Segmentation of Foveal Avascular Zone in Colour Fundus Images Based on Retinal Capillary Endpoints Detection

Authors

  • Hanung Adi Nugroho Department of Electrical Engineering and Information Technology, Faculty of Engineering, Universitas Gadjah Mada, Jalan Grafika 2, Kampus UGM, Yogyakarta 55281, Indonesia.
  • Dewi Purnamasari Department of Electrical Engineering and Information Technology, Faculty of Engineering, Universitas Gadjah Mada, Jalan Grafika 2, Kampus UGM, Yogyakarta 55281, Indonesia.
  • Indah Soesanti Department of Electrical Engineering and Information Technology, Faculty of Engineering, Universitas Gadjah Mada, Jalan Grafika 2, Kampus UGM, Yogyakarta 55281, Indonesia.
  • Widhia K.Z Oktoeberza Department of Electrical Engineering and Information Technology, Faculty of Engineering, Universitas Gadjah Mada, Jalan Grafika 2, Kampus UGM, Yogyakarta 55281, Indonesia.
  • Dhimas Arief Dharmawan Department of Electrical Engineering and Information Technology, Faculty of Engineering, Universitas Gadjah Mada, Jalan Grafika 2, Kampus UGM, Yogyakarta 55281, Indonesia.

Keywords:

Foveal Avascular Zone, Local Entropy Thresholding, Matched Filter, Retinal Capillary,

Abstract

Diabetic retinopathy (DR) is one of the diabetes complications affecting the retina. It can be detected by investigating foveal avascular zone (FAZ) since there is a correlation between enlargement of FAZ and DR progression. In this research work, a method of FAZ detection is developed. Firstly, pre-processing is conducted to enhance and improve image quality. Afterwards, segmentation of FAZ is conducted using matched filter and local entropy thresholding to extract retinal vessels. FAZ area segmentation is done based on retinal capillary endpoints detection. This work is validated using retinal fundus images from Messidor and DRIVE databases. The result of FAZ segmentation has been verified by measuring the correlation coefficient of determined FAZ areas between the capillary endpoints of the proposed method and that of detected by ophthalmologists. The correlation values achieved are 0.912 and 0.802 for two aforementioned databases, respectively. These results indicate that the proposed method has successfully detected and segmented FAZ area, due to the highly significant correlation coefficient obtained between the proposed FAZ and that of the ophthalmologists.

Downloads

Published

2017-11-30

How to Cite

Nugroho, H. A., Purnamasari, D., Soesanti, I., Oktoeberza, W. K., & Dharmawan, D. A. (2017). Segmentation of Foveal Avascular Zone in Colour Fundus Images Based on Retinal Capillary Endpoints Detection. Journal of Telecommunication, Electronic and Computer Engineering (JTEC), 9(3-8), 107–112. Retrieved from https://jtec.utem.edu.my/jtec/article/view/3107