What is it about?
This paper deal with an efficient and robust features extraction and classification of Electrocardiography (ECG) signal. The used method simply based on using the most efficient techniques for feature extraction and classification in order to accurately identify different heart arrhythmias. The main problem with the extraction of ECG signals is non-stationary, which means the decision about the behavior of these signals at any interval of time becomes difficult because of abnormality. For this reason, a specific model has been designed for reading the ECG signal. The results from this model are forwarded directly to be processed using LabVIEW using the biomedical toolkit for the purpose of feature extraction. For feature extraction, four features are extracted and applied to decision classifiers to determine the heart issue (problem). Two kinds of classification algorithms are applied: MLP and SVM. The signals dataset was collected from records that were manually classified using the generated database directory, The SVM classifier achieved the highest detection accuracy which is equal to 99.5% and the total number of identified cardiac arrhythmias is equal to eight.
Featured Image
Photo by Joshua Chehov on Unsplash
Why is it important?
One of the most important signal is Electrocardiography (ECG) signal that reveal electrical activity of human heart. This signal is also used to show the measurement and distribution of electrical activity through cardiac muscles. So, this signal has been utilized by cardiologists to show the important information that is related to rhythm and functioning of the human heart. The analysis of various Classification methods was based on the MIT-BIH ECG database and the generated dataset, which is internationally adopted. These strategies were developed in a simulated environment, eight heart disease cases were identified with these methods, and comparisons were made among the classifiers. The detected heart cases are: N, FDHB, WPW, LGL, RLBB, MI, B, and T.
Perspectives
Writing this article was a fulfilling experience, as it provided an opportunity to collaborate with esteemed colleagues who share a deep commitment to advancing healthcare technology. This work has been particularly rewarding, as it opened pathways to interactions with clinicians and researchers in cardiovascular health. Their feedback not only enriched the study but also highlighted the potential real-world impact of improving diagnostic accuracy using advanced ECG feature extraction and classification techniques. Ultimately, this article deepened my involvement in developing machine learning models that can address critical challenges in medical diagnostics.
Mustafa Ghanim
University of Technology
Read the Original
This page is a summary of: ECG feature extraction and classification using robust technique for critical cardiac cases, January 2024, American Institute of Physics,
DOI: 10.1063/5.0236830.
You can read the full text:
Contributors
The following have contributed to this page
