Related Experiment Video
Updated: Sep 10, 2025
05:56
Objectification of Tongue Diagnosis in Traditional Medicine, Data Analysis, and Study Application
Published on: April 14, 2023
2.6K
Tongue Image-Based Diagnosis of Acute Respiratory Tract Infection Using Machine Learning: Algorithm Development and
Qianzi Che1, Yuanming Leng2, Wei Yang1
1Institute of Clinical Basic Medicine of Chinese Medicine, Chinese Academy of Traditional Chinese Medicine, No.16, Nanxiao street, Dongzhimen, Dongcheng District, Beijing, 100700, China.
JMIR Medical Informatics
|August 25, 2025
Summary
Artificial intelligence can analyze tongue images to differentiate COVID-19 from human adenovirus (HAdV) infections. This noninvasive method shows promise for improving respiratory infection diagnostics.
Area of Science:
- Medical diagnostics
- Artificial intelligence in healthcare
- Traditional Chinese Medicine
Background:
- COVID-19 and human adenoviruses (HAdVs) present similar symptoms, complicating diagnosis without testing.
- Tongue diagnosis, a traditional Chinese medicine technique, shows potential for identifying respiratory infections.
- AI can enhance objective analysis of tongue images for improved diagnostic accuracy.
Purpose of the Study:
- Develop and validate AI models for differentiating COVID-19 from HAdV infections using tongue images.
- Integrate traditional diagnostic methods with modern AI technologies.
- Improve diagnostic accuracy for acute respiratory tract infections.
Main Methods:
- Collected 280 tongue images from COVID-19 patients, HAdV patients, and healthy controls.
- Applied deep learning to extract tongue features (color, coating, fissures, etc.).
- Developed and compared four machine learning classifiers (logistic regression, random forest, gradient boosting, extreme gradient boosting).
Main Results:
- Nine tongue features, including color, tooth marks, and moisture, significantly differed among groups (P<.05).
- Extreme gradient boosting model achieved the highest performance (AUC 0.84).
- Shapley analysis identified tongue color, moisture, and texture as key diagnostic features.
Conclusions:
- Tongue diagnosis, aided by AI, can help identify pathogens causing acute respiratory tract infections at admission.
- This approach may reduce clinician workload and enhance diagnostic accuracy.
- AI-powered tongue analysis offers a promising, noninvasive tool for respiratory infection diagnostics.
