Tongue Diagnostic System Uses Color Analysis and Artificial Intelligence in Detecting Diseases With High Accuracy

Two thousand years ago, Chinese herbalists performed the old practice of examining the human tongue to detect signs of disease. Computer scientists are now adopting this method using machine learning and artificial intelligence.


Tongue Color as Reflection of Health

Tongue diagnostic systems have been gaining the interest of experts due to increased global health monitoring. In a recent study, Iraqi and Australian scientists provide more evidence of the increasing accuracy of this method in monitoring diseases.

In this research, engineers from Middle Technical University (MTU) in Baghdad and the University of South Australia (UniSA) used a computer and a USB web camera to capture images of the tongue from 50 patients with anemia, renal failure, and diabetes. Professor Ali Al-Naji led the team to compare the colors with a database of 9,000 tongue images.

Using image processing techniques, the researchers correctly diagnosed the diseases in 94% of cases, compared to laboratory results. The patient or the nominated health provider also received a voicemail via text message specifying the identified tongue color and disease.

Thousands of years ago, Chinese medicine pioneered examining tongue color to detect an illness. This method has been long endorsed by conventional medicine, demonstrating that the signs of blood and heart diseases, circulatory and digestive problems, liver issues, and diabetes can be revealed by the tongue's color, shape, and thickness.

In modern times, this method for diagnosing disease from a tongue's appearance can be done using AI, a camera, or even a smartphone. Computerized tongue analysis shows high accuracy and can help diagnose illnesses remotely easily, painlessly, safely, effectively, and cost-effectively. This approach is especially relevant during a global pandemic such as COVID, with compromised access to health centers.

There were also previous studies where tongue diagnostic systems have accurately diagnosed thyroid disease and appendicitis. According to Al-Naji, diagnosing with 80% accuracy in over ten conditions showing visible tongue color changes is possible. As their study achieved 94% accuracy with three diseases, scientists plan to fine-tune this research even further.

Tongue Appearance in Every Disease

Various diseases manifest distinct changes in tongue appearance. For instance, diabetes patients usually have a yellow tongue, and cancer patients possess a purple language with a thick, greasy coating. Meanwhile, acute stroke patients have a red tongue that is often crooked.

In a separate study conducted in 2021, researchers from Ukraine used tongue color analysis to determine the health condition of COVID-19 patients. They were led by Liudmyla Horzov from the Department of Dentistry at Uzhhorod National University in analyzing the tongue images of 135 COVID patients using a smartphone. It was found that 64% of patients with a mild infection had a pale pink tongue, 62% of patients with a moderate COVID infection had a red language, and 99% of patients with severe viral infection had a dark red tongue.

SARS-CoV-2 infection does not manifest through signs and characteristics that target the tongue or are specific to the tongue. However, COVID-19 triggers changes in the color of the tongue and tongue plaque, similar to those in other internal pathologies.

Check out more news and information on Tongue in Science Times.

Join the Discussion

Recommended Stories

Real Time Analytics