Format

Send to

Choose Destination
Comput Med Imaging Graph. 2013 Jul-Sep;37(5-6):394-402. doi: 10.1016/j.compmedimag.2013.05.005. Epub 2013 Jun 15.

Simple hybrid method for fine microaneurysm detection from non-dilated diabetic retinopathy retinal images.

Author information

1
Faculty of Science and Arts, Burapha University, Chanthaburi Campus, 57 Moo 1, Kamong, Thamai, Chanthaburi 22170, Thailand. Electronic address: akara@buu.ac.th.

Abstract

Microaneurysms detection is an important task in computer aided diagnosis of diabetic retinopathy. Microaneurysms are the first clinical sign of diabetic retinopathy, a major cause of vision loss in diabetic patients. Early microaneurysm detection can help reduce the incidence of blindness. Automatic detection of microaneurysms is still an open problem due to their tiny sizes, low contrast and also similarity with blood vessels. It is particularly very difficult to detect fine microaneurysms, especially from non-dilated pupils and that is the goal of this paper. Simple yet effective methods are used. They are coarse segmentation using mathematic morphology and fine segmentation using naive Bayes classifier. A total of 18 microaneurysms features are proposed in this paper and they are extracted for naive Bayes classifier. The detected microaneurysms are validated by comparing at pixel level with ophthalmologists' hand-drawn ground-truth. The sensitivity, specificity, precision and accuracy are 85.68, 99.99, 83.34 and 99.99%, respectively.

KEYWORDS:

Diabetic retinopathy; Microaneurysms; Naive Bayes classifier

[Indexed for MEDLINE]

Supplemental Content

Full text links

Icon for Elsevier Science
Loading ...
Support Center