Melola Discovery

← Melola Discovery

Open access Jurnal SINTA 3

Sistem Deteksi Forensik Keaslian Gambar dan Video Menggunakan Metode K-Means dan Localization Tampering

SemanTIK: Teknik Informasi · Vol. 12 No. 1 (2026) · pp. 204–213

DOI
10.55679/semantik.v12i1.265
ISSN
2502-8928, 2460-1446
Published
2026-06-30
Publisher
Informatics Engineering Department of Halu Oleo University
Language
eng
License
creativecommons.org/licenses/by-sa/4.0

Abstract

Perkembangan pesat pengeditan media digital memungkinkan manipulasi gambar dan video yang sangat realistis, sehingga mengancam integritas bukti digital. Verifikasi manual dan deteksi metode tunggal memiliki keterbatasan fundamental, seperti subjektivitas tinggi dan tingkat positif palsu. Oleh karena itu, pengembangan sistem forensik otomatis dengan pendekatan multi-metode sangat diperlukan. Penelitian ini mengusulkan sistem forensik arsitektur jalur ganda yang mengintegrasikan K-Means Clustering dan Localization Tampering, didukung dengan Error Level Analysis (ELA) serta Scale-Invariant Feature Transform (SIFT). Jalur gambar menggunakan 19 tahapan pemrosesan untuk mendeteksi manipulasi spasial (copy-move dan splicing), sedangkan jalur video menggunakan 5 tahapan untuk anomali temporal (duplikasi, penyisipan, dan penghapusan frame). Sistem dievaluasi menggunakan kerangka kerja Risk-Based Testing pada 28 kasus uji gambar dan 23 video. Hasil pengujian menunjukkan sistem forensik gambar mencapai tingkat keberhasilan 100% dengan rata-rata keandalan 92,9%. Sistem forensik video mencapai tingkat keberhasilan 91,3%, menunjukkan ketahanan tinggi terhadap anomali halus. Kesimpulannya, pendekatan multi-metode terintegrasi ini memberikan solusi yang sangat andal dan efektif untuk verifikasi keaslian multimedia dalam investigasi hukum. Penelitian selanjutnya disarankan mengintegrasikan deep learning untuk mendeteksi deepfake. Rapid advancements in digital media editing enable highly realistic image and video manipulations, threatening digital evidence integrity. Manual verification and single-method detection have fundamental limitations, such as high subjectivity and false-positive rates. Thus, developing an automated, multi-method forensic system is crucially needed. This research proposes a dual-pipeline forensic system integrating K-Means Clustering and Localization Tampering, supported by Error Level Analysis (ELA) and Scale-Invariant Feature Transform (SIFT). The image pipeline utilizes 19 processing stages to detect spatial manipulations (copy-move and splicing), while the video pipeline uses 5 stages to detect temporal anomalies (frame duplication, insertion, and deletion). The system was evaluated using a Risk-Based Testing framework on 28 image and 23 video test cases. Testing results showed the image forensic system achieved a 100% success rate with 92.9% average forensic reliability. The video forensic system achieved a 91.3% success rate, demonstrating high robustness against subtle anomalies. In conclusion, this integrated multi-method approach provides a highly reliable and effective solution for multimedia authenticity verification in legal investigations. Future work should integrate deep learning to detect deepfakes.

Keywords

Collaboration network of the authors

14 authors · 36 co-authorship links

Hover or tap a name to highlight its co-authors. Double-click to open the author's page.

  • Universitas Halu Oleo
  • Affiliation unknown
  • Different organisations
  • Same organisation
  • Affiliation unknown
  • Thicker line = more joint articles · bigger circle = more articles
Show as table
Co-authorOrganisationJoint articles
LM Bahtiar AksaraUniversitas Halu Oleo1
SutardiUniversitas Halu Oleo1
Ageng Arya Khrysna DwipanggaUniversitas Halu Oleo1
IsnawatyUniversitas Halu Oleo1
Muhammad Ihsan SaritaUniversitas Halu Oleo1
Jumadil NangiUniversitas Halu Oleo1
Asri SamsudinUniversitas Halu Oleo1
Isnawaty IsnawatyAffiliation unknown1

Affiliations come from OpenAlex; journal OAI-PMH metadata carries none. Authors without an OpenAlex ID are matched by name, which can merge or split people.

Metadata source: OAI-PMH (metadata jurnal) https://semantik.uho.ac.id/index.php/journal/oai, first harvested 26 Okt 2026, last seen 26 Okt 2026.
Original metadata (oai_dc) 
<oai_dc:dc xmlns="http://www.openarchives.org/OAI/2.0/" xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:oai_dc="http://www.openarchives.org/OAI/2.0/oai_dc/" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xsi:schemaLocation="http://www.openarchives.org/OAI/2.0/oai_dc/  http://www.openarchives.org/OAI/2.0/oai_dc.xsd">
	<dc:title xml:lang="en-US">Sistem Deteksi Forensik Keaslian Gambar dan Video Menggunakan Metode K-Means dan Localization Tampering</dc:title>
	<dc:creator>Wahab, Agrian</dc:creator>
	<dc:creator>Isnawaty</dc:creator>
	<dc:creator>LM. Fid Aksara</dc:creator>
	<dc:creator>Aksara, Muh Yamin</dc:creator>
	<dc:creator>LM. Bahtiar Aksara</dc:creator>
	<dc:creator>Subardin</dc:creator>
	<dc:subject xml:lang="en-US">Deteksi Forensik</dc:subject>
	<dc:subject xml:lang="en-US">K-means</dc:subject>
	<dc:subject xml:lang="en-US">Localization Tampering</dc:subject>
	<dc:subject xml:lang="en-US">Error Level Analysis (ELA)</dc:subject>
	<dc:subject xml:lang="en-US">Scale-Invariant Feature Transform (SIFT)</dc:subject>
	<dc:description xml:lang="en-US">Perkembangan pesat pengeditan media digital memungkinkan manipulasi gambar dan video yang sangat realistis, sehingga mengancam integritas bukti digital. Verifikasi manual dan deteksi metode tunggal memiliki keterbatasan fundamental, seperti subjektivitas tinggi dan tingkat positif palsu. Oleh karena itu, pengembangan sistem forensik otomatis dengan pendekatan multi-metode sangat diperlukan. Penelitian ini mengusulkan sistem forensik arsitektur jalur ganda yang mengintegrasikan K-Means Clustering dan Localization Tampering, didukung dengan Error Level Analysis (ELA) serta Scale-Invariant Feature Transform (SIFT). Jalur gambar menggunakan 19 tahapan pemrosesan untuk mendeteksi manipulasi spasial (copy-move dan splicing), sedangkan jalur video menggunakan 5 tahapan untuk anomali temporal (duplikasi, penyisipan, dan penghapusan frame). Sistem dievaluasi menggunakan kerangka kerja Risk-Based Testing pada 28 kasus uji gambar dan 23 video. Hasil pengujian menunjukkan sistem forensik gambar mencapai tingkat keberhasilan 100% dengan rata-rata keandalan 92,9%. Sistem forensik video mencapai tingkat keberhasilan 91,3%, menunjukkan ketahanan tinggi terhadap anomali halus. Kesimpulannya, pendekatan multi-metode terintegrasi ini memberikan solusi yang sangat andal dan efektif untuk verifikasi keaslian multimedia dalam investigasi hukum. Penelitian selanjutnya disarankan mengintegrasikan deep learning untuk mendeteksi deepfake.
Rapid advancements in digital media editing enable highly realistic image and video manipulations, threatening digital evidence integrity. Manual verification and single-method detection have fundamental limitations, such as high subjectivity and false-positive rates. Thus, developing an automated, multi-method forensic system is crucially needed. This research proposes a dual-pipeline forensic system integrating K-Means Clustering and Localization Tampering, supported by Error Level Analysis (ELA) and Scale-Invariant Feature Transform (SIFT). The image pipeline utilizes 19 processing stages to detect spatial manipulations (copy-move and splicing), while the video pipeline uses 5 stages to detect temporal anomalies (frame duplication, insertion, and deletion). The system was evaluated using a Risk-Based Testing framework on 28 image and 23 video test cases. Testing results showed the image forensic system achieved a 100% success rate with 92.9% average forensic reliability. The video forensic system achieved a 91.3% success rate, demonstrating high robustness against subtle anomalies. In conclusion, this integrated multi-method approach provides a highly reliable and effective solution for multimedia authenticity verification in legal investigations. Future work should integrate deep learning to detect deepfakes.</dc:description>
	<dc:publisher xml:lang="en-US">Informatics Engineering Department of Halu Oleo University</dc:publisher>
	<dc:date>2026-06-30</dc:date>
	<dc:type>info:eu-repo/semantics/article</dc:type>
	<dc:type>info:eu-repo/semantics/publishedVersion</dc:type>
	<dc:type xml:lang="en-US">Peer-reviewed Article</dc:type>
	<dc:format>application/pdf</dc:format>
	<dc:identifier>https://semantik.uho.ac.id/index.php/journal/article/view/265</dc:identifier>
	<dc:identifier>10.55679/semantik.v12i1.265</dc:identifier>
	<dc:source xml:lang="en-US">SemanTIK : Teknik Informasi; Vol. 12 No. 1 (2026): Vol. 12 No. 1 (2026): SemanTIK Teknik Informasi; 204-213</dc:source>
	<dc:source>2502-8928</dc:source>
	<dc:source>2460-1446</dc:source>
	<dc:language>eng</dc:language>
	<dc:relation>https://semantik.uho.ac.id/index.php/journal/article/view/265/78</dc:relation>
	<dc:rights xml:lang="en-US">Copyright (c) 2026 Agrian Wahab, Isnawaty, LM. Fid Aksara, Muh Yamin Aksara, LM. Bahtiar Aksara, Subardin</dc:rights>
	<dc:rights xml:lang="en-US">https://creativecommons.org/licenses/by-sa/4.0</dc:rights>
</oai_dc:dc>