Deteksi Diabetic Eye Diseases dengan Arsitektur VGG19 Berbasis Transfer Learning

Kahfisin, Hafiz Ibnu (2025) Deteksi Diabetic Eye Diseases dengan Arsitektur VGG19 Berbasis Transfer Learning. Other thesis, Institut Teknologi Sepuluh Nopember.

Text 5002201090-Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only until 1 April 2027. Download (1MB) | Request a copy

Abstract

Diabetes merupakan gangguan metabolisme kompleks yang memiliki ciri-ciri tingginya gula darah dikarenakan kurangnya produksi insulin atau tubuh kurang maksimal dalam menggunakan insulin yang diproduksi tubuh. Secara umum, pasien yang menderita diabetes akan mengalami komplikasi. Salah satu komplikasi yang umum adalah diabetic eye diseases. Diabetic eye diseases merupakan kumpulan beberapa penyakit mata yang disebabkan oleh penyakit diabetes. Terdapat 4 penyakit yang termasuk dalam diabetic eye diseases yaitu diabetic retinopathy, diabetic macular edema, glaukoma, dan katarak. Pada 2019, 1,5 juta orang mengalami kematian diakibatkan diabetes, dan sebanyak 48% terjadi pada seseorang berusia dibawah 70 tahun. Oleh karena itu, dibutuhkan peran teknologi untuk mempercepat proses diagnosis diabetic eye disease oleh tenaga medis agar lebih efisien. Penelitian ini bertujuan untuk mendeteksi penyakit diabetic eye diseases menggunakan model deep learning dengan teknik transfer learning untuk memudahkan diagnosis tenaga medis. Penggunaan arsitektur VGG19 dimaksudkan untuk mengenali pola pada fundus mata pasien penderita diabetic eye diseases dan transfer learning digunakan untuk mempercepat proses pelatihan model. Terdapat tiga tahapan utama dalam penelitian ini. Tahapan pertama pada penelitian yaitu preprocessing dataset citra fundus mata, tahapan kedua yakni membuat dan memodifikasi model VGG19, dan tahapan terakhir adalah melakukan pengujian model. Berdasarkan hasil pengujian model VGG19 berbasis transfer learning dengan preprocessing, didapatkan hasil akurasi, precision, recall, dan F1-score deteksi diabetic eye diseases sebesar 0.9313, 0.9321, 0.9313, dan 0.9314 secara berurutan.
==================================================================================================================================
Diabetes is a complex metabolic disorder characterized by high blood sugar levels caused by insufficient insulin production or the body’s inability to effectively use the insulin produced. Generally, patients with diabetes are prone to complications, one of the most common being diabetic eye diseases. Diabetic eye diseases refer to a group of eye conditions caused by diabetes. These include four major conditions: diabetic retinopathy, diabetic macular edema, glaucoma, and cataracts. In 2019, 1.5 million people died due to diabetes, with 48% of these deaths occurring in individuals under the age of 70. To address this, technology plays a crucial role in expediting the diagnosis process of diabetic eye diseases for medical professionals to improve efficiency. This study aims to detect diabetic eye diseases using a deep learning model with transfer learning techniques to assist medical diagnoses. The VGG19 architecture is utilized to recognize patterns in the fundus images of patients with diabetic eye diseases, and transfer learning is employed to accelerate the model training process. This study involves three main stages. The first stage is preprocessing the fundus image dataset, the second is creating and modifying the VGG19 model, and the final stage is testing the model. Based on the test results of the VGG19 model with transfer learning and preprocessing, the accuracy, precision, recall, and F1-score for detecting diabetic eye diseases were 0.9313, 0.9321, 0.9313, and 0.9314 respectively.

Item Type:	Thesis (Other)
Uncontrolled Keywords:	Diabetic Eye Diseases, Fundus Mata, VGG19, Transfer Learning, Eye Fundus
Subjects:	Q Science > QA Mathematics > QA336 Artificial Intelligence Q Science > QA Mathematics > QA76.87 Neural networks (Computer Science)
Divisions:	Faculty of Science and Data Analytics (SCIENTICS) > Mathematics > 44201-(S1) Undergraduate Thesis
Depositing User:	Hafiz Ibnu Kahfisin
Date Deposited:	30 Jan 2025 00:49
Last Modified:	30 Jan 2025 00:49
URI:	http://repository.its.ac.id/id/eprint/116975

Actions (login required)

View Item