Automated Detection Of Exudates Using DBSCAN Clustering And Fuzzy Classifier

Shantala Giraddi, Dr.Jagadeesh Pujari

Abstract


Diabetic Retinopathy is a major cause of vision loss for diabetic patients, but early detection of its symptoms and treatment can prevent blindness. Exudates are the key indicators of diabetic retinopathy that can potentially be automatically quantified. In this paper the authors have attempted to detect exudates by a combination of DBSCAN clustering algorithm and Fuzzy classifiers. The DBSCAN algorithm produces many clusters that human cannot make out. In order to correctly identify exudates, Post processing is performed using fuzzy classifier to classify clusters as exudates or non-exudates. Exudates in training retinal images are marked by expert ophthalmologists. Various histogram based features are calculated for the regions marked. These features are used for training the Fuzzy classifiers. Optic disc is localized by the Circular Hough Transform. The publicly available diabetic retinopathy data set DIARETDB0 is used for evaluation .In addition to the above set; images from VASAN Eye Care Hospital (Reputed local Eye care centre) have been used. Our proposed algorithm achieved image based classification accuracy above 90%.

Keywords: Optic disc, Exudates, Diabetic Retinopathy, DBSCAN clustering, Fuzzy classifiers


Full Text:

PDF


DOI: https://doi.org/10.26483/ijarcs.v3i7.1440

Refbacks

  • There are currently no refbacks.




Copyright (c) 2016 International Journal of Advanced Research in Computer Science