Nurazis, Ikhsan and , Diah Priyawati, S.T., M.Eng. (2024) Ekstraksi Fitur Citra Retina Dengan Resnet50 dan Teknik Ensemble Untuk Klasifikasi Gangguan Penglihatan Dari Citra Fundus Retina. Skripsi thesis, Universitas Muhammadiyah Surakarta.
![[img]](https://eprints.ums.ac.id/style/images/fileicons/application_pdf.png) |
PDF (Naskah Publikasi,)
Naskah_Skripsi_EkstraksiFiturCitraRetina.pdf Download (458kB) |
|
PDF (Surat Pernyataan Publikasi)
UNGGAH Nurazis.pdf Restricted to Repository staff only Download (284kB) | Request a copy |
Abstract
Visual impairments such as glaucoma, cataracts, and diabetic retinopathy are significant health issues that require early detection for more effective prevention. Early detection can be achieved by leveraging machine learning technology, which can analyze retinal images automatically and in depth. Using a deep learning method based on the ResNet50 architecture combined with ensemble techniques, this system can recognize specific patterns and characteristics of each eye condition, including glaucoma, cataracts, and diabetic retinopathy. This study utilizes four types of retinal images as the dataset: normal, cataract, glaucoma, and diabetic retinopathy. During the process, features are extracted using ResNet50 and subsequently reduced in dimensionality using Principal Component Analysis (PCA). Classification is then performed by combining Random Forest, Support Vector Machine (SVM), and XGBoost algorithms through a Voting Classifier technique to enhance model accuracy. The results indicate that this approach successfully achieves an accuracy of 91% with an Area Under Curve (AUC) value of 0.99 on the ROC curve, signifying that this model has a solid potential to automatically detect eye disorders, thereby supporting early diagnosis and improving the quality of eye healthcare.
| Item Type: | Thesis (Skripsi) |
|---|---|
| Uncontrolled Keywords: | Ensemble learning, Retinal image classification, Principal Component Analysis (PCA), ResNet50, Voting Classifier. |
| Subjects: | T Technology > TZ Technical Information |
| Divisions: | Fakultas Ilmu Komunikasi dan Informatika > Teknik Informatika |
| Depositing User: | IKHSAN NURAZIS |
| Date Deposited: | 19 Nov 2024 08:54 |
| Last Modified: | 19 Nov 2024 08:54 |
| URI: | http://eprints.ums.ac.id/id/eprint/129030 |
Actions (login required)
 |
View Item |
