| Home  | About ScienceAsia  | Publication charge  | Advertise with us  | Subscription for printed version  | Contact us  
Editorial Board
Journal Policy
Instructions for Authors
Online submission
Author Login
Reviewer Login
Volume 50 Number 1
Volume 49 Number 6
Volume 49 Number 5
Volume 49S Number 1
Volume 49 Number 4
Volume 49 Number 3
Earlier issues
Volume  Number 

previous article next article

Research articles

ScienceAsia 35 (2009): 80-88 |doi: 10.2306/scienceasia1513-1874.2009.35.080


Automatic exudate detection for diabetic retinopathy screening


Akara Sopharaka,*, Bunyarit Uyyanonvaraa, Sarah Barmanb

 
ABSTRACT:     Exudates are one of the primary signs of diabetic retinopathy which is a main cause of blindness that could be prevented with an early screening process. Pupil dilation is required in the normal screening process but this affects patients' vision. Automatic computerized screening could facilitate the screening process, reduce inspection time, and increase accuracy. In this paper we propose an automatic method to detect exudates from low-contrast digital images of retinopathy patients with non-dilated pupils using a fuzzy c-means (FCM) clustering technique. Preprocessing of contrast enhancement was applied in order to enhance the quality of the input image before four features, namely, intensity, standard deviation on intensity, hue, and number of edge pixels, were selected to supply to the FCM method. The number of required clusters was optimally selected from a quantitative experiment where it was varied from two to eight clusters. The number of cluster optimization was based on sensitivity and specificity which were calculated by comparison of the detected results and hand-drawn ground truths from expert ophthalmologists. The positive predictive value and positive likelihood ratio were also used to evaluate the overall performance of this method. From the result of the subtracted cluster with the number of clusters equalling 2, it was found that the proposed method detected exudates with 92.18% sensitivity and 91.52% sensitivity.

Download PDF

3 Downloads 1279 Views


a Department of Information Technology, Sirindhorn International Institute of Technology, Thammasat University, 131 Moo 5, Tiwanont Road, Bangkadi, Muang, Pathumthani 12000, Thailand
b Faculty of Computing, Information Systems and Mathematics, Kingston University, Penrhyn Road, Kingston upon Thames, Surrey, KT1 2EE, UK

* Corresponding author, E-mail: akara@siit.tu.ac.th

Received 8 Sep 2008, Accepted 12 Feb 2009