Research Of Gray Image Watermark Algorithm Based On DWT

The rapid development of Internet, multi-media technology and communication technology has brought great convenience to people's lives, however, just as every coin has two sides, there are also some negative impacts such as information theft, piracy and so on. Under this circumstance, the digital watermark technology came into being. In recent years, the watermark technology has caught more and more people's attention and has become one of the most effective measures to settle the problems of copyright protection. In this paper, the digital watermarking algorithm of the gray image based on the discrete wavelet transform (DWT) domain is proposed.This paper proposes a blind digital watermarking algorithm with odd-even extraction algorithm based on discrete wavelet transform (DWT) domain. Before the embedding of the watermark image, the first step is scrambling it by Arnold transform and using Logistic chaos sequence to make the further encryption. In the process of embedding the watermark image, matrix norm calculation is introduced on the basis of the algorithm with odd-even extraction. A blind digital watermarking algorithm is designed with the combination of matrix norm and odd-even extraction. In this scheme, the vector image should be first transformed into secondary wavelet transform and the low-frequency coefficients are divided into several sub-blocks of the same size, with the number of the sub-blocks being more than or equal to the size of the watermark. Then the two matrix norm of each sub-block is calculated and the embedding position of the watermark is selected through Arnold transform. Finally, in order to achieve the purpose of watermark embedding, use odd-even extraction to embed the watermark to each sub-block matrix norm and modify the low-frequency coefficients of each sub-block according to the matrix norm. The extraction doesn't need the original image.Secondly, this paper proposes a non-blind digital watermarking algorithm of gray image under discrete wavelet transform (DWT) domain. This kind of algorithm also conducts the encryption with the application of chaos sequence before the embedding of the watermark. First, the vector image should be transformed into secondary wavelet transform and embed the watermark into the low-frequency wavelet coefficients. Make the multiple comparisons of the differences between the three consecutive values of each row in the secondary low-frequency wavelet coefficients of the vector image. Set up a new matrix to save the group of values, when the maximum of the differences is less than the threshold value. Extract and save the coefficient values of each row in turn, and then applying Arnold transform select the embedding position of the watermark in the matrix newly generated, with each pixel of watermark corresponding to two wavelet coefficients. Compare the two wavelet coefficients using to embedding watermark, then embed the watermark into the matrix newly generated with the application of the embedding formula. Then the changed coefficients are re-assigned to the wavelet coefficient matrix to conduct the inverse wavelet transformation and the vector images embedded with the watermark are thus formed. When extracting the watermark, the original vector image is needed to determine the location of the embedding, the watermark can be extract.The author conducted some simulation experiments and principle analysis to the two algorithms, and the results show that compared with the original odd-even extraction algorithm, the adding of norm calculation makes the invisibility and robustness of the watermark improved. The newly proposed non-blind watermarking algorithm also has good invisibility, and also has a good capability to resist the attacks such as noise, cropping, JPEG compression, filter and so on.