DWT-DCT Domain Digital Image Zero-Watermark Technology Research

As information technology moving forward at high speed, all kinds of multimedia works, which greatly facilitate and enrich people’s lives and strengthen the exchange of information between people on the network have been widely used. The development of technology brings us great convenience. However, it also brings many troubles like that digital works can be very easily edited, modified and copied, resulting in a lot of copyright and economic disputes. How to protect the security of digital media and to protect the legitimate rights and interests of people today has become an important research topic. Digital watermark is widely used as a kind of copyright protection technology. However, the watermark’s properities of transparency and robustness seriously restrict the development of it. Then, zero watermark technology is put forward, and some progress has been made about it. Because it solves the problem of the contradiction between transparency and robustness well, zero watermark has a broad development prospects.By studying the existing classic digital image watermarking algorithms and zero-watermarking ones, through simulation, statistics and analysis of a large number of experimental data, an image feature vector from Discrete Wavelet Transform (DWT) and Discrete Cosine Transform (DCT) which is resistant to geometric attacks is used to contruct zero watermark. In this paper, the main work content is as follows:First, a digital image zero-watermarking scheme based on DWT and DCT domain is presented. Based on the the characteristics of the relative stability of the coefficient symbol, the image feature vector which is used to contruct zero watermark is extracted in a certain area of the coefficient matrix.Second, test the zero-watermarking algorithm. To verify the performance of the algorithm, the article compares respectively five different images zero watermark to the ones extracted from the changed images which have been attacked by adding noise, zooming, cutting, JPEG compression, histogram average, median filtering and so on. Finally, through the comparison between different image zero watermark, the threshold of copyright identification is determined.Third, according to the above algorithm, a simple copyright identification system is designed, which can construct the zero watermark of selected images and show the size of the selected images’similarity.