Related Experiment Video
Updated: Apr 12, 2026
05:49
Author Spotlight: Advancing CBCT and Digital Dental Image Integration with AI-Assisted Digitization
Published on: February 23, 2024
1.7K
Artificial intelligence and image processing framework for automated airway invasion detection and residue
Luiza Araújo1,2, Enzo Rangel3, Anibal Cotrina-Atencio4
1Postgraduate Program in Neuroengineering, Edmond and Lily Safra International Institute of Neuroscience, Santos Dumont Institute, Zona Rural, Macaíba, 59288-899, RN, Brazil.
Scientific Reports
|April 10, 2026
Summary
This study introduces an AI framework to objectively analyze Fiberoptic Endoscopic Evaluation of Swallowing (FEES) videos. The system accurately detects aspiration, penetration, and pharyngeal residue, aiding clinical decision-making.
Area of Science:
- Medical Imaging
- Artificial Intelligence
- Otolaryngology
Background:
- Fiberoptic endoscopic evaluation of swallowing (FEES) is crucial for diagnosing dysphagia.
- FEES interpretation is subjective, relying on lighting and evaluator expertise.
- Objective analysis tools are needed to improve FEES reliability.
Purpose of the Study:
- To develop and validate an AI and image processing framework for objective FEES analysis.
- To automatically detect penetration, aspiration, and classify pharyngeal residues.
- To enhance clinical decision-making support for dysphagia management.
Main Methods:
- Developed an AI framework integrating anatomical tracking (epiglottis, arytenoids, vocal folds).
- Employed image processing techniques: contrast adjustment, color segmentation, enhancement filters.
- Validated the system using 60 FEES videos, assessing penetration, aspiration, and residue severity (YPR-SRS).
Main Results:
- Achieved high accuracy for penetration (0.90) and aspiration (0.87).
- Demonstrated strong performance in classifying pharyngeal residues in pyriform sinuses (accuracy 0.91-0.95) and valleculae (accuracy 0.95-1.00).
- Reported high sensitivities, specificities, and Kappa values across all assessed parameters.
Conclusions:
- The developed AI framework shows significant assistive potential for FEES analysis.
- This technology can objectively support clinical decision-making in identifying penetration, aspiration, and pharyngeal residues.
- This automated classification capability for FEES findings is a novel contribution.
