TinyML based Classification of Fetal Heart Rate using Mother’s Abdominal ECG Signal

Abstract

ECG plays a vital role in the field of medical science. The proposed machine learning model is able to classify the heart rate of the fetus through ECG signals that are collected from the mother’s womb. The noise and interference have been reduced in the ECG signal by preprocessing. The proposed Recurrent neural network (RNN) architecture is made on edge impulse. The potential efficiency of the designed model is studied by varying parameters such as kernel size, learning rate, epochs, etc using the abdominal and direct fetal ECG Database from the physionet. During training, the classification of heart rate using ECG signals has achieved 75% accuracy. The beauty of the model is that it is designed with tinyML for implementation on embedded systems.