ECG beat detection using filter banks

Area of Science:

• Biomedical Engineering
• Digital Signal Processing
• Cardiology

Background:

• Electrocardiogram (ECG) signal analysis is crucial for diagnosing cardiac conditions.
• Accurate and efficient heart beat detection is a fundamental challenge in ECG processing.
• Existing methods may face limitations in real-time performance and computational efficiency.

Purpose of the Study:

• To develop a novel multirate digital signal processing algorithm for accurate ECG heart beat detection.
• To leverage a filter bank (FB) for enhanced time-frequency analysis of ECG signals.
• To create a computationally efficient and real-time capable beat detection system.

Main Methods:

• Designed a multirate digital signal processing algorithm incorporating a filter bank (FB).
• Decomposed the ECG signal into uniform frequency subbands for independent time and frequency analysis.
• Fused decisions from multiple one-channel beat detection algorithms using features from subbands and a heuristic strategy.

Main Results:

• Achieved high performance with 99.59% sensitivity and 99.56% positive predictivity on the MIT/BIH database.
• Demonstrated minimal beat detection latency, enabling real-time application.
• The filter bank (FB)-based structure operates at the subband rate, ensuring computational efficiency.

Conclusions:

• The developed FB-based algorithm provides a highly sensitive and specific method for ECG heart beat detection.
• The algorithm is suitable for real-time applications due to its minimal latency and computational efficiency.
• The FB-based structure offers potential for performing multiple ECG processing tasks with a single set of preprocessing filters.