IMPLEMENTASI DEEP LEARNING UNTUK KLASIFIKASI PENYAKIT MATA DENGAN STUDI PERBANDINGAN ARSITEKTUR INCEPTIONV3 DAN VGG-16 PADA CONVOLUTIONAL NEURAL NETWORK (CNN)

ABDILLAH, MUHAMMAD ARIF and Sutarno, Sutarno (2025) IMPLEMENTASI DEEP LEARNING UNTUK KLASIFIKASI PENYAKIT MATA DENGAN STUDI PERBANDINGAN ARSITEKTUR INCEPTIONV3 DAN VGG-16 PADA CONVOLUTIONAL NEURAL NETWORK (CNN). Undergraduate thesis, Sriwijaya University.

Preview	Image RAMA_56201_09011182126008_cover.jpg - Cover Image Available under License Creative Commons Public Domain Dedication. Download (550kB) | Preview
	Text RAMA_56201_09011182126008.pdf - Accepted Version Restricted to Repository staff only Available under License Creative Commons Public Domain Dedication. Download (2MB) | Request a copy
	Text RAMA_56201_09011182126008_TURNITIN.pdf - Accepted Version Restricted to Repository staff only Available under License Creative Commons Public Domain Dedication. Download (7MB) | Request a copy
	Text RAMA_56201_09011182126008_0201117802_01_front_ref.pdf - Accepted Version Available under License Creative Commons Public Domain Dedication. Download (787kB)
	Text RAMA_56201_09011182126008_0201117802_02.pdf - Accepted Version Restricted to Repository staff only Available under License Creative Commons Public Domain Dedication. Download (918kB) | Request a copy
	Text RAMA_56201_09011182126008_0201117802_03.pdf - Accepted Version Restricted to Repository staff only Available under License Creative Commons Public Domain Dedication. Download (301kB) | Request a copy
	Text RAMA_56201_09011182126008_0201117802_04.pdf - Accepted Version Restricted to Repository staff only Available under License Creative Commons Public Domain Dedication. Download (856kB) | Request a copy
	Text RAMA_56201_09011182126008_0201117802_05.pdf - Accepted Version Restricted to Repository staff only Available under License Creative Commons Public Domain Dedication. Download (127kB) | Request a copy
	Text RAMA_56201_09011182126008_0201117802_06_ref.pdf - Bibliography Restricted to Repository staff only Available under License Creative Commons Public Domain Dedication. Download (152kB) | Request a copy
	Text RAMA_56201_09011182126008_0201117802_07_lamp.pdf - Accepted Version Restricted to Repository staff only Available under License Creative Commons Public Domain Dedication. Download (395kB) | Request a copy

Abstract

Retinal image based eye disease classification is an important approach to support automated diagnosis in the medical field. This study aims to classify four types of eye diseases normal, cataract, glaucoma, and diabetic retinopathy using Convolutional Neural Network (CNN) models with the InceptionV3 and VGG-16 architectures. The dataset was obtained from the Kaggle platform, consisting of 4,217 images, and was augmented to 11,531 images using techniques such as flipping, zooming, and scaling. The preprocessing stage included normalization adapted to each architecture and data splitting into training, validation, and testing subsets. The models were evaluated based on varying hyperparameters, including the number of epochs, batch sizes, and learning rates. The results show that the bestperforming model using the InceptionV3 architecture (100 epochs, batch size 64, learning rate 0.001) achieved a testing accuracy of 98.6%, with precision, recall, and F1-score all reaching 0.99. Meanwhile, the best VGG-16 model achieved a maximum accuracy of 89.8% with an F1-score of 0.90. In conclusion, the InceptionV3 architecture outperforms VGG-16 in classifying retinal images, and the selection of appropriate hyperparameters significantly influences the final model performance.

Item Type:	Thesis (Undergraduate)
Uncontrolled Keywords:	Klasifikasi Penyakit Mata, CNN, InceptionV3, VGG-16, Deep Learning, Citra Retina
Subjects:	Q Science > Q Science (General) > Q334-342 Computer science. Artificial intelligence. Algorithms. Robotics. Automation.
Divisions:	09-Faculty of Computer Science > 56201-Computer Systems (S1)
Depositing User:	Muhammad Arif Abdillah
Date Deposited:	01 Sep 2025 01:37
Last Modified:	01 Sep 2025 01:37
URI:	http://repository.unsri.ac.id/id/eprint/183471

Actions (login required)

View Item