Tongue Colour Diagnosis System Using Convolutional Neural Network

Abstract Tongue diagnosis is known as one of the effective and yet noninvasive techniques to evaluate patient’s health condition in traditional oriental medicine such Chinese Korean medicine. However, due ambiguity, practitioners may have different interpretation on tongue colour, body shape texture. Thus, research automatic system needed for aiding recognizing features diagnosis. In this paper, a based Convolution Neural Network (CNN) classifying colours proposed. The extracts all relevant information (i.e., features) from three-dimensional digital image classifies into (i.e. red or pink). Several pre-processing data augmentation methods been carried out evaluated, which include salt-and-pepper noises, rotations flips. proposed achieves accuracy up 88.98% 5-fold cross validation. Compared reported baseline Support Vector Machine (SVM) method, method using CNN results roughly 30% improvement recognition accuracy.