed which is based on the machine-supporting robust water-marking and the chaotic-sequence-mapping fragile watermarking. SVM can learn the relationship between selected pixel and its 3×3 neighboring pixeis with supported vector regression. Through the way of adjusting the output value between selected pixel the trained SVR, we embed or extract the robust watermarks. After that we combine LSB and the chaotic sequence mapping to embed the fragile watermark in the existed robust watermarks. The algorithm not only implements the copyright protection of the images and the tamper localization of the contents, but also improves the security of the watermarking system.