Newton-Cotes-based FIR smoothing for robust ECG segmentation: Integration with the Modified Varri method

Reliable segmentation of electrocardiographic (ECG) signals is often compromised by noise and non-stationary artifacts, which can lead to unstable boundary detection. This study proposes a preprocessing strategy to enhance the robustness of the Modified Varri method using Newton–Cotes Moving Average (NCMA) filters. NCMA filters are symmetric, normalized FIR operators derived from closed Newton–Cotes quadrature rules. Their construction ensures unit DC gain, linear-phase behavior, and exact preservation of low-order polynomial trends in an integral sense. These properties make them suitable as structured smoothing operators for ECG signals, reducing noise variance while maintaining the underlying signal morphology. NCMA filters with 3, 5, and 7 points were evaluated on both synthetic signals and real recordings from the MIT-BIH Arrhythmia Database using standard denoising metrics (MSE, SNR, and PRD) and a segmentation-specific Boundary Stability Index (BSI). Results show that the 5-point NCMA filter provides the best trade-off between noise reduction and signal fidelity, outperforming Savitzky–Golay smoothing of comparable length on synthetic data. On real recordings, while the 5-point filter balances signal fidelity and noise reduction, the 7-point configuration provides the highest segmentation stability (lowest BSI), despite greater morphological smoothing. Statistical validation across multiple noise realizations confirms the robustness of the proposed approach. These findings demonstrate that NCMA-based preprocessing offers a reliable and lightweight alternative to conventional SG filtering, enhancing ECG segmentation reliability in noisy environments.