A Graph-constrained Changepoint Detection Approach for ECG Segmentation.

Area of Science:

• Biomedical Signal Processing
• Cardiovascular Disease Assessment
• Time-Series Analysis

Background:

• Electrocardiogram (ECG) signals are crucial for non-invasive cardiovascular disease assessment.
• Accurate R-peak detection is essential for ECG signal segmentation and analysis.
• Existing algorithms often require preprocessing, limiting real-time application and reliability with noisy data.

Purpose of the Study:

• To introduce a novel graph-based optimal changepoint detection (GCCD) method for R-peak detection.
• To develop a reliable R-peak detection algorithm that does not require preprocessing steps.
• To provide a generic method applicable to various time-series biomedical signals.

Main Methods:

• A graph-based optimal changepoint detection (GCCD) model was developed.
• The method computes globally optimal changepoint detection solutions.
• The algorithm was evaluated on the MIT-BIH arrhythmia (MIT-BIH-AR) database.

Main Results:

• The GCCD method achieved high performance metrics on the MIT-BIH-AR database.
• Achieved sensitivity (Sen) of 99.76%, positive predictivity (PPR) of 99.68%, and detection error rate (DER) of 0.55%.
• Performance is comparable to state-of-the-art R-peak detection approaches.

Conclusions:

• The proposed GCCD method offers reliable R-peak detection without preprocessing.
• The algorithm provides globally optimal solutions and is robust to noisy ECG data.
• The generic nature of the method allows for application to other biomedical time-series signals.