Iris Diagnosis – A Quantitative Non-Invasive Tool for Diabetes Detection

Journal: GRENZE International Journal of Computer Theory and Engineering
Authors: Mithun B S, R Sneha, Vinay Raj K, Basavaraj Hiremath, B. Ragavendrasamy
Volume: 3 Issue: 4
Grenze ID: 01.GIJCTE.3.4.78 Pages: 526-532

Abstract

Iris Diagnosis is a novel approach to assess various pathological conditions based on the Iris patterns. Iris Image analysis is a non-invasive technique for determining the health condition of an individual. Correct and timely diagnosis is critical, yet is the absolute requirement of medical science. In general, current approaches fail to diagnose various diseases correctly. An attempt is being made in the current research to explore the possibility of diagnosing diabetes from the representations of the Iris. Initially the images of eye are captured using Iridoscope and a database is created with their clinical history. Various algorithms are developed to assess the quality of the Iris image, and then segmentation and feature extraction techniques such as GLCM and DCT are applied. Feature extraction plays a vital role in assessment of the individual to be diabetic or not. In order to assess the presently proposed approach, 30 patient data were acquired for which the present approach was able to detect diabetic or not with an accuracy of 83%.

Download Now << BACK

GIJCTE