Advancing Fetal Heart Rate Monitoring: AI-Based Approach for Baseline Detection and Classification

Automated fetal heart rate monitoring (fhr) plays a crucial role in minimizing inconsistencies among experts, thereby reducing harmful birth complications and unnecessary medical procedures. This research proposes an algorithm for dynamic calculation of baseline variability through the utilization of an exponential moving average, tailored for the classification of cardiotocography (CTG) data. The algorithm categorizes fhr values into baseline data, acceleration, deceleration or noise, with the primary objective of identifying distinct segments within the CTG recordings. Subsequently, the segmented data serves as essential training input for a deep learning neural network, named NexoCTGNet, designed for AI-based baseline detection. The integration of this proposed method aims to enhance the accuracy and efficiency of fetal heart rate monitoring, thereby contributing to advancements in diagnostic capabilities. Testing results reveal an accuracy of 0.94, with notable precision (0.99) and recall (0.93) for baseline detection, highlighting the algorithm's effectiveness in improving baseline classification accuracy.