M-SIFT: A detection algorithm for copy move image forgery

摘要

Forgery is not a new concept but from the very start it has been there whether it is in documents, images, art, or literary etc. Threats against digital images has increased to that extent that they require serious attention. Copy move forgery being the most dangerous one, is affecting world of digital images as this is hardest to detect. We first discuss and demonstrate image forgery types, discuss detection methods and their shortcomings, and then discuss the idea of using Keypoint based method M-SIFT which is an improved version of SIFT. Additionally, we provide implementation and performance analysis of suggested algorithm.