Over the past three years, medical technology experts have been working to leverage artificial intelligence (AI) to quickly and accurately narrow down a patient’s illness.
From AI predicting lung disease with over 90% accuracy to general AI models predicting the likelihood of disease before a traditional diagnosis is made, technology is rapidly developing today.
Now, engineering researchers from the Middle Technical University (MTU) in Baghdad, Iraq, and the University of South Australia (UniSA) in Adelaide, Australia, have achieved a breakthrough in training a machine learning algorithm to detect disease through tongue colour analysis.
An AI algorithm looked at the color of a person's tongue and was 98 percent accurate in predicting various diseases.
AI diagnoses diabetes, stroke, COVID-19
The AI model diagnosed diabetes, stroke, anemia, asthma, liver and gallbladder diseases, COVID-19, and a range of vascular and gastrointestinal problems, the scientists said in a statement.
Senior author Ali Al-Naji, from MTU and adjunct associate professor at UniSA, said the AI replicates a 2,000-year-old practice widely used in traditional Chinese medicine of examining the tongue to look for signs of disease.
“The color, shape and thickness of the tongue can reveal a variety of health conditions,” he added.
“Typically, people with diabetes have a yellow tongue, people with cancer have a purple tongue with a thick greasy film, and people with acute stroke have an abnormally shaped red tongue.”
Al Naji also noted that a white tongue can be a sign of anemia, while patients with severe COVID-19 are likely to have a bright red tongue, while a deep blue or purple tongue indicates vascular or gastrointestinal problems or asthma.
AI model trained on 5260 images
So experts trained a computer vision system with a new image processing system using 5,260 images differentiated by color classes: red, yellow, green, blue, gray, white and pink.
Six machine learning algorithms were used to train an AI-based computer algorithm that predicts tongue color in any lighting condition.
These systems include Naive Bayes (NB), Support Vector Machines (SVM), K-Nearest Neighbors (KNN), Decision Trees (DT), Random Forests (RF), and Extreme Gradient Boosting (XGBoost).
The authors state in their study that the study proposes a novel AI imaging system to analyze and extract tongue color features from five color space models (RGB, YcbCr, HSV, LAB, YIQ) at different color saturation and under different light conditions.
To test the system, 60 images of the tongues of patients with specific health conditions were collected from two teaching hospitals in the Middle East.
The statement highlighted that in most cases the AI model was able to successfully match tongue colour with the disease a patient had been diagnosed with.
Additionally, to capture images of the patients' tongues, the camera was placed 20 centimetres away from the fleshy muscular organ.
Real-time diagnostics
This confirms that AI can indeed advance the medical field by detecting diseases earlier and providing an on-the-spot diagnosis through instant scrutiny of the tongue colour.
Real-time diagnosis could reduce wait times and lead to shorter wait times at hospitals. Diagnosis of a patient's illness may still need to be done by a human, but AI models could help experts confirm the diagnosis sooner.
Co-author and professor Javan Shaar from UniSA said that in future smartphones could be used to diagnose diseases in this way.
“These results confirm that computerized tongue analysis is a safe, efficient, easy-to-use and affordable disease screening method that supports modern methods in a centuries-old practice,” he says.
The research was published in the journal Technologies.
Newsletter
Blueprint Daily
Get the latest news in engineering, technology, space and science with The Blueprint.
By clicking sign up, you agree to this site's Terms of Use and Privacy Policy.
About the Editor
Shubhangi Dua As an eccentric and imaginative multimedia journalist with a Masters in Magazine Journalism, I am constantly coming up with fresh ideas and finding innovative ways to tell a story. I have dabbled in various media fields, from holding the pen as a writer to capturing moments as a photographer and even strategizing on social media. With a creative mind and attention to detail, I have worked in the dynamic field of multimedia journalism and written on sports, lifestyle, arts, culture, health and wellbeing for Further Magazine, Alt.Cardiff and The Hindu. I am on a mission to create a media landscape that is as diverse as a Spotify playlist. From India to Wales and now England, my journey has been full of adventures that inspire me to paint, cook and write.
