Research on the Nonlinear Algorithm of Real-Time Extraction of ECG Based on Multifractal Theories

Two steps are often used during FECG extraction: MECG estimation and MECG suppression. However, the use of fast independent component analysis (FastICA) for MECG estimation suffers from the uncertainty of permutation order, making it difficult to perform real-time processing and determine the channel corresponding to MECG. A nonlinear algorithm for automatic identifying of MECG (AIMECG-NA) was proposed to address this problem. The algorithm adopted nonlinear method multifractal theory and multifractal detrend fluctuation analysis (MFDFA). SET A database were analyzed, the results show that the channel where MECG is located can be identified using the multifractal spectral width. Experiments were carried out on the NIFECGDB and SET B databases, and the results show that the proposed algorithm can identify the channel where MECG is located. AIMECG-NA can be applied to intelligent devices for ECG analysis, providing crucial physiological information.