Research On Object-based Detection Method For Image Copy-move Tampering

In recent years,with the rapid development of multimedia technology,and the widespread using of more and more powerful image editing softwares(such as Photoshop?isee),the security issue of the image content has become a hot research topic in information security circles.In this paper,we study the object-based detection method for image copy-move attack by using the theory of image salient region extraction and spectral clustering based super pixel segmentation.The main works are as follows:We proposed an object-based digital image tampering detection method against Copy-Move attack.In the proposed method,we firstly extract the salient objects of image by using a context-based salient objects extraction approach,and then isolate the objects in the graph,finally,calculate the ratio of Hu moment and Zernike moment of each image object.Then we detect Copy-Move attack according to a defined rule..Simulation experimental results show that the proposed method is of high accuracy,and good timeliness.It is robust for geometric transformations such as rotation and scaling.We propose a Copy-Move tampering detection method by using super pixel image segmentation and feature point matching.In the proposed method,the algorithm of spectral clustering is applied to super pixels image segmentation,it is more efficient than others.For each piece of image block,we extract the SUFR feature points,and then match the feature points to find potential repeated areas.The proposed method can accurately detect the Copy-Move attack.It realizes the compromise between robustness and tampering localization ability,and has the robustness against geometric distortion.