Enhanced copy-paste image forgery detection based on archimedean spiral and coarse-to-fine approach

Duplicated region detection is one of the most common blind image forgery detection techniques to detect evidence of tampering and this is done by scrutinizing clues in a copy-paste image forgery. Two main issues for detecting copy-paste image forgery are robust feature extraction and computational...

Full description

Saved in:

Bibliographic Details
Main Author:	Sekeh, Mohammad Akbarpour
Format:	Thesis
Language:	English
Published:	2013
Subjects:	QA75 Electronic computers Computer science
Online Access:	http://eprints.utm.my/id/eprint/43961/5/MuhammadAkbarpourSekehPFSKSM2013.pdf
Tags:	Add Tag No Tags, Be the first to tag this record!

id	my-utm-ep.43961
record_format	uketd_dc
spelling	my-utm-ep.439612017-06-22T01:09:49Z Enhanced copy-paste image forgery detection based on archimedean spiral and coarse-to-fine approach 2013-10 Sekeh, Mohammad Akbarpour QA75 Electronic computers. Computer science Duplicated region detection is one of the most common blind image forgery detection techniques to detect evidence of tampering and this is done by scrutinizing clues in a copy-paste image forgery. Two main issues for detecting copy-paste image forgery are robust feature extraction and computational complexity. The major and specific challenges are to improve robustness especially against rotation for small size duplicated regions and improve time complexity of block similarity detection due to blindly matching in current methods. In this study, a copy-paste image forgery detection model is enhanced by including two proposed algorithms. The algorithms are Spiral Unique Sequence feature (SUS) based on Archimedean spiral to address the robustness issue and Coarse-To-Fine (CTF) block-matching algorithm based on sequential straightforward block clustering technique to resolve the time complexity issue. For evaluating the performance of SUS and CTF, MICC-F220 dataset from University of Florence and FC2010 dataset from Universiti Teknologi Malaysia were used. To measure the robustness of SUS, two sizes of blocks including �� pixels and �� pixels were analysed and the results were compared with Zernike moment’s robustness. For the first blocksize, the robustness improvement of SUS against noise and compression were 9.6% and 1.7% respectively but, was -2.9% against rotation. However, for the second blocksize, the robustness of SUS against noise, compression, and rotation were improved by 21.3%, 18.9%, 30.8% respectively. Next, the performance of CTF computational time was analysed in different cases of the number of clusters and compared with Lexicographical-sorting method. When the number of clusters exceeded a specific threshold, the computational time of CTF matching was significantly reduced. In conclusion, the experimental results and mathematical analysis demonstrated that SUS feature with coarse-to-fine block matching algorithm have made considerable improvements in terms of robustness and time complexity thus contributing to the area of duplicated region detection in forensic science. 2013-10 Thesis http://eprints.utm.my/id/eprint/43961/ http://eprints.utm.my/id/eprint/43961/5/MuhammadAkbarpourSekehPFSKSM2013.pdf application/pdf en public phd doctoral Universiti Teknologi Malaysia, Faculty of Computing Faculty of Computing
institution	Universiti Teknologi Malaysia
collection	UTM Institutional Repository
language	English
topic	QA75 Electronic computers Computer science
spellingShingle	QA75 Electronic computers Computer science Sekeh, Mohammad Akbarpour Enhanced copy-paste image forgery detection based on archimedean spiral and coarse-to-fine approach
description	Duplicated region detection is one of the most common blind image forgery detection techniques to detect evidence of tampering and this is done by scrutinizing clues in a copy-paste image forgery. Two main issues for detecting copy-paste image forgery are robust feature extraction and computational complexity. The major and specific challenges are to improve robustness especially against rotation for small size duplicated regions and improve time complexity of block similarity detection due to blindly matching in current methods. In this study, a copy-paste image forgery detection model is enhanced by including two proposed algorithms. The algorithms are Spiral Unique Sequence feature (SUS) based on Archimedean spiral to address the robustness issue and Coarse-To-Fine (CTF) block-matching algorithm based on sequential straightforward block clustering technique to resolve the time complexity issue. For evaluating the performance of SUS and CTF, MICC-F220 dataset from University of Florence and FC2010 dataset from Universiti Teknologi Malaysia were used. To measure the robustness of SUS, two sizes of blocks including �� pixels and �� pixels were analysed and the results were compared with Zernike moment’s robustness. For the first blocksize, the robustness improvement of SUS against noise and compression were 9.6% and 1.7% respectively but, was -2.9% against rotation. However, for the second blocksize, the robustness of SUS against noise, compression, and rotation were improved by 21.3%, 18.9%, 30.8% respectively. Next, the performance of CTF computational time was analysed in different cases of the number of clusters and compared with Lexicographical-sorting method. When the number of clusters exceeded a specific threshold, the computational time of CTF matching was significantly reduced. In conclusion, the experimental results and mathematical analysis demonstrated that SUS feature with coarse-to-fine block matching algorithm have made considerable improvements in terms of robustness and time complexity thus contributing to the area of duplicated region detection in forensic science.
format	Thesis
qualification_name	Doctor of Philosophy (PhD.)
qualification_level	Doctorate
author	Sekeh, Mohammad Akbarpour
author_facet	Sekeh, Mohammad Akbarpour
author_sort	Sekeh, Mohammad Akbarpour
title	Enhanced copy-paste image forgery detection based on archimedean spiral and coarse-to-fine approach
title_short	Enhanced copy-paste image forgery detection based on archimedean spiral and coarse-to-fine approach
title_full	Enhanced copy-paste image forgery detection based on archimedean spiral and coarse-to-fine approach
title_fullStr	Enhanced copy-paste image forgery detection based on archimedean spiral and coarse-to-fine approach
title_full_unstemmed	Enhanced copy-paste image forgery detection based on archimedean spiral and coarse-to-fine approach
title_sort	enhanced copy-paste image forgery detection based on archimedean spiral and coarse-to-fine approach
granting_institution	Universiti Teknologi Malaysia, Faculty of Computing
granting_department	Faculty of Computing
publishDate	2013
url	http://eprints.utm.my/id/eprint/43961/5/MuhammadAkbarpourSekehPFSKSM2013.pdf
_version_	1747817123678257152

Enhanced copy-paste image forgery detection based on archimedean spiral and coarse-to-fine approach

Similar Items