Research On Color Image Double Watermarking Algorithm Against Geometric Attack

The digital watermarking technology becomes an important method to solving the copyright problem because of its good robustness and invisibility.However,most of the traditional watermarking algorithms were aimed at gray images and were limited in the application of the world dominated by color images.The traditional watermarking algorithm has a single function and weak anti-geometric capability.Therefore,this article aims at the problems above and proposes two algorithms of color image double watermarking algorithm against geometric attack based on the traditional watermarking algorithm.The main research content and research results of this paper are as follows:(1)The development and research status of digital watermarking are described in detail,and system model,watermark classification and the evaluation of watermark performance are described too,which provides a theoretical support for the improvement of the algorithm.(2)In order to solve the problem that the watermarking algorithm has a single function and is incapable of resisting the geometric attacking,a dual watermarking algorithm for copyright identification and protection is proposed.This algorithm embeds the binary visible watermarking to identify copyright according to the HSV space model and Discrete Cosine Transformation(DCT);embeds the invisible color watermarking to strengthen copyright protection by Discrete Wavelet Transformation(DWT)and Singular Value Decomposition(SVD),and applies Zernike moments for rotation correction.The copyright and carrier images are both color images,which greatly increase the capacity of the embedded watermark and the robustness of the watermark are enhanced.The experimental results showed under the premise of guaranteeing the quality of the image,visible watermarking has better visual visibility,and the ability of anti-geometric-attacking of invisible watermarking has been improved,and the interaction and complementary function of double watermark improves the copyright protection(3)There are deficiencies in the extraction of directional information for wavelet transforms,and in order to improve the ability of the watermark algorithm to resist geometric attacks,the algorithm of dual watermark in color image based on Contourlet transform and compressive sensing is proposed.The color watermark wasembedded into the color carrier image based on Contourlet transform(CT)and SVD decomposition,and then the second watermark was generated by Block-based Compressive Sensing(BCS)from the color image embedded the copyright watermark,and was saved as zero watermark.When the copyright identification was needed,the zero watermark was reconstructed by the reconstruction algorithm of Contourlet and Smoothed Projected Landweber(Contourlet-SPL),then Scale Invariant Features Transform(SIFT)principle was applied to correct the detection image subjected to a geometric attack,finally,the copyright watermarking was extracted to verify the copyright.The experimental results also showed that the algorithm of dual watermark in color image based on Contourlet transform and compressive sensing had good invisibility,and in addition to the better robustness to conventional attacks,the ability to resist geometric attacks was improved.