A Novel Short-Term Event Extraction Algorithm for Biomedical Signals.

Area of Science:

• Biomedical Signal Processing
• Nonlinear Filtering Techniques
• Computational Physiology

Background:

• Accurate extraction of short-term events from noisy biomedical signals is crucial for diagnosis and monitoring.
• Existing methods may lack robustness or efficiency in diverse biomedical applications.
• Need for advanced filtering techniques to enhance subtle physiological events.

Purpose of the Study:

• To introduce a fast, novel nonlinear filtering method, Relative-Energy (Rel-En), for robust short-term event extraction.
• To evaluate the performance of Rel-En across multiple biomedical signal types and event detection tasks.
• To demonstrate the adaptability and effectiveness of Rel-En in improving existing signal processing pipelines.

Main Methods:

• Development of a Rel-En algorithm that extracts short- and long-term signal energies.
• Application of a derived coefficient vector to selectively amplify events of interest.
• Validation on benchmark datasets for ECG QRS-complex, EEG K-complex, and iPPG peak detection.

Main Results:

• Rel-En successfully identified target events in ECG, EEG, and iPPG signals.
• Achieved state-of-the-art or comparable performance in QRS-complex and K-complex detection.
• Significantly improved iPPG peak detection when used as a preprocessing step.

Conclusions:

• The Rel-En method provides a robust and efficient means for extracting short-term biomedical events.
• The algorithm is easily implemented, with intuitive parameter selection.
• Rel-En shows potential for broad application in various biomedical signal processing tasks requiring event detection.