Application of wavelet transform and Shannon energy on R peak detection algorithm

The R peak detection algorithm is a necessary tool for monitoring and diagnosing the cardiovascular disease. This paper presents the R peak detection algorithm based on continuous wavelet transform (CWT) and Shannon energy. We evaluate the proposed algorithm with the ECG data record 108 from MIT-BIH arrhythmia database. The record 108 results in high detection error rate (DER) in many previous R peak detection algorithms because it consists of muscle noise, baseline shifts, and abnormal heart beats. Results show that the proposed algorithm gives very good DER (0.79%-0.85%) compared to those from previous publications (0.57%-4.71%). We demonstrated that the use of the CWT with a single scaling parameter is capable of removing noises. In addition, we found that Shannon energy cannot improve the DER value but it can highlight the R peak from the low QRS complex in ECG beat leading to the improvement in the robustness of the R peak detection algorithm.