The Role of Tongue Examination in Traditional Chinese Medicine and Modern Technology

Throughout history, practitioners of traditional Chinese medicine have considered tongue examination an integral part of a comprehensive medical check-up. They focus on the tongue’s color, shape, and coating to detect diseases. With technological advancements, researchers are now attempting to integrate this traditional method into disease diagnosis using modern artificial intelligence techniques.

The Importance of Tongue Color in Traditional Chinese Medicine

Traditional Chinese medicine views tongue color as closely linked to the state of the blood and vital energy, making it a key indicator in assessing a patient’s overall health. However, tongue examination heavily relies on the practitioner’s individual observation and analysis, rendering it a highly subjective process.

In 2022, the World Health Organization included traditional Chinese medicine diagnoses in the International Classification of Diseases, despite ongoing debate within the global scientific community. Nonetheless, there is strong academic interest in this field.

Challenges in Using Technology for Disease Diagnosis Through Tongue Color

Despite significant advances in computing technologies, there remain substantial challenges in using technology to diagnose diseases through tongue color. One major challenge is the perception bias caused by varying lighting conditions. To overcome this hurdle, a research team developed a standardized lighting system within a booth setup where patients place their heads in a box illuminated by LED lights, providing a stable and controllable lighting environment.

The researchers collected over 5,260 images to train machine learning models to recognize specific colors under different lighting conditions. They used these models to predict medical conditions associated with tongue color, such as diabetes, asthma, COVID-19, and anemia, achieving a test accuracy of 96.6%.

Results and Potential Applications of the Technology

The most accurate model among six was applied to tongue images taken in the standardized booth setup at two hospitals in Iraq, where the experimental diagnoses were compared to patients’ medical records. The system successfully identified 58 out of 60 images correctly.

Researchers are currently focusing on diagnosing the center and edges of the tongue and plan to use a new image database to examine tongue shape and oral conditions such as ulcers and cracks using the deep learning algorithm YOLO.

Conclusion

While tongue color can serve as a useful biological indicator of a person’s health status, it cannot be solely relied upon for making precise clinical decisions. Current tongue analysis systems require further improvements to become part of conventional diagnoses. Challenges remain in collecting usable data and standardizing tongue examination in clinical settings, but with ongoing technological progress, we may see a closer integration between traditional medicine and modern technology in the future.