Comparative Study of Iris and Retinal Images for Early Detection of Diabetic Mellitus
The recent increase in the number of diabetic cases due to genetic reasons or sedentary lifestyle, necessitates urgent need for an effective glucose monitoring system. Certainly, periodic glucose level monitoring in the blood will prevent from entering chronic diabetic condition and
a noninvasive monitoring tool leads to a simple and automated diagnosis procedure. In this present work, an iridology-based diagnosis of diabetes has been discussed and is compared with a standard retinal imaging modality. Two subject groups, one group of 30 subjects without diabetes, the
other group with 20 subjects of controlled diabetes with less than two years duration and 25 subjects with more than two years of uncontrolled diabetes were evaluated. Iris images are acquired using an iriscope and subsequently compared it with that of retinal spectral domain optical coherence
tomography (SDOCT) images of the same subjects. The segmentation of the pancreas region in the iris images and the retinal layers in the SDOCT retinal images are performed automatically to predict Diabetic Mellitus (DM). The selected features of the segmented region are given as input to k-Nearest
Neighbor, Support Vector Machine and Random Forest classifiers for discriminating diabetic and non-diabetic normal cases. The results showed that iris images are able to reveal the diabetic condition of the subjects even before their retina could reveal the same. However, elaborate extensive
studies have to be carried out in order to validate iridology as a best noninvasive tool for detection of diabetes at the early stage itself.
Keywords: IRIDOLOGY; RANDOM FOREST CLASSIFIER; SDOCT RETINAL IMAGE; SUPPORT VECTOR MACHINE; k-NEAREST NEIGHBOR CLASSIFIER
Document Type: Research Article
Publication date: 01 February 2020
- Journal of Medical Imaging and Health Informatics (JMIHI) is a medium to disseminate novel experimental and theoretical research results in the field of biomedicine, biology, clinical, rehabilitation engineering, medical image processing, bio-computing, D2H2, and other health related areas.
- Editorial Board
- Information for Authors
- Subscribe to this Title
- Ingenta Connect is not responsible for the content or availability of external websites
- Access Key
- Free content
- Partial Free content
- New content
- Open access content
- Partial Open access content
- Subscribed content
- Partial Subscribed content
- Free trial content