Research On Image Copy-move Passive Forensics Algorithm Based On Feature Points And Local Search

With the rapid development of science and technology and the wide application of computers.,digital images can be easily tampered.Among the types of images forgery,copy-move forgery is one of the most common tampering methods.Copy move forgery which tampering with the same image is detected difficultly because of the similarity of Illumination angle and Imaging principle.The reliability of digital image is greatly impacted..Copy-move passive forensics identifies the copy-move forgery of the same image through similar regions,This paper mainly focuses on some problems in the traditional passive copy-move forensics technology.The algorithm based on image block has high time complexity and weak robustness to affine transformation such as scale and rotation,and the algorithm based on feature points can not locate the tampered area well through feature points,and so on.In order to overcome the shortcomings of traditional algorithms,a new passive forensics algorithm based on SIFT and search algorithm is proposed.The SIFT algorithm is used to obtain the feature points and corresponding local features in the image.The image is segmented by an adaptive scale segmentation algorithm,and the feature points are matched between the segmented image blocks.A two-stage image search algorithm is proposed,In the first stage,the local search algorithm runs in the image block stage,ZNCC algorithm is used to get the contour of the tampered area,and then edge matching algorithm is used to match the edge blocks of the tampered area at the pixel level.Finally,the details of the image are enriched by search the edge pixel level of the tampered area.The experimental results show that the scheme has better performance than the existing copy-move forgery detection algorithm,and the algorithm also has good performance under various attacks,but it does not perform very well in the presence of high noise.In order to solve the problem of noise and overcome the shortcomings of traditional algorithms,a new copy-move passive forensics algorithm based on BRISK and search algorithm is proposed,which uses BRISK as the image feature.In order to better deal with some post-processing transformations,the local search algorithm is modified to match the source matching detection unit with the surrounding eight neighborhoods.In the imageblock-based search algorithm,PCT feature is used as the feature of similarity matching,ZNCC algorithm is used to calculate edge matching,and finally the details of tampered area are processed by the search algorithm based on pixel level.The experimental results show that the proposed algorithm is effective.This algorithm can identify tampered area very well.It performs well under various attacks.Especially in the aspect of noise,even in the high noise environment,it can get better experimental results.