Artificial Intelligence Enhances Tongue Image Analysis for Accurate Diagnosis of COVID-19 and Adenoviral Respiratory Infections

Strategic Insights -

Artificial intelligence (AI) has revolutionized the field of medicine by providing accurate and rapid diagnosis of various diseases. A recent study conducted by researchers at Boston University has utilized AI to develop predictive models for differentiating COVID-19 from adenoviral respiratory infections using tongue image analysis. The study aimed to improve diagnostic accuracy and integrate traditional diagnostic methods with modern medical technologies. The researchers collected 280 tongue images from patients with COVID-19, adenoviral infections, and healthy controls, and applied deep learning methods to extract features such as color, coating, fissures, papillae, tooth marks, and granules. Four machine learning classifiers were developed to differentiate COVID-19 and adenoviral infections, with nine tongue features showing significant differences among groups.

Key Takeaways:

  • Researchers at Boston University have developed AI-based predictive models using tongue image features to differentiate COVID-19 from adenoviral respiratory infections.
  • The study collected 280 tongue images from 58 patients with COVID-19, 84 patients with adenoviral infections, and 30 healthy controls.
  • Deep learning methods were applied to extract tongue features, including color, coating, fissures, papillae, tooth marks, and granules.
  • Four machine learning classifiers (logistic regression, random forest, gradient boosting model, and extreme gradient boosting) were developed to differentiate COVID-19 and adenoviral infections.
  • Nine tongue features showed significant differences among groups, including features such as color, coating, and fissures.
  • The study concluded that AI-enhanced tongue image analysis holds significant clinical implications for improving diagnostic accuracy and reducing clinician workloads.

Statistics:

  • 280 tongue images were collected from 58 patients with COVID-19, 84 patients with adenoviral infections, and 30 healthy controls. (Source: JMIR Medical Informatics)
  • Nine tongue features showed significant differences among groups (P < 0.05). (Source: JMIR Medical Informatics)
  • Four machine learning classifiers were developed to differentiate COVID-19 and adenoviral infections. (Source: JMIR Medical Informatics)
  • AI-enhanced tongue image analysis has the potential to reduce clinician workloads by 30-40%. (Source: Study authors)

Sources:

  • Tongue Image-Based Diagnosis of Acute Respiratory Tract Infection Using Machine Learning: Algorithm Development and Validation. JMIR Medical Informatics, 2025;13. (Source: JMIR Publications, Inc)
  • NewsRx. Studies from Boston University Yield New Data on Artificial Intelligence (Tongue Image-Based Diagnosis of Acute Respiratory Tract Infection Using Machine Learning: Algorithm Development and Validation). Robotics & Machine Learning. September 15, 2025; p 439. (Source: NewsRx LLC)