Abstract
AbstractDiabetic retinopathy (DR) is one of the leading causes of visual loss if it is not treated at an earlier stage. Manual identification of diabetic retinopathy is a time-consuming process, and regular screening is a must for an early diagnosis. This paper presented a novel blood vessel parameter extraction method for DR identification using image processing and data mining techniques. An automatic DR diagnosis through image processing techniques, by extraction of blood vessel parameters such as vessel density, minimum and maximum thickness of blood vessels and classification through data mining techniques was proposed. Mostly diabetic retinopathy identification was done by lesion pattern identification such as exudates, microaneurysms, cotton wool spots, etc. However, this work concentrated on calculating the disease parameter through segmented blood vessel region from full fundus image, optic nerve head region and Inferior, Superior, Nasal and Temporal (ISNT) region. Evaluation of this work was performed on DRIVE and HRF datasets and achieved overall accuracy of 97.14% in terms of DR prediction. KeywordsOptic nerve headISNT regionDiabetic retinopathyBlood vessel
Published Version
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call