Staff View: Forgery detection algorithm based on texture features

Forgery detection algorithm based on texture features

Any researcher's goal is to improve detection accuracy with a limited feature vector dimension. Therefore, in this paper, we attempt to find and discover the best types of texture features and classifiers that are appropriate for the coarse mesh finite differenc (CMFD). Segmentation-based fract...

Full description

Saved in:

Bibliographic Details
Main Authors:	Ahmed I.T., Hammad B.T., Jamil N.
Other Authors:	57193324906
Format:	Article
Published:	Institute of Advanced Engineering and Science 2023
Tags:	Add Tag No Tags, Be the first to tag this record!

id	my.uniten.dspace-25968
record_format	dspace
spelling	my.uniten.dspace-259682023-05-29T17:05:49Z Forgery detection algorithm based on texture features Ahmed I.T. Hammad B.T. Jamil N. 57193324906 57193327622 36682671900 Any researcher's goal is to improve detection accuracy with a limited feature vector dimension. Therefore, in this paper, we attempt to find and discover the best types of texture features and classifiers that are appropriate for the coarse mesh finite differenc (CMFD). Segmentation-based fractal texture analysis (SFTA), local binary pattern (LBP), and Haralick are the texture features that have been chosen. K-nearest neighbors (KNN), na�ve Bayes, and Logistics are also among the classifiers chosen. SFTA, LBP, and Haralick feature vector are fed to the KNN, na�ve Bayes, and logistics classifier. The outcomes of the experiment indicate that the SFTA texture feature surpassed all other texture features in all classifiers, making it the best texture feature to use in forgery detection. Haralick feature has the second-best texture feature performance in all of the classifiers. The performance using the LBP feature is lower than that of the other texture features. It also shows that the KNN classifier outperformed the other two in terms of accuracy. However, among the classifiers, the logistic classifier had the lowest accuracy. The proposed SFTA based KNN method is compared to other state-of-the-art techniques in terms of feature dimension and detection accuracy. The proposed method outperforms other current techniques. � 2021 Institute of Advanced Engineering and Science. All rights reserved. Final 2023-05-29T09:05:49Z 2023-05-29T09:05:49Z 2021 Article 10.11591/ijeecs.v24.i1.pp226-235 2-s2.0-85116528024 https://www.scopus.com/inward/record.uri?eid=2-s2.0-85116528024&doi=10.11591%2fijeecs.v24.i1.pp226-235&partnerID=40&md5=20c52d967245f33e970ab66a01532061 https://irepository.uniten.edu.my/handle/123456789/25968 24 1 226 235 All Open Access, Gold, Green Institute of Advanced Engineering and Science Scopus
institution	Universiti Tenaga Nasional
building	UNITEN Library
collection	Institutional Repository
continent	Asia
country	Malaysia
content_provider	Universiti Tenaga Nasional
content_source	UNITEN Institutional Repository
url_provider	http://dspace.uniten.edu.my/
description	Any researcher's goal is to improve detection accuracy with a limited feature vector dimension. Therefore, in this paper, we attempt to find and discover the best types of texture features and classifiers that are appropriate for the coarse mesh finite differenc (CMFD). Segmentation-based fractal texture analysis (SFTA), local binary pattern (LBP), and Haralick are the texture features that have been chosen. K-nearest neighbors (KNN), na�ve Bayes, and Logistics are also among the classifiers chosen. SFTA, LBP, and Haralick feature vector are fed to the KNN, na�ve Bayes, and logistics classifier. The outcomes of the experiment indicate that the SFTA texture feature surpassed all other texture features in all classifiers, making it the best texture feature to use in forgery detection. Haralick feature has the second-best texture feature performance in all of the classifiers. The performance using the LBP feature is lower than that of the other texture features. It also shows that the KNN classifier outperformed the other two in terms of accuracy. However, among the classifiers, the logistic classifier had the lowest accuracy. The proposed SFTA based KNN method is compared to other state-of-the-art techniques in terms of feature dimension and detection accuracy. The proposed method outperforms other current techniques. � 2021 Institute of Advanced Engineering and Science. All rights reserved.
author2	57193324906
author_facet	57193324906 Ahmed I.T. Hammad B.T. Jamil N.
format	Article
author	Ahmed I.T. Hammad B.T. Jamil N.
spellingShingle	Ahmed I.T. Hammad B.T. Jamil N. Forgery detection algorithm based on texture features
author_sort	Ahmed I.T.
title	Forgery detection algorithm based on texture features
title_short	Forgery detection algorithm based on texture features
title_full	Forgery detection algorithm based on texture features
title_fullStr	Forgery detection algorithm based on texture features
title_full_unstemmed	Forgery detection algorithm based on texture features
title_sort	forgery detection algorithm based on texture features
publisher	Institute of Advanced Engineering and Science
publishDate	2023
_version_	1806423500507840512
score	13.211869

Forgery detection algorithm based on texture features

Similar Items