Penerapan algoritma DenseNet dalam mendeteksi foto asli dan foto palsu

Adzkia, Muhammad Fasya (2025) Penerapan algoritma DenseNet dalam mendeteksi foto asli dan foto palsu. Sarjana thesis, UIN Sunan Gunung Djati Bandung.

	Text 1_cover.pdf Download (254kB)
	Text 2_abstrak.pdf Download (254kB)
	Text Surat pernyataan karya sendiri (22).pdf Download (229kB)
	Text 4_daftarisi.pdf Download (274kB)
	Text 5_bab1 (19).pdf Download (328kB)
	Text 6_bab2.pdf Restricted to Registered users only Download (523kB) | Request a copy
	Text 7_bab3.pdf Restricted to Registered users only Download (467kB) | Request a copy
	Text 8_bab4.pdf Restricted to Registered users only Download (704kB) | Request a copy
	Text 9_bab5.pdf Restricted to Registered users only Download (274kB) | Request a copy
	Text 10_daftarpustaka.pdf Restricted to Registered users only Download (199kB) | Request a copy
	Text 11_lampiran.pdf Restricted to Repository staff only Download (4MB) | Request a copy

Abstract

ENGLISH: The advancement of digital manipulation technologies such as Generative Adversarial Networks (GANs) has introduced significant threats to the authenticity of visual content, particularly in images and photographs. This research implements the Dense Convolutional Network (DenseNet) algorithm to detect and distinguish between real and fake images. DenseNet is chosen due to its ability to enhance feature propagation and model efficiency through dense layer connections. The dataset used in this study consists of a combination of real and GAN-generated fake images. After undergoing preprocessing, the data was trained using a DenseNet-based architecture and evaluated using performance metrics including Accuracy, Precision, Recall, and F1-Score. Experimental results show that the DenseNet model achieved accuracy 81%, with precision 82%, recall 81%, and F1-score 81 % .These results demonstrate the effectiveness of DenseNet in detecting fake images and highlight its potential as a reliable solution for digital content verification and visual security systems. Keywords: DenseNet, Fake Image Detection, GAN, Real Image INDONESIA: Perkembangan teknologi kecerdasan buatan (Artificial Intelligence) telah mendorong berbagai metode baru dalam memanipulasi media digital, salah satunya adalah penggunaan tools yang memanfaatkan Generative Adversarial Network (GAN) yang mampu menghasilkan foto palsu yang sangat mirip dengan foto asli. Fenomena ini menyebabkan keresahan dimasyarakat terhadap foto yang beredar di internet apakah asli atau palsu. Penelitian ini bertujuan untuk menerapkan algoritma DenseNet dalam mendeteksi foto asli dan foto palsu. DesnseNet digunakan karena memiliki kemampuan dalam meminimalkan masalah vanishing gradient dan meningkatkan efisiensi fitur melalui koneksi antar-layer yang padat. Dataset yang digunakan teridiri dari foto asli dan foto palsu yang dihasilkan oleh GAN. Hasil dari penelitian menunjukan penggunaan algoritma DenseNet dalam mendeteksi foto asli dan foto palsu memiliki nilai precision 82%, f1-score 81%, recall 81% dan accuracy 81% yang menunjukan bahwa model mampu untuk mendeteksi foto asli dan foto palsu. Kata kunci : DenseNet, Deteksi foto palsu, GAN, Foto Asli

Actions (login required)

View Item