AI spots diseases with 98% accuracy via tongue color
A team of researchers from Middle Technical University and the University of South Australia has developed an artificial intelligence (AI) system that can analyze tongues to detect signs of disease with 98% accuracy. The system, inspired by traditional Chinese medicine, uses computer vision and machine learning algorithms to identify changes in tongue color, shape, and thickness that are associated with various health conditions.
According to Ali Al-Naji, senior author and MTU Adjunct associate professor, “The color, shape, and thickness of the tongue can reveal a range of health conditions.” For example, people with diabetes often have yellow tongues, while those with cancer may have purple tongues with a thick, greasy coating. Acute stroke patients may present with unusually shaped red tongues.
To develop the system, researchers trained computer vision systems using 5,260 images of tongues categorized into different colors. They also used six machine-learning algorithms to enable the system to predict tongue color in various lighting conditions. The algorithms included Naïve Bayes, Support Vector Machine, K-nearest neighbors, Decision Trees, Random Forest, and Extreme Gradient Boost.
In a test of the system, researchers used it to examine 60 patient tongue images. The results showed that the AI model was able to match tongue colors in most cases. This suggests that the system has the potential to advance the field of medicine by enabling faster disease detection.
The researchers hope to make the AI model available on smartphones in the future. Javaan Chahl, a co-author from UniSA, stated that “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.”
