Modified Multi-scale Feature Extraction for Copy-Move Forgery Detection Based on CMD-SIFT

The copy-move attack is a more common method in digital tampering. When copy-move forgery image occur, many important objects add or remove from the image. In order to implement forensic of the images, In the literatures many methods of copy-move forgery detection (CMFD) have been improved. The different approaches of CMFD feature-based was proposed in recent years. but, still more place to enhancement performance further. the problem of Many methods are suffering insufficient matched key points, but forgeries performance on the mirror transformed, then many feature-based methods when the forged region is of uniform texture it might hardly expose the tempering. In this paper we proposed a now scheme. In this scheme the criteria of block and keypoint features will integrate to gather in a new scheme, then multiple copy-move regions or objects will be work very well and especially when regions and objects are different sizes and contain both detailed textures and smooth.