Research On Digital Watermarking Algorithm Based On Spread Transform

With the rapid development of multimedia technology,the management and transmission of digital information has become more convenient.However,it also brings about a series of information security problems,such as illegal copying,piracy and copyright disputes.Therefore,the application of content authentication and copyright protection for digital products is becoming more and more extensive and urgent.Digital watermarking technology uses specific algorithms to embed mark information into digital products to protect their ownership,which provides an effective solution to the copyright problem.To better realize the copyright authentication of digital watermarking,based on the spread transform watermarking algorithm,this paper studies robust image watermarking algorithm.The main work is as follows:1.Aiming at the problem that the uniform quantization algorithm may causes local distortion of non-uniform distributed carrier information,a logarithmic watermarking algorithm based on spread transform is studied.The appropriate logarithmic transformation is used to transform the carrier information of non-uniform distribution into uniform distribution,and the spread transform adaptive quantization method is used to embed the watermark.In order to improve the performance of the watermark against JPEG compression attack,a JPEG quantization table is used to improve the Watson visual model,and the improved model is used to calculate the quantization step.The experiments show that the logarithmic watermarking algorithm based on spread transform is better than the spread transform quantization algorithm.2.Based on the analysis of DCT domain watermarking algorithm,a multilevel mixed block DCT domain watermarking algorithm based on visual model is studied.The image will get more large DCT coefficients after the multi-level DCT transformation.According to this characteristic,the multi-level DCT transform is performed on the carrier image,and the coefficients of the last-stage DCT transform are used to embed watermark,which not only improves the security of the watermark system,but also enhances its robustness.3.Aiming at the problem that the DCT algorithm fails in extracting the watermark accurately after the geometric attacks,an anti-geometric attack method of watermarked image based on improved scale invariant feature transform(SIFT)is proposed.Firstly,the SIFT algorithm is improved,and then the improved SIFT algorithm is used to preprocess the watermarked carrier image on the watermark detection side.Then the corrected watermarked image is used to extract the watermark.The experiment shows that the improved SIFT algorithm not only ensures the accuracy of correction but also reduces the computational complexity.What's more,the watermarking algorithm based on improved SIFT geometric correction significantly improves the accuracy of watermarking extraction after geometric attacks.4.Based on the analysis of DWT-DCT transform watermarking algorithm,a DWT multi-subgraph watermarking algorithm based on particle swarm optimization is studied.The algorithm performs multilevel DWT decomposition on the carrier image,and then selects all the subgraphs obtained by the last-level DWT transform for watermark embedding.The particle swarm optimization algorithm is introduced to find the optimal subgraph combination weights.The optimized watermarking algorithm is proved to have good imperceptibility and strong robustness.5.In order to enhance the anti-noise performance of the algorithm,a DWT-DCT domain watermarking algorithm based on compressed sensing noise reconstruction is proposed.According to the sparse characteristic of noise,the paper analyzes the watermarked carrier information before and after the noise attack,and reconstructs the noise information by using the orthogonal matching pursuit algorithm.Then the watermark is extracted after the watermarked carrier is de-noised.Experiments show that the proposed algorithm not only shows strong robustness against various kinds of noise attacks,but also improves the robustness against attacks such as JPEG compression and filtering.