A Novel Approach for Detecting Fetal Electrocardiogram (FECG) Signals: Integration of Convolutional Neural Network (CNN) with Advanced Mathematical Techniques

This paper presents an innovative method for recognizing fetal electrocardiogram (ECG) signals using a single-channel abdominal lead. The method combines the capabilities of Convolutional Neural Network (CNN) with advanced analytical techniques, including independent component analysis (ICA), Singular Value Decomposition (SVD), and Nonnegative Matrix Factorization (NMF), to reduce the dimensionality of the data. A crucial aspect of the fetal heart rate, which distinguishes it from the mother’s heart rate, is the necessity for a time-scale representation that effectively captures the fetal electrical activity in terms of energy. Additionally, by separating the various components of the fetal ECG, it becomes possible to utilize them as inputs to the CNN model, thereby optimizing the reconstruction of the actual fetal ECG via the proposed method. The experimental findings demonstrate the effectiveness of this innovative technique, indicating the potential for real-time extraction of FECG signals. This method holds promise for enhancing fetal monitoring and healthcare applications.