Research And Implementation Of Copy-move Forgery Passive Forensics For Digital Image

The rapid development of communication technology and digital equipment has promoted the spread of digital images.As an important carrier for transmitting information through social communication modes such as the Internet,the information expressed by digital images is far richer and more attractive to eyes than words.At the same time,digital images are often used as available evidence in the fields of justice,news and so on.Influenced by traditional concept,people are more inclined to believe that all images they see are real,and often without considering the possibility that these images have been tampered or forged.However,with the emergence of various image editing applications,the cost of image editing and forgery is getting lower and lower.Authentic images and forged image are mixed together on the network,some images with maliciously tampered will even bring huge impact and damage to our social life.Therefore,how to protect the integrality and authenticity of digital images has become the concern of many researchers.Compared with active forensics technology which require digital watermarks,digital signatures or other protection information inserted in the image in advance,the passive forensics that only starts from the image itself has more application value and is also one of present research hotspots.This paper mainly study the copy-move forgery which is the most common way of tampering.The main work and contributions of this paper are as follows:(1)A copy-move forgery detection scheme based on color local binary pattern and discrete cosine transform is proposed.In this scheme,the color local binary pattern is used to replace the traditional local binary pattern to retain the color information of the image,and fuse the texture features of the image.Secondly,divide the established color LBP image into overlapped blocks and apply discrete cosine transform on each block to obtain the coefficients matrix,which the feature vectors of image blocks are extracted.Then feature vectors have been sorted by the radix sort method and obtain matched image blocks preliminarily.finally,false positives are eliminated by computing the actual distance of the matched image blocks on the image.In real life,the tampering of using a smooth region to hide a target in the image is very common.The experimental results show that the scheme can effectively detect such tampering,and locate the tampered regions accurately.At the same time,it can also obtain good detection results when there are multiple pasted regions in the image.(2)A copy-move forgery detection scheme based on SURF feature points and HSI color features is proposed.SURF algorithm is a widely used classic feature point extraction algorithm.The SURF features are invariant to rotation and scaling,and are robust to postprocessing operations such as JPEG compression and blurring.Before extracting the feature points,SURF algorithm grayed the color image and ignoring most of the color information,which leads to the mismatched caused by inconsistent color information in the detection results.In this scheme,after extracting the SURF feature points of the image and achieving the feature point descriptors,the feature points matching process is completed by the Reversed generalized 2 Nearest Neighbor algorithm and hierarchical clustering.Then the HSI color features of the feature points are used to eliminate the mismatched caused by inconsistent color information.The experimental results show that although the number of matching feature points are less,the scheme can still point out forged regions.And it is robust to both geometric tampering operations and post trace hiding operations.(3)With the aid of the GUI development tool in Matlab R2018 b,a simulation system for digital image copy-move forgery detection is designed and implemented.