KLASIFIKASI PENYAKIT MATA PADA CITRA RETINA MENGGUNAKAN CONVOLUTIONAL NEURAL NETWORK DAN GRAD-CAM

MUHAMMAD, DAMAR FADHIL and Rini, Dian Palupi and Rachmatullah, Muhammad Naufal (2024) KLASIFIKASI PENYAKIT MATA PADA CITRA RETINA MENGGUNAKAN CONVOLUTIONAL NEURAL NETWORK DAN GRAD-CAM. Undergraduate thesis, Sriwijaya University.

	Text RAMA_55201_09021182025019.pdf - Accepted Version Restricted to Repository staff only Available under License Creative Commons Public Domain Dedication. Download (4MB) | Request a copy
	Text RAMA_55201_09021182025019_TURNITIN.pdf - Accepted Version Restricted to Repository staff only Available under License Creative Commons Public Domain Dedication. Download (9MB) | Request a copy
	Text RAMA_55201_09021182025019_0023027804_0001129204_01_front_ref.pdf - Accepted Version Available under License Creative Commons Public Domain Dedication. Download (2MB)
	Text RAMA_55201_09021182025019_0023027804_0001129204_02.pdf - Accepted Version Restricted to Repository staff only Available under License Creative Commons Public Domain Dedication. Download (613kB) | Request a copy
	Text RAMA_55201_09021182025019_0023027804_0001129204_03.pdf - Accepted Version Restricted to Repository staff only Available under License Creative Commons Public Domain Dedication. Download (451kB) | Request a copy
	Text RAMA_55201_09021182025019_0023027804_0001129204_04.pdf - Accepted Version Restricted to Repository staff only Available under License Creative Commons Public Domain Dedication. Download (577kB) | Request a copy
	Text RAMA_55201_09021182025019_0023027804_0001129204_05.pdf - Accepted Version Restricted to Repository staff only Available under License Creative Commons Public Domain Dedication. Download (898kB) | Request a copy
	Text RAMA_55201_09021182025019_0023027804_0001129204_06.pdf - Accepted Version Restricted to Repository staff only Available under License Creative Commons Public Domain Dedication. Download (236kB) | Request a copy
	Text RAMA_55201_09021182025019_0023027804_0001129204_07_ref.pdf - Bibliography Restricted to Repository staff only Available under License Creative Commons Public Domain Dedication. Download (254kB) | Request a copy
	Text RAMA_55201_09021182025019_0023027804_0001129204_08_lamp.pdf - Accepted Version Restricted to Repository staff only Available under License Creative Commons Public Domain Dedication. Download (204kB) | Request a copy

Abstract

Eye diseases can be detected by examining the retina of the eye. Therefore, this research develops a system using Convolutional Neural Network (CNN) and Gradient-Weighted Class Activation Map (Grad-CAM) methods that can help diagnose eye diseases based on retinal images. The CNN method is used to classify whether there is a disease or not, so that the system can diagnose the disease suffered by the patient. However, this method has the disadvantage that it cannot display a visual explanation of the classification results, to cover this deficiency this method is combined with Grad-CAM. Grad-CAM can provide a visual explanation of the classification results in the form of a heatmap, so that users of this system can understand the reasons behind the CNN method classifying to a certain class. This research compares the architecture of InceptionV3, MobileNetV2, VGG-16, and various configurations on epoch, learning rate, and batch size in building the best CNN model. The dataset used in this study consists of 4 classes and totals 4217 data. The test results in this study produced the best CNN model using InceptionV3 architecture, epoch = 50, learning rate = 0,0001, and batch size = 8 with an accuracy value of 96,3%.

Item Type:	Thesis (Undergraduate)
Uncontrolled Keywords:	Penyakit Mata, Citra Retina, Convolutional Neural Network, Grad-CAM
Subjects:	R Medicine > RE Ophthalmology > RE75-79 Examination. Diagnosis
Divisions:	09-Faculty of Computer Science > 55201-Informatics (S1)
Depositing User:	Damar Fadhil Muhammad
Date Deposited:	15 Jul 2024 07:46
Last Modified:	15 Jul 2024 07:46
URI:	http://repository.unsri.ac.id/id/eprint/151006

Actions (login required)

View Item