ECG Signal Processing for Arrhythmias Monitoring using Artificial Intelligence Techniques

Rhythmology is one of the most important and sensitive fields in the cardiovascular area, this last deals mainly with the heart rhythm troubles or arrhythmias. The most difficult challenge lies on the location of those troubles in time series, sometimes the cardiologist gives Holter ECG to the patient for 24 or 48 hours depending on each case. After the recording completion time the patient comes again to the hospital in order to get the analysis of his cardiac rhythm. When there is a lot of patients, the delay of results interpretation increases significantly. In the present study, a deep learning model was developed to classify the ECG recordings and predict the cardiovascular abnormalities with two aims. Firstly, to reduce the time of analysis and avoid more delays afterwards. Secondly, to make ECG analysis more accurate. for that purpose, Python programming language was used to develop and simulate our algorithm. The training and validation data was downloaded from physionet ATM database.