The Period-Aware Attention Network (PA²Net) achieved excellent fetal ECG detection performance, with positive predictive values of 98.47% to 99.74% across multiple benchmark databases.
Does the Period-Aware Attention Network (PA²Net) improve the detection of fetal ECG signals from abdominal ECGs?
The proposed PA²Net algorithm demonstrates high accuracy in detecting noninvasive fetal ECG signals from abdominal recordings across multiple benchmark datasets.
The noninvasive fetal ECG (FECG) is used to monitor fetal well-being at prenatal and intrapartum. However, it is challenging to detect the FECG signal from the abdominal ECG due to the following problems: 1) the FECG signal is weak and often masked by noise; 2) the FECG signal is mixed or overlapped with the maternal ECG (MECG) signal. To solve such problems, a Period-Aware Attention Network (PA2Net) is proposed for FECG detection by designing three modules, where a FECG period-aware attention module (FPAM), which is designed to suppress the noise interference by modeling the periods and features of signals, is first employed to detect FECG signals masked in noise; next, a MECG period-aware attention module, which is generated from the FPAM by the KL-divergence-based weight sharing module, can collaborate with the FPAM to search for mixed ECG signals of FECG and MECG; finally, an anti-aliasing signal separation module is developed to estimate FECG signals from mixed ECG signals. Experiments conducted on the benchmarks show that the proposed PA2Net can achieve the excellent performance with the PPV of 98.83%, 99.74%, 99.67%, and 98.47% on PCDB, ADFECGDB, NIFECGDB, and LHDB databases, respectively, allowing obstetricians to use PA2Net to monitor fetal health and diagnose diseases.
Wang et al. (Sat,) conducted a other in Fetal ECG detection. Period-Aware Attention Network (PA²Net) was evaluated on Positive Predictive Value (PPV) for fetal ECG detection. The Period-Aware Attention Network (PA²Net) achieved excellent fetal ECG detection performance, with positive predictive values of 98.47% to 99.74% across multiple benchmark databases.