RDCNET: Convolutional neural networks for classification of retinopathy disease in unbalanced data cases

Bambang Krismono Triwijoyo, Boy Subirosa Sabarguna, Widodo Budiharto, Edi Abdurachman

Research output: Contribution to journalArticle

Abstract

Retinopathy disease is a type of retinal disorder, which often occurs, includ- ing hypertensive retinopathy and diabetic hypertension. Detection of retinopathy can be by analyzing the retinal image, using a deep learning approach, but the problem that is often faced is unbalanced data. In this study, a convolutional neural network architec- ture is proposed for the classification of retinopathy using the MESSIDOR database that has been labeled, by duplicating and augmentation of sample images in classes with low numbers of samples using a data generator to overcome the problem of unbalanced data. The experimental results show that the validation and testing accuracy performance on the model with two output classes are 100%, and 87.50%, while on the model with four output classes are 99.38%, and 76.47%. ICIC International

Original languageEnglish
Pages (from-to)635-641
Number of pages7
JournalICIC Express Letters
Volume14
Issue number7
DOIs
Publication statusPublished - Jul 2020

Keywords

  • Convolutional neural network
  • Deep learning
  • Image classi_cation
  • Retinopathy diseases
  • Unbalanced data

Fingerprint Dive into the research topics of 'RDCNET: Convolutional neural networks for classification of retinopathy disease in unbalanced data cases'. Together they form a unique fingerprint.

  • Cite this