A computer algorithm may identify a person’s health condition in real-time — with 98% accuracy — just by analysing the colour of their tongue. Engineering researchers from the University of South Australia (UniSA) in Adelaide, Australia, and Middle Technical University (MTU) in Baghdad, Iraq, conducted tests using artificial intelligence (AI) to analyse tongue colour for illness diagnosis.
Tongue colour is an important health indicator for identifying disorders and monitoring their course. The tongue’s colour, shape and coating are among its many properties that indicate various medical issues.
Research findings have been published in the Technologies journal. Senior study author Ali Al-Naji explained that typically, people with acute stroke present with an abnormally shaped red tongue; those with diabetes have a yellow tongue; and cancer patients have a purple tongue with a thick greasy coating.
Al-Naji went on, “A white tongue can indicate anaemia; people with severe cases of COVID-19 are likely to have a deep-red tongue. An indigo- or violet-coloured tongue indicates vascular and gastrointestinal issues or asthma.”
To train machine learning algorithms for tongue colour detection, engineering researchers utilised 5260 photos.
Two Middle Eastern teaching institutions provided 60 tongue photos from individuals with a variety of medical issues. In nearly all cases, the AI model was able to correlate the colour of the tongue with the illness.
AI, according to Ali Al-Naji, is imitating a 2000-year-old technique that is extensively employed in traditional Chinese medicine: looking for sickness symptoms on the tongue.
In the experiment, a patient’s tongue colour was recorded by cameras 20 centimetres away, and an imaging system used this data to forecast the patient’s health status in real-time.
As per the scientists’ statement, the AI model identified several medical illnesses like diabetes, asthma, anaemia, liver and gallbladder disorders, COVID-19, and various gastrointestinal and vascular problems.
According to the study’s authors, a novel AI image system is proposed in order to analyse and extract tongue colour characteristics from five distinct colour space models (RGB, YcbCr, HSV, LAB, and YIQ) under various colour saturations and lighting circumstances.
In the future, a smartphone will be utilised for this kind of illness diagnosis, according to co-author and UniSA professor Javaan Chahl.
“These results confirm that computerized tongue analysis is a secure, efficient, user-friendly and affordable method for disease screening that backs up modern methods with a centuries-old practice,” Prof Chahl says.
Comments
0 comment