Copy-Move Forgery Detection and Localization Using Novel Technique

Abstract

Nowadays, Image Forgery is the most extensively exploited security vulnerability in real-time applications. It necessitates the approach to discover such vulnerabilities using computer vision mechanisms. The Copy-Move Forgery has commonly appeared the image forgery in multimedia applications. Several Copy-Move Forgery methods have already been proposed for forgery detection and localization. The majority of the approaches fell short of achieving maximal precision with correct forgery localization. Other techniques have suffered from a significant computation burden. To end this, we proposed novel image forgery detection and localization framework. The proposed framework is called GFGIF (Guided Filtering with Geometric Invariant Features) for robust and accurate forgery detection and its localization. The GFGIF consists of two phases such as forgery detection and forgery localization. For forgery detection, we pre-processed the input image using a guided filter and then decomposed it into non-overlapping blocks. The Geometric Invariant features are extracted from each block. Using Euclidean Distance measure, we first discovered the candidate blocks and then detected the forgery blocks. Localization of forged blocks efficiently is another challenge for further analysis purposes. We accurately localize the forged objects in the image using the detected forged blocks. The simulation results show the efficiency and robustness of the proposed model compared to state-of-art methods.