Software System for Arrhythmia Classification in Real Time of Segmented ECG Signals

This article introduces online software that enables users to classify ECG signals in short period of time. To achieve this, a neural network was fine-tuned by adjusting its hyperparameters and training it to classify six types of arrhythmias found in the PTB database. Prior to training, the signals were preprocessed by removing noise, normalizing them using cubic spline, and segmenting each signal into smaller 2-second segments. The performance of this neural network was assessed using several metrics, including accuracy, precision, recall, f1 score, and loss. The corresponding values for these metrics were 89.179%, 88.805%, 89.179%, 88.882%, and 0.407, respectively. Subsequently, this neural network was utilized to develop a user-friendly software interface that allows individuals to upload their ECG signals and view the results in a straight forward manner.