Research On Image Digital Watermark Algorithm Based On Wavelet Transform

Digital watermark is an important application of information hiding technology.It make possible to embed digital information (watermark) that is not visible in the multi-media information (such as text, images, sound and video).The main purpose of watermark processing technology is to hide the personal information in order to protect the copyright of digital products, or to prove the reliability of products.After embedding watermark,the multimedia data as carriers must not be degraded as obvious and not affect its value.This paper has been researched on image digital encryption technology and image digital watermarking algorithm based on wavelet-domain.In a number of classical learning algorithm study,it has been pointed out the insignificance,and proposed the improved algorithm.First,the technology of traditional image digital encryption will been taken a given image into a chaotic image by a scrambling algorithm.This method is usually only one encryption key,so its security is poor.In this paper,the image will be encrypted by integration and scrambling with another three-assisted image.This algorithm has three image key,two numeral key,so its security has been improved.Second,this paper has been introduced a image digital watermarking algorithm based on threshold classification in the wavelet domain . Carrier image of the wavelet decomposition that its embed region was classified by the threshold.Different regions were embedded watermark according to the different embedding strength.The simulation results show that this algorithm can be seen good invisibility and robustness.Third,this paper has been studied on an improved wavelet-domain image digital watermarking algorithm that based on the greatest expectation algorithm (EM) classificati- on.The embedded region was classified by EM algorithm of Gaussian Mixture Model.The encrypted watermark were embedded into the carrier image with appropriate intensity by different characteristics of the classification layer.Thus this greatly enhanced invisibility and robustness of the algorithm.