ECG Heart-rate Estimation
Signal-processing pipeline for robust heart-rate estimation from short ECG segments.
Overview
This project analyzed 200 ECG segments, each 30 seconds long and sampled at 200 Hz, to estimate heart rate under signal-quality and peak-detection variability.
Approach
- Established a baseline estimator with an MAE of 12.38 beats per minute.
- Investigated peak-detection, autocorrelation, and spectral estimates.
- Used physiological plausibility constraints and signal-quality checks.
- Compared errors across records rather than relying only on an overall average.
Technical considerations
Peak detection can fail because of noise, baseline wander, ectopic beats, or incorrect thresholds. Autocorrelation and spectral estimates provide complementary evidence, but each can also select harmonics. A robust estimator therefore needs agreement checks and explicit handling of low-confidence segments.
Limitations
The project dataset and short-window design constrain generalization. Clinical use would require validated acquisition protocols, uncertainty reporting, external cohorts, and medical-device quality controls.
Tools: Python, NumPy, SciPy, digital signal processing, peak detection, spectral analysis