Research On Watermarking Algorithm Of Identification For Digital Image Authenticity

The arrival of digital age makes some view and literary works be recorded and stored as digital images, the transmission of digital images also becomes very convenient due to the rapid development of Internet technology. However, when people enjoy the convenience and joy brought about by digital age, the digital images also impose some negative impacts on people's lives. For example, in order to reap commercial benefits, a variety of actions of illegal copying and pirating on the digital images commonly occur, for some personal preferences or economic purposes, a computer user can use lots of image editing softwares to modify and forge a digital image easily. As a result, more and more people begin to question the things which are so-called "What I Saw", the authenticities of digital images are continuously suspected.The authenticity of a digital image generally consists of two aspects, one is the authenticity of the digital image's source, the other is the authenticity of the digital image's content. The authenticity of an image's source refers to the copyright of the digital image, the authenticity of the ownership, which indicates whether an image is illegally copied, pirated, and etc. The authenticity of an image's content refers to the originality of the image's content, in other words, the authenticity indicates whether the image has been maliciously tampered, including compression, scaling, cropping, pixel-level changes and other operations. At present, the identification of a digital image's authenticity commonly involves the following three techniques, the methods based on digital signature technology, the methods based on digital watermarking technology, and the blind image identification technology which is emerging gradually. The digital watermark hiding technology has developed for long time, it embeds the watermark in the host image without affecting the image quality, and realizes the authenticity certification of the source or content of the digital images. This method's maneuverability is better, and the method can resist a variety of malicious modification. Because of the high robustness, the algorithm has become an effective technology in the field of digital image authenticity.The watermarking algorithms which are used for identifying the source authenticity of a digital image often first embed watermark information in digital images which are needed to be protected, finally complete image source tracking works by utilizing the integrity of the extracted watermark. The watermark image will be used as an invisible data hiding in the original image, this will protect the image's ownership without changing the visual effect and content integrity of the image. The robustness of current identification algorithms on image source authenticity using watermarking method is prevalently not very high, so this chapter presents a SVD-based robust watermarking algorithm. The algorithm first extracts the DC coefficients of the original DCT transformed image, then executes 2×2 block-SVD decomposition on the collections of DC coefficients, at last, embeds watermark information in the max singular value of every block utilizing the parity quantification principle. Arnold and Logistic chaotic scrambling encryption algorithms are used before watermark is embedded, ensuring the security of the algorithm better. Experiments show that the algorithm greatly increasing the watermark embedding quantization step and it does not significantly reduce the transparency of the watermark while improving the robustness of the algorithm. The important parts of the extracted watermark images are obvious in this algorithm, the copyright, ownership and other reliable sources information of host images are confirmed, tracking the sources of digital images better.Because fragile watermarking technique is very sensitive to image tampering operations, it is often used to identify the content authenticity of digital images. With the development of watermarking technology, the semi-fragile watermarking techniques which possess certain robustness gradually rise. Such algorithms can effectively increase the capacity of resistance to acceptable handlings while ensuring the localization accuracy of tampering areas.In order to increase the robustness of the semi-fragile watermarking technique, we propose a semi-fragile digital watermarking algorithm for identifying the authenticity of image content based on Slant transformation. First, divide the host image into three groups, select a 8×8 pixel block from one of the groups to generate the watermark. The process of watermark embedding combines with Slant transform, and embeds the generated watermark information into the two blocks which are related to the former block. The algorithm proposed the "relevant blocks group" concept innovatively, ensuring that the distance between any two related pixel blocks was long enough, what is more, the introduction of positioning reference blocks improve the positioning accuracy when the embedded images are tampered. Simulation results show that this method has high robustness against occasional attacks, such as JPEG compression, noise, meeting the requirements that semi-fragile watermarking algorithm must have certain robustness. The algorithm is simple and gets high precision for digital image tampering areas detection and localization, and is very significant for the protection of content authenticity information of digital images.