EYE DISEASE DETECTION USING FUNDUS IMAGES WITH CONVOLUTIONAL NEURAL NETWORK METHOD ON A WEB-BASED PLATFORM

MUHAMMAD, FATIH MAULANA (2025) EYE DISEASE DETECTION USING FUNDUS IMAGES WITH CONVOLUTIONAL NEURAL NETWORK METHOD ON A WEB-BASED PLATFORM. Tugas Akhir thesis, Informatics.

Text 5210411279_FatihMaulanaMuhammad_ABSTRAK.pdf Download (19kB)

Abstract

Eye disease diagnosis is typically performed by ophthalmologists. However, the high cost of specialist consultations, the time-consuming nature of diagnosis, and its occasional lack of accuracy present significant challenges. These barriers often prevent patients from undergoing early screening, which is a crucial step in preventing severe eye conditions. This study aims to develop an early detection system for eye diseases using fundus images. Medical data used in this research were obtained during an internship at PT. Stechoq Robotika Indonesia, consisting of 3,300 images categorized into three classes: Glaucoma, Diabetic Retinopathy, and Normal. Each class was further divided into training (1,100 images), testing (300 images), and validation (300 images) sets. Data preprocessing involved noise removal and labelling using bounding boxes to localize affected areas. The system was developed using the Convolutional Neural Network (CNN) approach, specifically leveraging the YOLO (You Only Look Once) architecture, version 11 by Ultralytics. Tools such as GitHub and Kaggle were used for data management, while Python served as the programming language. After model training, the system was deployed using Streamlet, enabling web-based access by connecting the model from GitHub. The model achieved an accuracy rate between 70% and 75%, indicating a satisfactory performance for an initial-stage prototype of an automated eye disease detection system. Keywords: Diabetic Retinopathy, Glaucoma, Convolutional Neural Network, YOLO, Streamlet

Item Type:	Thesis (Skripsi, Tugas Akhir or Kerja Praktek) (Tugas Akhir)
Subjects:	T Technology > T Technology (General)
Divisions:	Fakultas Sains Dan Teknologi > S1 Informatika
Depositing User:	Kaprodi S1 Informatika UTY
Date Deposited:	17 Jul 2025 01:17
Last Modified:	17 Jul 2025 01:17
URI:	http://eprints.uty.ac.id/id/eprint/18187

Actions (login required)

View Item