ApneaDetector: Detecting Sleep Apnea with Smartwatches
summary
A smartwatch app uses wrist accelerometer data, signal preprocessing and machine learning to flag apnea events, not just sleep stage guesses. It applies signal denoising and trend calibration, detects spikes with a maximum standardized residual, and extracts peak, interpeak and cross-axis covariance features. A 20-subject clinical study against PSG yields precision 0.9674, recall 0.9625 and F1 0.9649, showing the built system reliably detects obstructive and hypopnea patterns.
AI-written summary · about