Research On Digital Image Copy-move Forgery Detection Technology

Digital images have become one of the most important media to carry and spread information. At the same time, with the rapid development of image processing and Internet technology, people can tamper the content of images in any form easily, and distribute them over the world through the Internet. When these images are used in journalism, law, finance, insurance and other industries, it will bring immeasurable loss. In order to identify the authenticity and integrity of the image, the digital image forgery detection technology at the historic moment, and became a research focus in recent years.The main content of this paper is to explore the copy-move forgery detection methods of blind digital image authentication technologies. Copy-move forgery is a common way of image tampering, it's usually referred to the image manipulation that a region of image is copied and pasted to another region of the same image. As the copied area and pasted area are from the same image, they have better concealment and not easy to be perceived, but easy to operate. For copy-move forgery, first the existing detection methods are introduced and analyzed, and then two detection methods are proposed in this paper:Firstly, in order to solve the problem that some existing methods have low efficiency and storage consumption, it puts forward a copy-move forgery detection algorithm based on Discrete Wavelet(DWT) and ORB algorithm(o FAST and rBRIEF). Fristly DWT is applied on an image to decompose it into LL, LH, HL and HH; next ORB algorithm is applied on LL part to extract the key points, and construct binary descriptor vector—r BRIEF; finally, to achieve the purpose of tamper detection, hamming distance is applied to match similar key points. The experiment results show that the method can effectively detect this type of tampering for images, and the matching time is faster and storage type is less than SIFT algorithm.Secondly, in order to solve the problem that some existing methods are not sensitive to rotation and scaling, it puts forward a copy-move forgery detection algorithm based on invariant feature detector of multi-scale Gabor. First multi-scale Gabor filter bank is applied on an image to extract the key points, and the direction of key point is computed by intensity centroid; then binary descriptor vector are constructed to describe key points, and hamming distance is applied to match similar key points; finally, the false matches are removed by Random Sample Consensus(RANSAC) algorithm. The experiment results show that the method can effectively detect copy-move forgery, and has better real-time performance. At the same time, it has better robustness against rotation and scaling.Finally, the remaining problems of the paper and the next research direction are discussed, and urgent problems in this field are listed.