Feature extraction and calculating error rate in diabetic retinopathy images using neural network

  Anupam Singh,  Harishchandra Maurya

In Developing countries, Most of the working age populace are affected from eye diseases. Retinal image analysis plays an important role in medical diagnosis. Most common eye Diseases are Diabetic Retinopathy , Glaucoma, Retinal Detachment, Retinal Disorder, Age Related macular degeneration etc are of much significance today. Around 90% of the people are affected by Diabetic Retinopathy. Diabetic Retinopathy (DR) is one of the most common problems that may lead to blindness and it can be prevented or cured if detected at the earlier stage. The presence of DR can be detected by its different symptoms; the most distinctive is the presence of cotton wool and exudates which are bright lesions in the eye. It is necessary to detect the presence of bright lesions. They play a very important role in the accurate detection and classification of cotton wools and exudates. This research proposes a computer aided system which can be used for detection of lesions in the diabetic retinopathy. The research presents various algorithms on fundus retinal images by following steps like preprocessing, lesion detection, features extraction using fuzzy logic and using neural network. Neural network uses Back Propagation Algorithm which is the most efficient algorithm to deduct the results accurately. The developed method is tested on standard benchmark STARE [] database and the performance metrics is calculated.

Keywords- Diabetic Retinopathy, Lesions , Cotton Wools , Exudates

Unique Identification Number - IJEDR1702010
Page Number(s) - 65-68
Pubished in - Volume 5 | Issue 2 | April 2017
DOI (Digital Object Identifier) -   
Publisher - IJEDR (ISSN - 2321-9939)

  Anupam Singh,  Harishchandra Maurya,   "Feature extraction and calculating error rate in diabetic retinopathy images using neural network", International Journal of Engineering Development and Research (IJEDR), ISSN:2321-9939, Volume.5, Issue 2, pp.65-68, April 2017, Available at :http://www.ijedr.org/papers/IJEDR1702010.pdf