Research On Image Watermark Embedding Intensity Adaptive Algorithm Based On Differential Evolution

In the era of rapid development of Internet information technology,the features like the low-cost and easy-to-transmit of digital information such as images,videos,and texts make it vulnerable to illegal attacks such as illegal copying,falsification and cutting,and make it difficult to maintain the legitimate rights and interests of the original and copyright holders.Therefore,the digital watermarking technology,in which the naked eyes can't recognize information embedded in the carrier,is proposed and used for copyright protection,ticket security,and illegal copy protection.Due to its wide practical value and important theoretical research significance,it has become one of the current research hotspots.At present,the digital watermarking technology'development history is not long,and its theoretical system is not yet complete,at the same time,the fixed embedding intensity makes the balance between invisibility and robustness unstable,resulting in weaker ability to confront stronger attacks and weakening the practicality and effectiveness of copyright certification and other aspects.This thesis mainly studies the watermark embedding algorithm that can confront stronger illegal attacks.The large change in the transform domain of Discrete Wavelet Transform(DWT)has very little change in the pixel space of the image,and the singular value matrix after Singular Value Decomposition(SVD)has great stability,so the two domains of DWT and SVD have great robustness.Due to the Differential Evolution Algorithm's characteristics such as strong robustness,global optimization and fast convergence speed,so based on the analysis of the evolutionary process of the differential evolution algorithm,this thesis combines it with the DWT-SVD to adopt a watermark embedding intensity adaptive algorithm based on differential evolution to balance and enhance the invisibility and robustness of the image watermark embedding algorithm.This thesis systematically expounds the principle,characteristics,common algorithms and research status of digital watermarking technology,and describes the watermark embedding process in two domains of DWT and SVD,and by comparing with common similar algorithms the experimental data verifies that the watermark embedding algorithm based on DWT-SVD domain has better robustness and invisibility.Based on the analysis of the basic principle,algorithm features,implementation steps and the setting of controlled parameters of the differential evolution algorithm,this thesis systematically elaborates the evolutionary theory of "the natural selection of the survival of the fittest" through the study of its application in image segmentation,and at the same time briefly introduces its commonly used fields,and goes on the feasibility analysis of the differential evolution algorithm in DWT-SVD domain.In the end,this thesis establishes the DWT-SVD domain watermark embedding system based on the differential evolution algorithm,and realizes the watermark embedding intensity adaptive algorithm by the introduction of the differential evolution idea,and by comparing a variety of similar algorithms through MATLAB simulation experiments to verify this algorithm's balance and increase in robustness and invisibility.The result shows that the image watermark embedding intensity adaptive algorithm based on the differential evolution algorithm has been greatly improved in robustness and invisibility.