Development of Prognosis Tool for Type-II Diabetics using Tongue Image Analysis

Tongue image analysis is a method to identify the conditions of internal organs of the human body (i.e., pancreas). It has certain limitations due to subjective, qualitative and experimental analysis. A person with diabetics can have a grey colour coating at the center of the tongue. Thus the abnormalities of tongue images exhibit the condition (i.e., whether the person is diabetic or not). Based on that, this paper represent a computerised image analysis to improve the efficiency of the classification. Likewise, the three quantitative features like geometry, colour and texture are measured using MATLAB, then the SVM classifier and CNN are used for the classification images from the dataset and the results of both methods are compared. Based on the accuracy of prediction CNN has better results than SVM.