Health evaluation device using tongue analysis based on sequential image analysis

Abstract

This study employs digital image processing as a tool to reduce the complexity in tongue analysis. Through visual inspection, classification of tongue coat and body color is entirely based on the verdict of the medical personnel thus not providing an absolute truth towards chronic and acute irregularities the patient is suffering from. A firmware which can extract the natural color of the tongue and yield an accurate list of symptoms that might lead to a conclusion of anomalous working on an individual’s internal organ was developed. Its system framework comprises of: (1) image acquisition (2) image segmentation; (3) body and coat segmentation; (4) color classification and; (5) neural network. Moreover, the proposed device is an open source technology developed using Python programming language and Linux operating system on a Raspberry Pi 2 B+. The experiment revealed that the proposed method produced significant results with an accuracy of 93.33%.