Spatial domain image dual watermarking based on SVM and chaotic-sequence-mapping
Tang Xiao-fen · Jisuanji gongcheng yu sheji · 2009
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 pixels 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.