Intelligent maternal-fetal healthcare monitoring and abnormality detection using deep belief networks optimized by bacterial foraging algorithm
K. Nandini, K. Rahimunnisa · Biomedical Signal Processing and Control · 2025
Ensuring the health of pregnant women and their fetuses is vital for a safe pregnancy. This study introduces the Intelligent Maternal-Feature Healthcare Monitoring System using Optimal Deep Learning (IMFHMS-ODL) algorithm. Employing an array of sensors, including respiration, ECG, FECG, GSR, PPG, fetal movement, temperature, maternal and fetal heart rates, the system provides real-time monitoring throughout pregnancy. Utilizing a Deep Belief Network (DBN) model and Bacterial Foraging Optimization Algorithm (BFOA) for abnormality detection and model optimization, the IMFHMS-ODL promptly alerts healthcare practitioners upon identifying abnormalities. The study demonstrates superior performance over recent algorithms, achieving a notable accuracy, precision, recall, F-score, and AUC score of 97.00 %, 97.06 %, 96.99 %, 97.00 %, and 96.99 %, respectively, on 30 % of the testing phase. However, the computational time (CT) results reveal a longer processing time for the IMFHMS-ODL system, recording a CT of 0.68 s. In discussions, it is essential to consider the feasibility of implementing the method, conducting experiments with volunteers, and potentially developing a product based on the proposed technique, highlighting the practical maturity of the work.