Cardiac Analysis with Advanced ECG Feature Extraction and Reduction
Jamjoom, Jude Qashqari, Maha Alzahrani, Maria
Jamjoom, Jude
Qashqari, Maha
Alzahrani, Maria
Type
Supervisor
Subject
Date
2025-11-20
Collections
Research Projects
Organizational Units
Journal Issue
Abstract
This paper introduces an innovative Electrocardiogram (ECG) feature extraction and reduction system designed to elevate cardiac analysis capabilities. By integrating cutting-edge multi-domain feature extraction techniques with dimensionality reduction methods, the system offers a comprehensive approach to cardiac data analysis. Notably, it combines advanced transform-based features, such as the Hilbert transform, Mel-frequency Cepstral Coefficients (MFCC), and Multiple Signal Classification (MUSIC) algorithm analyses, with a range of effective feature reduction techniques, including Principal Component Analysis (PCA), Independent Component Analysis (ICA), Multi-Dimensional Scaling (MDS), Uniform Manifold Approximation and Projection (UMAP), and autoencoders. Processing two-channel ECG recordings from the (MIT-BIH) arrhythmia database in 60-s intervals, the system extracts and condenses 59 diverse features, capturing subtle nuances and prominent cardiac characteristics. PCA and ICA techniques yielded 19 reduce
Department
Publisher
Sponsor
Effat University
Copyright
CC0 1.0 Universal