Tongue colour diagnosis system using convolutional neural network

Abstract

Tongue diagnosis is known as one of the effective and yet noninvasive technique to evaluate patient’s health condition in traditional oriental medicine such as traditional Chinese medicine. However, due to ambiguity, practitioners may have different interpretation on the tongue colour, body shape and texture. Thus, research of automatic tongue diagnosis system is needed for aiding practitioners in recognizing the features for tongue diagnosis. In this project, a tongue diagnosis system based on Convolution Neural Network for classifying tongue colours is proposed. The system extracts all relevant information (i.e., features) from three-dimensional digital tongue image and classifies the image into one of the colour (i.e. red or pink). To increase the accuracy of the proposed system, a number of pre-processing and data augmentation are carried out and evaluated. Augmentation techniques evaluated consists of salt-and-pepper noises, rotations and flips. Synthetic one-sided flip has that proven that it increases the average accuracy from 75.41% to 75.72%. Thus, this technique is proposed for data augmentation in tongue diagnosis applications. The proposed system achieved accuracy up to 88.98% and average of 75.72% from 5-fold cross validation, and 0.05 seconds in processing time.