An Effective and Fast Model for Characterization of Cardiac Arrhythmia and Congestive Heart Failure

Area of Science:

• Biomedical Engineering
• Cardiology
• Signal Processing

Background:

• Cardiac arrhythmia (ARR) and congestive heart failure (CHF) are serious heart conditions with potentially fatal consequences.
• Early and accurate detection of ARR and CHF is crucial for effective patient management and improved outcomes.

Purpose of the Study:

• To develop a fast and accurate automatic detection system for distinguishing between normal sinus (NS) rhythms and ARR.
• To develop a fast and accurate automatic detection system for distinguishing between NS rhythms and CHF.

Main Methods:

• Utilized higher amplitude coefficients (HACs) derived from the discrete cosine transform (DCT) of electrocardiogram (ECG) signals as key features.
• Employed statistical classifiers, specifically the k-nearest neighbors (k-NN) algorithm, for signal discrimination.
• Evaluated system performance using ten-fold cross-validation.

Main Results:

• The DCT effectively compressed ECG signals, and derived HACs demonstrated significant differences between normal sinus (NS) and abnormal rhythms (ARR, CHF).
• The k-NN classifier achieved high accuracy: 97% for ARR vs. NS (99% sensitivity, 90% specificity, 0.63s processing time) and 99% for CHF vs. NS (99.7% sensitivity, 99.2% specificity, 0.27s processing time).
• The DCT-kNN system demonstrated superior accuracy and speed compared to recent related works.

Conclusions:

• DCT-based HACs serve as effective biomarkers for identifying ARR and CHF.
• The developed computer-aided diagnosis (CAD) system is efficient, accurate, and does not require ECG signal pre-processing or segmentation.
• The proposed system shows significant promise for practical implementation in clinical settings.