Skip to content
SDGs Center UDINUS

Evidence record · UDINUS · SDGs

Filter Gabor and SMOTE Method-Based Convolutional Neural Network for Diabetic Retinopathy Classification

Rektorat UDINUS 2024 Published 2026-10-30

Evidence record · UDINUS · SDGs

Unit
Rektorat UDINUS
Period
2024
Goals
Record
6a448195-0a00-4516-826d-91391345221a
Published
2026-10-30

Universitas Dian Nuswantoro conducted research on the classification of diabetic retinopathy using retinal fundus images. The activity involved Mulyani, Asri; Muljono; Purwanto; and Soeleman, Moch Arief as participants. Diabetic retinopathy is described as a disorder of the retina that occurs in people with diabetes, with an increasing risk associated with age. The research addressed the need for treatment and early detection of eye fundus images to prevent the worst risk of blindness.

The activity focused on improving the performance of an image-classification model for diabetic retinopathy retinal fundus images. The method employed deep learning, described as a problem-solving method for computer learning systems that uses the concept of hierarchy. A convolutional neural network was used as a deep learning algorithm for detection or identification problems. The research also applied a Gabor Filter to extract features from retinal fundus images and SMOTE to overcome unbalanced datasets.

The study results showed that the use of Gabor and SMOTE filters significantly increased the accuracy value from 40% to 78%. The reported increase in performance was influenced by preprocessing procedures, namely image resizing, conversion to grayscale, and the setting of Gabor filter parameters. Through this activity, UDINUS addressed a health-related research topic involving diabetic retinopathy identification and the processing of retinal fundus images. The activity provides institutional evidence of research using image-classification methods to examine a health-related condition.

Reference https://www.scopus.com/pages/publications/85213339424?origin=resultslist