Preclinical Diagnosis of Diabetes with Tongue Infrared Thermography and PSO Algorithm

Abstract

The tongue use as an initial biomarker for disease diagnosis in Greek, western and Chinese culture. The tongue provides a non-invasive method to diagnose diseases in health care. The tongue shows gamut of symptoms due to disease and disorder. The traditional tongue diagnosis requires quantitative experience. In western medicine the tongue symptom use as one of the criteria for disease diagnosis. The tongue reflexology shows functioning of internal organs in body. Recent studies attribute chronic pancreatitis to saliva secretion. In this paper, we propose to diagnose diabetes with tongue thermal image. The Particle swarm optimization (PSO) algorithm to cluster thermally active pixels with similar intensity. The clustered tongue thermal region show drastic change for normal and diabetic person. The study involved 25 normal person and 25 type 1 diabetic person. The proposed approach diagnoses diabetes with 86% accuracy.