Diabetes Detection Based on Iris Abnormal
Iris image analysis is one of the most effective non-invasive diagnosis methods for determining the health condition of organs in clinical diagnosis. Though accurate and quick diagnosis is vital, it is an absolutely necessary requirement of medical science. According to the literature review, several modern technologies also fail to appropriately identify disease. This endeavorinvestigates the field of diagnosis from several angles. Irido diagnosis is an area of medical science that can be used to detect a variety of disorders. Initially, photos of the eyes are recorded, a database with their clinical history is formed, features are identified, and lastly, a classification is made to determine whether or not a diabetic is present. Here For training and classification, a machine learning KNN (K-nearest neighbors)model is used.
For the detection of diabetics using iris images, a novel framework has been built. Enhancement is necessary for the extraction of deep layer features for clinical feature analysis. One eye picture is captured and kept in the project’s database. Apply the DCT transform, then segment and normalizethe iris using the hough transform model to extract the only iris area, and the daugman rubber sheet model to normalizethe iris. Later, using an iris chart and a normalizediris picture, the liver portion is removed. Then it extracts textural characteristics like mean, contrast, energy, and correlation. Finally, the test image is classified as normal or abnormal using the machine learning model KNN algorithm.
Authors: V.Ashok Kumar, A. Komini, V.Sudheer, A.Venkatesh, D.Ravi Kishore
Download full abstract: https://iridology-research.com/pdf/fin_irjmets1656673077.pdf