Fast Mask Recurrent Convolutional Neural Network for IoT-Based Maternal and Fetal Monitoring in High-Risk Pregnancies

Abstract

Ensuring the well-being of fetuses and their timely diagnosis for potential abnormalities is a critical aspect of healthcare. Early identification of intrauterine growth restriction can facilitate appropriate interventions and improve neonatal outcomes. This study presents a novel approach incorporating the Internet of Things (IoT) and Artificial Intelligence (AI) in the medical domain for the automatic detection of fetal abnormalities. IoT sensors were employed to gather maternal clinical data, including temperature, blood pressure, oxygen saturation levels, and fetal heart rate. A Fast Mask Recurrent Convolutional Neural Network (FMRCNN) was proposed to predict and accurately classify a range of conditions affecting pregnant women and their unborn children. The developed FMRCNN model learns, segments, and classifies fetal abdominal images to identify abnormalities. Additionally, a unified fetal abnormality prediction model was established to process and classify both fetal abdomen and brain ultrasound images. Comparative performance analysis was conducted using Convolutional Neural Networks (CNN), Random Forest (RF), and Support Vector Machine (SVM) algorithms. Evaluation metrics, such as F1score, accuracy, precision, recall, and sensitivity, were employed to assess the effectiveness of the proposed approach. The results indicate that the presented FMRCNN model holds promise for IoT-based maternal and fetal monitoring in high-risk pregnancies.