Research On Capacity And Energy Of Digital Watermarking

Along with the development of information society in twenty-first Century,the information security is becoming increasingly prominent.As one mean to solve the copyright protection of digital multimedia works,digital watermarking technology has caused extensive concern.At present,for most of the digital watermarking algorithms,the optimal capacity and energy of embedded digital watermarking are determined by repeated experiments.In order to overcome this limitation,the following main research is carried out in this paper:(1)The watermarking model based on filling ball is put forward in this paper.The formula of capacity and the quantitative relationship between capacity and peak signal to noise ratio(PSNR)are proposed.There are not any probability statistical assumptions about the attack channel in this model.The attack behavior is quantified as a geometric distortion.(2)The capacity formula of upper and lower bounds of the general watermarking model based on bit replacement are proposed in this paper,and the simulation of capacity of the watermarking model is realized.As the formula of capacity is a non analytical form,the result can be used to estimate the capacity.(3)The watermarking algorithm based on vector inner product is put forward in this paper.The embedding method of watermarking bite is based on additive embedding.The detection method of watermarking bite is based on vector inner product.The formula of PSNR of the watermarking algorithm is proposed.(4)The improved watermarking algorithm based on vector inner product is put forward in this paper.The formula of PSNR and the formula of bit error rate(BER)are proposed.Conclusions are confirmed by multi group experiments in this paper.For common attacks such as Gaussian white noise,amplitude scaling,and Gaussian filtering,the improved watermarking algorithm based on vector inner product has good robustness.The introduction of digital image scrambling technology can effectively resist the destruction of cropping attack on the content of digital watermarking image.The introduction of feature point matching can effectively resist geometric attacks.The research of this paper can be applied to the concrete digital watermarking scheme,which has a certain practical value.