Deep learning-based multi-level airway collapse site localization on obstructive sleep apnea patients based on snoring sounds
Researchers used deep learning models, specifically AST and ResNet-50, to locate upper airway collapses in obstructive sleep apnea patients by analyzing snoring sounds from drug-induced sleep endoscopy. The AST model performed better than ResNet-50 for identifying obstructions at the Velum (V) and Retropalatal (RP) levels. This suggests a promising non-invasive tool for tracking overnight airway dynamics.