Based On The Transform Domain Image Robustness Of Watermark Algorithm And Realization Of The Statistical Properties

As digital and network technology is developing rapidly, digital products and information are widely dispersed online, making information sharing and access easier and quicker. However, the problem of intellectual property protection and information security is surfacing. Thus how to solve this problem is one of research focuses over a long period of time.The digital watermarking has sprung up as required and has been widely studied and used. It is to add specific information to digital products without affecting the quality of original products. The information can be obtained for testing to protect the copyright.The author in this thesis does research on the technology of digital watermarking based on DTC and Contourlet transform.The thesis first gives a brief analysis of the research background and significance of digital watermarking.It also introduces the present condition home and abroad, framework of watermarking system, typical watermarking algorithms and the evaluation criteria of its performance.It gives a detailed description of relevant theories and statistic characteristics of DTC transform. In terms of watermarking as weak signal, statistic distribution model is set up by means of Gaussian distribution. The dualistic hypothesis is used to analyse testing problems in watermarking processing system. Combined with theories of weak signal testing, the author deduces the testing algorithm where optimized multiplication is embedded, realizes the algorithm through matlab, and conducts the experiment to assure its robustness.The thesis focuses on relevant theories of Contourlet transform and watermarking algorithm. In the process, watering embedding is adopted in high frequency component according to HVS system. The author chooses the greatest sequence in variance as the embedding point to make sure its excellent invisibility. Blind detection together with inferential theory of statistics is applied to define watermarking testing as problems in dualistic hypothesis to solve them. Matlab has proved this algorithm robust.In the end, the thesis draws a conclusion and expects further research on digital watermarking.