Research Of Patchwork Algorithm For Image Watermarking Based On Transform Domain

Nowadays, the widespread of huge multimedia digital products on the network brings great convenience to our study and life. But it is very easy to do a digital copy and modify which is a serious violation of the author’s copyright and interests. The advent of digital watermarking technology for digital products is a good way to solve copyright issues provide. This paper studies the color image digital watermarking technology and proposes the algorithm to embed a color watermark into a color image at transform domain by improved patchwork method.This paper proposes a method for improved patchwork combination of discrete wavelet transform. The improved patchwork method use two sets groups: one collection is about RGB color of images, the other group is about the high-frequency, medium-frequency and lowfrequency of images. In order to ensure the robustness of the watermark, the watermark is embedded into the two-level discrete wavelet decomposition of the low-frequency of vector image which is three-level discrete wavelet decomposition of low, medium and high frequency; in order to ensure the transparency of the watermark, the watermark on the green component on the increase or decrease in brightness is opposite with blue and red component on the increase or decrease in brightness is the when embedding watermark, and the low-frequency wavelet decomposition on three levels of brightness changes is opposite with the high-frequency wavelet decomposition brightness changes on the same level. In order to ensure the same degree of similarity to extract watermark and embedding watermark as high as possible, the low-frequency part of the watermark embedded into vector image four times. When extracting the watermark, remove the four embedded low-frequency components of watermark and take the average, and use the average as the final watermarked image extracted low frequency component. The experimental results and contrasts show that the algorithm has good robustness and transparency, but for the low frequency of Gaussian low-pass filtering and JPEG compression attacks, due to the similarity lead to lower for background color deepened, so this article also proposed another method.In order to improve the similarity of the extracted watermark and the original watermark by Gaussian low-pass filtering and JPEG compression attacks. Discrete cosine transform has a feature of high concentration of the energy. The paper uses the feature to presents a combination of discrete cosine transform and discrete wavelet transform method for improving patchwork method. In order to ensure the robustness of the watermark, the algorithm will remove each of the discrete cosine transform block low-frequency components to re-form a new matrix, then does level discrete wavelet transform for the new matrix. In order to guarantee transparency and similarity of extracted watermark image and original image, the algorithm uses the improved patchwork method to embed the low-frequency part of RGB components of the watermark image after discrete wavelet transform into RGB components of vector images after discrete cosine transform and discrete wavelet transform correspondingly; extraction time, removes the four low-frequency components embedded and averages for the ultimate low-frequency component as a watermark. Experimental results and contrasts show that the algorithm has a good robustness for variety of attacks, and has better robustness for Gaussian low-pass filtering and JPEG compression attack than previous algorithm.Finally, this paper probably summarizes the study of the contents, and suggests the future research directions.