Diabetic Retinopathy: Identification and Classification using Different Kernel on Support Vector Machine

Ahmad Zoebad Foeady, Dian Candra Rini Novitasari, Ahmad Hanif Asyhar

2018

Abstract

Diabetic Retinopathy (DR) is one complication of diabetes that characterized by high glucose levels in the eyes that ultimately lead to blindness. To minimize the occurrence of blindness from DR, required diagnosis on the eye to possible for early treatment. In this paper, identified and classified DR using Gray Level Co-occurrence Matrix (GLCM) as feature extraction and Multiclass SVM with different kernel functions. The purpose of this study is to provide a breakthrough for patients in diagnosing the severity of the DR. The components identified in DR images include blood vessels, microaneurysms, and hemorrhages with contrast, energy, correlation, and homogeneity as feature extraction data on the GLCM method. The feature data will be classified using the Multiclass SVM method with 4 different kernel functions such as quadratic, linear, gaussian, and polynomial. The feature data will be classified using the Multiclass SVM method with 4 different kernel functions. Identification and classification of the DR image have an accuracy from each of quadratic, linear, Gaussian, and polynomial kernels functions are 72.72%, 22.72%, 63.64%, and 90.91%. From that accuracy, it has seen polynomial kernel function is more suitable for DR classification.

Paper Citation

in Harvard Style

Foeady A., Novitasari D. and Asyhar A. (2018). Diabetic Retinopathy: Identification and Classification using Different Kernel on Support Vector Machine.In Proceedings of the International Conference on Mathematics and Islam - Volume 1: ICMIs, ISBN 978-989-758-407-7, pages 72-79. DOI: 10.5220/0008517400720079

in Bibtex Style

@conference{icmis18,
author={Ahmad Zoebad Foeady and Dian Candra Rini Novitasari and Ahmad Hanif Asyhar},
title={Diabetic Retinopathy: Identification and Classification using Different Kernel on Support Vector Machine},
booktitle={Proceedings of the International Conference on Mathematics and Islam - Volume 1: ICMIs,},
year={2018},
pages={72-79},
publisher={SciTePress},
organization={INSTICC},
doi={10.5220/0008517400720079},
isbn={978-989-758-407-7},
}

in EndNote Style

TY - CONF

JO - Proceedings of the International Conference on Mathematics and Islam - Volume 1: ICMIs,
TI - Diabetic Retinopathy: Identification and Classification using Different Kernel on Support Vector Machine
SN - 978-989-758-407-7
AU - Foeady A.
AU - Novitasari D.
AU - Asyhar A.
PY - 2018
SP - 72
EP - 79
DO - 10.5220/0008517400720079