Abstract:
Among the cruelest life-taking diseases around the world, heart diseases top the chart, and it is critically important to have a timely diagnosis of these heart diseases for improving the health conditions of people.
Providing information about the rate and rhythms of heart beats, an Electrocardiogram (ECG) is one of the most basic, easy-to-use, and efficient ways of determining the condition of a human heart. Due to its non-invasive nature, it is very easy for physicians to detect any possible abnormalities in the person’s heart by merely analyzing the ECG waveforms. Arrhythmia is the most basic type of heart abnormality, that has affected countless hearts all across the world. Arrhythmia has many types and some of them can lead to worse severe heart diseases which might be uncurable if not detected timely.
With the advancement in technology in the medical industry, Machine Learning techniques have lately been applied to various medical datasets to automate the process of the prediction of abnormal heart rhythms in patients.
This project focuses on developing a prototype for a cost-effective, portable system that will be exploiting appropriate Machine Learning Techniques on 12-Lead ECG signals and predict the possibility of the presence of arrhythmic rhythms in the heart. This ML-based arrhythmia classification system will improve and automate the process of detecting several heart arrhythmias and reducing physicians or doctors’ workload.
The proposed device will implement a prediction function on the real-time ECG signal that has been acquired from the subject, predicts the possible arrhythmic or normal heart rhythm, and display the results on a useful medium.
