Research On Tracing To Tamper With The Information Of Original Image Digital Watermarking Algorithm

In modern society, the digital media is widely used in various fields, the digital image is an important way for human beings to gain and exchange information, network image information explosive growth and spreading brought a great challenge of digital copyright protection, and caused many by digital image has been illegally tampered harm. The existing detection technology can not be the majority of digital image user master, professional identification means high cost and can not take into account the timeliness. Therefore, the digital image watermark embedding and extraction, has become very necessary and valuable to realize tracing back to the original information and copyright in technology.This paper according to the interactive relationship between the watermark information capacity and digital image embedded watermark invisibility and robustness of watermark attack, and according to the existing technology, the current mainstream based on spatial domain and the typical algorithm based on frequency domain transform is analyzed and studied, and on the current network the most widely used means of investigation and analysis of tamper, proposed a new watermarking method for tracing the tampered image of original information.Through the QR code watermarking introduced more information capacity, at the same time reduce the embedding capacity to ensure invisibility, block DCT transform domain AC variance to find out the hidden information in the images to the position, repeated not superposition of QR code is embedded, to ensure that as long as there is a watermark can effectively scan can get complete information. Different degree of tamper attacked algorithm to generate the vector image contains the watermark, the watermark extraction and reduction proved this algorithm has a good ability to resist different tamper attack, on the premise of ensuring imperceptibility and robustness, which greatly improved the information capacity of the watermark bearing. The main work of this paper are as follows:(1) Analysised and Researched on the spatial domain and transform domain digital watermarking embedding algorithm is typical, coded simulation analysised of existing lowest embedded mode and transform domain embedding location selection correlation algorithm results, summarized all kinds of transform, the embedded position selection and preprocessing of the image characteristics.(2) Put forward "The QR code related to the original watermark information-Discrete cosine transform DCT positioning the embedding position-Adaptive watermark generation-Multilayer embedded" algorithm:The block DCT transformation vector image is adapted to the size of watermark, and through the AC variance ranking to find the best embedding position;Introduced QR code instead of traditional watermark embedding positions, combined with the embedded position pixel gray value two value watermark adaptively generated;Aiming at the RGB layer of gray value characteristics in different layers to select multiple non overlapping embedding position multiple embedded.(3) Based on adaptive QR DCT positioning code watermarking algorithm process and key steps of implementation are described in detail, and used the Matlab to realize the embedding and extraction of adaptive QR DCT location codes based on watermarking, the core code algorithm are described, finally according to the analysis of simulation results the performance of the algorithms. Prove that embed51watermark in an image still has good invisibility, and can extract effective watermarking associated to the original information in different extent of superposition attack. This paper combined the frequency domain transform and the spatial embedding a bit put forward new ideas, the use of adaptive QR DCT positioning code watermarking algorithm can effectively solve the current network can’t track the original information and investigate the copyright problem, provided a new idea for the analysis of the same types of problems. At the end of the article, the work of this paper are summarized, and pointed out that need to be improved and perfected in practical applications of the algorithm in the future.