Tongue Image Analysis for Hepatitis

In this study, we proposed an automated methodology to analyze and detect and analyze hepatitis by using tongue images analysis based on genetic algorithm–support vector machine. There is a strong association in between the characteristics of tongue and human health diagnosis. GA-SVMs are used to establish a relationship between diseases like appendicitis, diabetes and characteristics of tongue. All the tongue samples are captured with the help of TDA-2 device. GA-SVMs are trained and tested with features like texture and chromatic information obtained from tongue image samples. GA-SVM classification accuracy of hepatitis is enhanced by obtaining unbiased features of tongue image samples. Proposed methodology achieved high accuracy by providing relevant data in training phase and removing irrelevant data. Tongue is very important organ in modern medical treatment as well as in Ayurveda medical treatment which is useful to detect and analyze diseases like appendicitis, pancreas, liver related diseases and diabetes. Experiments are performed on data set consisting of 1298 healthy and 2057 disease tongue samples. The accuracy of tongue images related to hepatitis is 93.74%.