The Research For Tampering Detection In A Forged Image

With extensively using and spreading the powerful and increasing mature digital image processing software, digital image is modified easier than ever. Although most people just do some daily modifications, there are still some law breakers taking advantage of the powerful and convenient digital image processing software to maliciously tamper with the image and spread counterfeit and falsified digital image everywhere, which has reached ulterior motive. Currently, the Internet and the Media are flooded with heavily altered image, seriously affecting people’s daily lives. What’s more, if tampered image is largely used for official media, scientific discovery and exhibits of court, etc., which might impact on the political and social stability. Therefore, image tamper technology brings new challenge to practical application such as judiciary forensics and police detection, and the identification and detection of tampered image forensics technology has become a hot research topic.Usually, geometric transformation such as copy-paste and size direction changing and image blurring processing operations and so on will be used in the process of tampering image. This thesis focuses on depth researching and analyzing above typical tamper operation. The main content includes as follows:Firstly, do a research on detecting copy-paste related theories and methods. This thesis introduces the current condition of copy-paste at home and abroad, analyzes algorithm thought and algorithm implementation, and compares and summarizes the detection model of such methods.Secondly, When different images needed to be altered such as copying, pasting and sheering, etc., partial image modified often need to change scale and rotate, etc., to achieve that the altered image not to be visually suspected. This thesis introduces the basic operating principle of the resample, analyzes the current resample detection technology, and summarizes the problems occurred. The new altered image detection basic on image resample is proposed. This method can automatically detect and locate tampered regions. Thirdly, in order to eliminate the obvious trace of synthetic edge and cover up the truth, the interpolator always blurs the image when tampering the synthetic image. This thesis introduces the basic principle of fuzzy operations and is done a research on the methods of fuzzy trace detection methods, and combined with the needs of image tamper detection, proposing a simple and effective automatic detection, which is good for JPEG compression with better robustness, moreover, the computing complexity is lower.Finally, the paper summarizes the work done, and with methods of image tamper diversifying, we are prospecting for the future of the image forensics technology.