The Image Watermarking Based On Sparse And Redundant Representations

In the Internet era, efficient, accurate digital watermark technology enables that multimedia has become more important than ever. Watermarking method is that characteristical watermark embedded in the image or audio file in order to protect ownership, watermark is usually is stable resistance to aggressive, however, media pirates may try to delete it. In order to verify the ownership of digital watermarking algorithm design, we must detect the presence of a particular watermark. Under normal circumstances, we chose a simple low complexity watermark embedding/detection algorithm..In addition to copyright protection, digital watermarking can also be used for content verification to prevent the spread of illegal media.We often use the specific watermarking method to suit specific application requirements. But the most important themes is the accurate detection in all the applications of watermark. Here, we propose a detection method that not only can a high probability to the digital watermark is detected (under specified conditions), but doing so can no error estimate. Therefore, ownership protection is established in the watermark is detected correct. Based on the above characteristics, we made the following research and exploration:1.First, We research the basic concept of the digital watermark, classification, the basic model of the watermark and the performance evaluation of the digital watermark.2.Second,we research compressed sensing and Sparse-land signal model, and then focus on the structure of the dictionary in the Sparse-land model of the signal processing, Based on this basis,we put forward construction algorithm of MOD and K-SVD dictionary, so as to effectively sparse and represent the image.3. Based on the MOD and K-SVD dictionary algorithm, we research digital watermarking technology applied the sparse redundant representation Respectively, we apply the MOD and K-SVD dictionary to sparse and represent the image.and the we embed the digital watermark, detect the digital watermark.finally, we can know that K-SVD dictionary with respect to the MOD dictionary algorithm has strong stability.