Study On Key Technology For Digital Image Forensics

With the rapid development of the Internet and the spread of low priced digital image capturing devices such as digital cameras, people can access to a large number of digital photos and digital images through the network, and create digital images by digital cameras and other digital image capturing devices easily. The continuously emerging of image forgery events subvert the traditional concept of seeing is believing and cause serious negative effects on political, cultural, news, scientific researches and so on. Therefore, the problem of authentication for digital image forgery attracts much more attention and becomes one of the most important issues needing to be resolved. Studies on this problem is not only far-reaching scientific significance to regularize media images and to ensure the authenticity of scientific achievements but also of very important political significance to ensure trust public order and fight against crime as well as safeguard judicial justice.In this paper, we discuss the background and significance of the blind forensic technology for digital images as well as the current situation and existing problems at home and abroad in detail. Besides, the current system architecture of digital image forensics and key blind forensic technologies for digital image forgery are analyzed systematically. The changes of invariant features and partly statistic properties caused by image tampering are studied by using the methods of mathematical statistics and pattern recognition. Consequently, we propose a new forensic method for image tampering based on moment invariants and an improved tamper forensic technology based on the statistical characteristics caused by double JPEG compression. The research results and technology innovations can be seen as follows:Firstly, after studying the common tampering means and the basic architecture of digital image forensics proposed by the research team with the core of Shih-Fu Chang at Columbia University as well as the architecture of an digital image forensics proposed by Zhou Linna, a more comprehensive and more perfect system architecture of digital image forensics is proposed in this paper, which consists of the common tampering means, the common forensic technologies and the standard database of digital forgery images.Secondly, After studying the definitions and properties of Radon transform and analytical Fourier-Meilin transform, we propose a new blind forensic method for image forgery using moment invariants based on Radon transform and analytical Fourier-Mellin transform, and give concrete steps of using this method, then analyze their feasibility.Thirdly, After studying the tamper forensic technology based on the changes of statistical characteristics of DCT coefficients caused by double JPEG compression and the double JPEG compression detection technology based on searching JPEG Ghosts, we propose an improved tamper forensic technology based on the statistical characteristics caused by Double JPEG Compression, which can solve the problem that the existing double JPEG compression detection technologies fail in detecting double JPEG compression whose DCT blocks are different during the two compresses. In addition, the method proposed in this paper can accelerate the speed of detection and increase the availability and precision of detection.Finally, the paper is summarized and the prospective developing direction of the digital image forensic technology is also put forward.