Towards real-time QRS feature extraction for wearable monitors

Area of Science:

• Biomedical Engineering
• Signal Processing
• Wearable Technology

Background:

• Wearable physiologic monitors require computationally compact algorithms due to limited energy and memory.
• Traditional ECG analysis methods may be too resource-intensive for remote monitoring devices.
• Efficient extraction of QRS complexes is crucial for accurate ECG interpretation in ambulatory settings.

Purpose of the Study:

• To evaluate computationally compact algorithms for QRS complex detection in ECG signals.
• To compare the performance of a Short Time Fourier Transform (STFT) algorithm against a more complex Stockwell Transform (SWT).
• To assess algorithm suitability for resource-constrained wearable ECG monitors.

Main Methods:

• Generated synthetic and used measured ECG signals with added noise.
• Implemented and evaluated a Short Time Fourier Transform (STFT) based ECG analysis algorithm.
• Compared the STFT algorithm's sensitivity and specificity against the Stockwell Transform (SWT).

Main Results:

• The STFT algorithm demonstrated comparable sensitivity and specificity to the SWT.
• The STFT algorithm possesses a significantly smaller computational and memory footprint.
• This indicates feasibility for application in low-power wearable devices.

Conclusions:

• A compact STFT algorithm is effective for QRS complex detection in ECG signals.
• The STFT algorithm offers a viable, resource-efficient alternative for wearable physiologic monitors.
• This advancement supports the development of more capable and accessible remote health monitoring solutions.