Efektivitas RANSAC dan Outlier Detection dalam Mendeteksi Keaslian pada Gambar Foto Produk Digital Hamako Eco Babywear

Anhar Abul Gani (Universitas Kebangsaan Republik Indonesia)
Deni Suprihadi (Universitas Kebangsaan Republik Indonesia)

JuSTISe: Journal Data Science, Technology, Informatics and Security
Vol 2 No 1 (2024): Journal Data Science, Technology, Informatics and Security (Juni 2024)

This study addresses the increasing risk of digital crimes, particularly image forgery, as a result of advancements in information and communication technology. The research focuses on comparing two methods, RANSAC and Outlier Detection, for analyzing the authenticity of digital images related to Hamako Eco Baby Wear products, which potentially violate Intellectual Property Rights (IPR). The case involves the misuse of product logos and attributes. The Integrated Digital Forensic Investigation Framework (IDFIF) is employed as the main framework, supplemented by tools such as the Image Hash Generator and RANSAC Detection. This study also examines metadata from sample and suspect images, providing crucial information about the time and tools used for capturing or editing the images. The findings reveal that the Outlier Detection method is effective in quickly identifying image anomalies, while RANSAC generates a mathematical model that is robust against outliers, enabling deeper analysis. These two methods complement each other in proving image forgery or misuse. This research contributes significantly to the development of digital forensic techniques, particularly in analyzing the authenticity of digital images in the modern era.