A comparative analysis of copy-move forgery detection algorithms. (13th September 2022)

• Record Type: Journal Article
• Authors: Ahmad, Mohassin; Khursheed, Farida
• Is Part Of: International journal of electronic security and digital forensics. Volume 14:Number 6(2022)
• Page Start: 547
• Page End: 564
• Publication Date: 2022-09-13
• Subjects: copy-move forgery detection -- CMFD -- digital image forensics -- discrete cosine transform -- DCT -- scale invariant feature transform -- SIFT -- PatchMatch; Computer crimes -- Periodicals; Computer crimes -- Investigation -- Periodicals; Computer security -- Periodicals; Electronic commerce -- Law and legislation -- Periodicals; Information warfare -- Periodicals; 005.805
• Languages: English
• ISSNs: 1751-911X

Abstract

Copy-paste/copy-move image forgery is also known as image cloning, in which a portion of an image or entity is copied and pasted to another region of a certain picture. This category of image manipulation has the intent either to conceal the entity or to fabricate the image details. Thus, the authenticity of the photographs in different real-world implementations becomes challenging. The number of cases of image tampering is raising with the simple accessibility of image manipulation tools. Therefore, robust, precise and effective approaches to digital image forgery detection are increasingly required. A study on copy-move forgery detection (CMFD) is performed in this paper using three common schemes and their efficiency is checked and compared on images with rotation and scaling in the copied region. First, we will cover DCT-based CMFD, then adaptive over-segmentation and matching feature point. Finally, CMFD, based on PatchMatch, is addressed. The findings show the very good performance of each system.