Related Experiment Videos
[An algorithm study on telecardiogram diagnosis based on multivariate autoregressive model and two-lead ECG signals]
Ding-fei Ge1, Yu-quan Shao, Hui-zhong Jiang
1Department of Information & Electrical Engineering, Zhe-jiang University of Science and Technology, Hangzhou, China. gedingfei@hotmail.com
Summary
A new multivariate autoregressive (MAR) model technique directly classifies electrocardiogram (ECG) signals, achieving 98.3%-100% accuracy for telediagnosis. Combining two ECG leads improves classification results compared to single leads.
Area of Science:
- Biomedical Engineering
- Signal Processing
- Cardiology
Background:
- Telediagnosis of electrocardiogram (ECG) signals faces challenges due to time delays in remote signal reconstruction.
- Accurate and efficient ECG analysis is crucial for timely medical intervention.
Purpose of the Study:
- To develop a direct classification method for ECG signals suitable for telediagnosis.
- To evaluate the effectiveness of a multivariate autoregressive (MAR) model-based technique for ECG classification.
Main Methods:
- Utilized the MIT-BIH database with 300 samples each of normal sinus rhythm (NSR), atria premature contraction (APC), premature ventricular contraction (PVC), ventricular tachycardia (VT), ventricular fibrillation (VF), and superventricular tachycardia (SVT).
- Developed a MAR model technique to combine signals from two ECG leads for direct classification.
- Employed quadratic discrimination function (QDF) for classification using MAR coefficients and K-L MAR coefficients.
Main Results:
- The proposed MAR modeling based technique achieved high classification accuracy, ranging from 98.3% to 100%.
- The method offers quick and convenient diagnosis, beneficial for telediagnosis applications.
Conclusions:
- The MAR modeling based technique is well-suited for telecardiogram diagnosis.
- Combining two-lead ECG signals yields superior classification results compared to single-lead ECGs.