Fetal QRS Complexes Detection Using Deep Learning Technique

The fetal Q, R, and S peaks complex detection in Non-invasive fetal ECG is an important procedure to ensure the fetal condition during the pregnancy. However, the detection process is quite complex because of the presence of large-amplitude maternal ECG signals. While conventional approaches lag, detecting devices should deliver data with low accuracy and low sensitivity. As a result of the findings in the current study, an architecture based on a convolutional neural network model -LeNet is proposed for reliable detection of fetal QRS complexes. The proposed deep learning model is experimented with non-invasive fetal electrocardiogram (NI-fECG). NI-FECG physio net data and compared with conventional, support vector machine (SVM), Naive Bayes, k-nearest neighbor (KNN), and convolutional neural network (CNN) algorithms. With maximum accuracy of 99.46% the proposed model attains maximum performance for all other parameters like precision, recall and F-measure compared to existing state of art of techniques.