An Analysis of Copy-Move Forgery Detection based on Discrete Cosine Transform against JPEG Compression and Brightness Changes Attack

Copy-move forgery (CMF) is an image manipulation in which a component of the image is copied and pasted onto another portion, usually to hide undesirable aspects of the image. The objective of copy-move forgery detection (CMFD) is to identify image regions that are identical or very similar. However...

Full description

Saved in:
Bibliographic Details
Published in2024 1st International Conference on Cyber Security and Computing (CyberComp) Vol. 1; pp. 45 - 50
Main Authors Hamidon, Nurul Rihhadatul'aisy, Md Salleh, Siti Fadzlun, Abd Warif, Nor Bakiah
Format Conference Proceeding
LanguageEnglish
Published IEEE 06.11.2024
Subjects
Brightness
CoMoFoD
copy-move forgery detection
DCT
Discrete cosine transforms
Forgery
GRIP
Hands
Image coding
MATLAB
Measurement
Object recognition
Software
Transform coding
Online AccessGet full text
DOI10.1109/CyberComp60759.2024.10913676

Cover

More Information
Summary:Copy-move forgery (CMF) is an image manipulation in which a component of the image is copied and pasted onto another portion, usually to hide undesirable aspects of the image. The objective of copy-move forgery detection (CMFD) is to identify image regions that are identical or very similar. However, the tampered region in copy-move forgery is difficult to identify because the forge image appears to be identical from the original. In this research, Discrete Cosine Transform (DCT) which is a block based method, will be evaluated using F-measure metric to spot forged region on images with JPEG compression and brightness changes attack. To represent both attacks, this research used ten images in two available datasets which are CoMoFoD and GRIP datasets. These two attacks will be applied in image using GIMP editing software to three different parameters on compression quality and brightness manipulation. Experiments for this research were conducted using MATLAB. The results show that DCT-based copy-move forgery detection is able to detect at least 93% score for 90% compression quality but decreased to 88% for 70% quality. Meanwhile for brightness changes, DCT-based can detect at least 93% for (0.01,0.95) but decreased to 92% for (0.01,0.80) parameter.
DOI:10.1109/CyberComp60759.2024.10913676