Quality estimation of the electrocardiogram using cross-correlation among leads

Area of Science:

• Biomedical Engineering
• Signal Processing

Background:

• Quality estimation of electrocardiogram (ECG) signals is crucial for telemedicine, especially when data is collected by untrained personnel.
• Existing methods struggle to accurately differentiate between clinically acceptable and unacceptable ECG records.

Purpose of the Study:

• To develop a novel, simple, and accurate algorithm for 12-lead ECG quality estimation.
• To address the limitations of current ECG quality assessment techniques.

Main Methods:

• The algorithm exploits the cross-covariance matrix structure of different ECG leads.
• Eigenvalues from the ECG signal covariance matrix are used as input for supervised binary classifiers.

Main Results:

• The best performing classifier achieved an accuracy of 0.898 on the test set.
• The algorithm demonstrates significantly lower computational complexity compared to existing methods.

Conclusions:

• The proposed ECG quality estimation algorithm is accurate and efficient.
• Its low complexity makes it suitable for implementation on mobile devices for remote patient monitoring.