Detection Of Anthropogenic Changes In Digital Image

With the development of society, digital image plays a more and more important part in our daily life. Digital image serves as medium of communication and involves a great deal of significant information. For instance, a credible digital photo may be taken as evidence at court; everyday newspaper and magazines have a close connection with digital images; of course, it depends on digital images that doctors diagnose the patients. However, sophisticated digital cameras and photo-editing software packages are becoming ubiquitous and have strong functions. As a result, it has become relatively easy to manipulate digital images and create forgeries that are difficult to distinguish from authentic photographs. The reality and integrality of digital images are suffering badly threats. In consequence, it is vital and urgent to discover the skill to distinguish the truth from the false images. Naturally, digital image forensics technology is a new research area under this background. It can be divided into active evidence and passive-blind evidence based on whether the additional information is embedded into digital images in advance or not. As active evidence is mainly has many disadvantages, it is limited in practical application. Active evidence requires the content providers to pre-process the images. What is embedded with watermark makes the quality of the picture a decline. At present, many of the digital cameras have not been imbedded the function of watermark. Watermark is easy to attack and destroy. When we obtain the evidence, we need a third institute to be involved in, of course, the third institute needs accepting by the recognition of the two sides and the court. Therefore, only by depending on the content itself will the passive-blind evidence that is distorted and detected become a research topic of digital image forensics technology.This article takes my existing dissertation as the foundation, use the digital image passive-blind detection technology to analyze and research the most commonly forgery—copy-move forgery. The mainly research work includes the following:Firstly, introduce the research background and significance of this article, summarize the related research situations of the digital image active evidence and passive-blind evidence technologies, and point out the limitations of the digital image active evidence technology. Secondly, introduce the correlation theories of the digital image passive-blind detection technology and summarize six common digital image tampered methods, define the digital image passive-blind detection technology. Propose the detecting problem about the digital image copy-move forgery which is the problem needed to study in this article. Recommend the commonly detecting methods and compare the advantages and disadvantages respectively.Thirdly, introduce the digital image copy-move forgery detecting methods based on singular value decomposition and euclidean distance which are the key of the article and research on the key technologies about this algorithm. The key technologies of this algorithm mainly include discrete wavelet transform, singular value decomposition, euclidean distance, mathematical morphology and so on. To enhance the algorithm's detecting efficiency, I did some improvement on singular value decomposition. Finally, propose this algorithm's examination flow and summarize the steps of algorithm.Fourthly, to confirm the correctness, feasibility and validity of the algorithm in this article, use the matlab software to do some simulation experiments on the copy-moved images. Give the detecting results of gray images, color images, images under the different JPEG compression factor and adding Gaussian white noise with different SNR and analyze the experimental results at last. The experimental results show that this algorithm has the strong examination ability, anti-noise and anti-JPEG compression ability.Fifthly, summarize the research content in this article, point out the paper's deficiency and forecast the later research work.In summary, the work in this article mainly concentrates on the method research of digital image copy-move forgery and gets some progress in the theory and application. These achievements will have the positive promotion effect to the development of digital image passive-blind detection technology.