Research Of Robust Watermarking Algorithms Based On HVS Of Image Transformation Domain

Digital watermarking confirms a work's ownership by embedding copyright information into the digital product(digital image, video,audio and so on). It becomes a new resolution for copyright protection of the digital media and the security of the multimedia and information. Robust Digital Watermarking is considered as an alternative tool increased security for protecting the copyright of multimedia and enhancing the effectiveness of intellectual property rights. This article researches the digital watermarking techique. At first,we introduce the basic theory of the digital watermark and the model of the Human Visual System. Then the main work of this article is the ralgorithms research of robust digital watermarking based on transformation domain and the model of Human Visiual System.This article mainly researches the robust digital watermarking based on Human Visual System of the transform domain. The main research contents can be summarized as follows.Firstly, A robust image watermarking algorithm based on Human Visual System (HVS) and Discrete Wavelet Transform (DWT) is proposed,and the experimental results show that the algorithm has more effective of the robustness. Host image was decomposed through the discrete wavelet transform. The JND threshold was calculated according to the human visual system characteristics, and the significant coefficients of wavelet were selected. Watermark image was embedded to the different frequency of the significant coefficients of wavelet. And watermark image is embedded by the rule of the improved binary operation, changing wavelet coefficients part of the decimal to improve the quality of the image after the watermark embedded. The experimental results show that our proposed algorithm has Enhanced the robustness effectively under the premise of guaranteeing watermark invisibility, and it is robust against JPEG compression, additive noise, zooming, cut and median filtering attack.Furthermore, an image blind watermarking algorithm based on Waston's visial model and DCT domain is presented. The original carrier image is splitted into sub-blocks, and each sub-blocks is transformed by DCT to get the DCT frequencies of the sub-image. Then according to the mask charachers of the visual model of Waston, the algorithm quantizes the frequencies by DCT transform of each sub-blocks. The locations embeded the watermarks are selected by the quantization steps which calculated based on the Waston's visiual model. The watermark image is setted chaotically and reduced the dimension, then embeded into the DC and parts of low frequencies coefficients in DCT domain of the sub-blocks with the method of quantizing the coefficients. This algorithm can extract the watermarks without the original host image and the original watermark image. The experimental results indicate that this watermarking algorithm has perfect invisible effect and robust to the common attacks, the simulation experimental results demonstrate the effectiveness of the proposed algorithm.Finally, summarized the work which the paper mainly completes and pointed out the further research directions.