Authors: M. Tamilarasi; K. Duraiswamy
Addresses: Department of Computer Science and Engineering, K.S.R. College of Engineering, Tiruchengode, Tamilnadu, India ' Department of Computer Science and Engineering, K.S.R. College of Engineering, Tiruchengode, Tamilnadu, India
Abstract: Diabetic retinopathy is one of the most common causes of blindness in the working-age population. Screening for diabetic retinopathy represents a good clinical practice and cost-effective healthcare. Early detection and treatment of retinopathy has been shown to be effective in preventing visual impairment. This paper provides an automated system for the identification of exudates for early diagnosis of diabetic retinopathy. The grey scale version of colour retinal fundus image was first produced. It was then enhanced using pre-processing techniques. The template matching (TM) algorithm was proposed to segment the exudates regions from diabetic retinopathy retinal images. This was experimented with a dataset, DIARETDB0 which consists of 130 colour retinal fundus images. The sensitivity and specificity achieved were 99.45% and 95.68%, respectively. It had shown 98.72% accuracy.
Keywords: decision support systems; DSS; image analysis; template matching; edge detection; image segmentation; connected components; diabetic retinopathy; retinal exudates; retinal fundus images; retinal images; diabetes; blindness; retinopathy screening; visual impairment; retinopathy diagnosis.
International Journal of Computer Applications in Technology, 2013 Vol.48 No.2, pp.136 - 143
Received: 08 May 2021
Accepted: 12 May 2021
Published online: 24 Aug 2013 *