The Frequency Modulated Möbius (FMM) decomposition method accurately estimated ECG segment lengths, achieving low misclassification errors of 3% to 6% for pathological QRS and QT criteria.
Does the Frequency Modulated Möbius (FMM) decomposition approach improve the accuracy and robustness of peak width and wave duration estimation in ECG and spectral signals compared to traditional methods?
The novel FMM-based method provides a robust, mathematically sound approach for accurately estimating ECG wave durations and spectral peak widths, outperforming standard algorithms especially in noisy, asymmetric, or multi-channel data.
Accurately estimating peak width and wave duration (WD) is a significant challenge across various scientific disciplines. Traditional methods, such as Full Width at Half Maximum (FWHM), often encounter difficulties with overlapping peaks, asymmetric patterns, noisy data, and, especially, multi-channel data where these challenges multiply. This paper introduces a novel approach for estimating peak width and WDs in curve-fitting applications, leveraging the oscillatory nature of signals through Frequency Modulated Möbius (FMM) decomposition. We derive a parametric expression for FWHM and propose a novel WD measure. Beyond being both mathematically and physiologically sound, this method offers several advantages, including a straightforward parametric formulation, robust estimation, and flexibility to handle single and overlapping peaks, as well as peaks recorded across multiple channels. While potential applications extend across disciplines, we demonstrate this method’s effectiveness in addressing critical challenges in electrocardiogram (ECG) signal and spectroscopic analysis. In ECG analysis, the WD measure effectively estimates ECG segments that capture critical aspects of cardiac electrical activity. In spectroscopy, we evaluate the new FWHM estimator, a key parameter for determining spectral resolution and material properties. Extensive testing confirms the suitability and robustness of the new measures, outperforming standard techniques in both applications.
Rueda et al. (Sat,) conducted a other in ECG signal analysis and X-ray Photoelectron Spectroscopy (n=19,922). Frequency Modulated Möbius (FMM) decomposition vs. Traditional methods (12SL algorithm, Gaussian, Lorentzian, Voigt models) was evaluated on Accuracy of QRS and QT segment length estimation and FWHM estimation. The Frequency Modulated Möbius (FMM) decomposition method accurately estimated ECG segment lengths, achieving low misclassification errors of 3% to 6% for pathological QRS and QT criteria.