Electrocardiograph (ECG) R-wave detection from abdominal ECGs presents significant challenges due to signal distortion and interference. This study introduces a novel approach using K-Mean Clustering technique to detect ECG R waves from ECG signal. Leveraging the quad periodicity of R waves, we developed an algorithm to identify and correct misaligned and missed R wave detection in the RR time series.
Our detection process comprises two stages. Initially, we detect ECG R waves that do not overlap with QRS complexes. The algorithm then corrects misaligned and missed R waves while predicting approximate regions where ECG R waves overlap with QRS. In the second stage, we detect overlapping ECG R waves within these predicted regions.
We construct QRS complex from non-overlapping detected complexes using K-Mean Clustering.
The algorithm then precisely locates overlapping ECG R waves by finding the optimal correlation between the actual QRS and the superposition of QRS within the predicted regions. The results suggest that our method significantly enhances the accurate extraction of ECG R waves, making it potentially valuable for both clinical and commercial applications in monitoring.