Key result
QRS-MoMaMeMi-MOA filter cuts detection errors ~89% vs existing filters while lowering FPGA power usage.
Why the study?
ECG signals are contaminated by noise affecting the QRS profile, and existing detection methods yield lower accuracy with higher error rates.
A novel QRS detector using a modified MaMeMi filter optimized with a mayfly optimization algorithm demonstrated higher accuracy, lower detection error rates, and reduced power consumption compared to existing algorithms.
Electrocardiogram (ECG) is considered as the important diagnostic tests in medical field for detecting the cardiac anomalies. But, the ECG signals are polluted with numerous noise from power line intrusion, muscle noise, baseline wander, motion artifacts, low frequency noise signals, high frequency noise signals and T-wave, which automatically affects the QRS profile. The existing method provides the result in lesser accuracy with higher rate of error detection. To overcome these issues, QRS detector using modified maximum mean minimum (MoMaMeMi) filter optimized with mayfly optimization algorithm (QRS-MoMaMeMi-MOA) is proposed in this paper for less computational cost along with resource requirements. The proposed filter design consists of two phases for detecting QRS detector, such as filtering process associated to the enhancement and detection phase. Initially, the ECG data are taken from MIT/BIH arrhythmia dataset (MIT-AD). For eradicating the baseline wander in ECG data, MaMeMi filter is used. For expanding the performance of the modified MaMeMi filter, filter parameters, such as [Formula: see text] and [Formula: see text] are optimized by MOA to accomplish the best values and measure the performance of the whole QRS detector. For high frequency noise suppression in ECG data, the range function, noise subtractors, modified triangular detector are used. Then, heart beat detection can be done with the help of adaptive thresholding technique. The proposed filter design is carried out in MATLAB and implemented on field programmable gate arrays (FPGAs). The proposed QRS-MoMaMeMi-MOA filter design had 0.93%, 0.12% and 0.19% higher accuracy and 89.32%, 50% and 62% low detection error rate, compared to the existing filters, like Kalman filtering based adaptive threshold algorithm for QRS complex detection (QRS-KF-ATA), QRS detection of ECG signal utilizing hybrid derivative with MaMeMi filter by efficiently removing the baseline wander (QRS-HD-MaMeMi), and knowledge-based QRS detection operated by cascade of moving average filters (QRS-CAF). Then, the device utilization of the proposed FPGA implementation of the QRS-MoMaMeMi-MOA filter provides 95.556% and 71.428% lower power usage compared with the existing algorithms, like Kalman filtering based adaptive threshold algorithm for QRS complex detection in FPGA (FPGA-QRS-KF-ATA), and efficient architecture for QRS detection in FPGA utilizing integer Haar wavelet transform (FPGA-QRS-IHWT).
No takes yet. Share an insight, caveat, or question.
Pankaj Kumar (2022) studied Cardiac anomalies (arrhythmia). QRS-MoMaMeMi-MOA filter design vs. Existing filters (QRS-KF-ATA, QRS-HD-MaMeMi, QRS-CAF) and FPGA implementations was evaluated on Accuracy, detection error rate, and power usage. The proposed QRS-MoMaMeMi-MOA filter design achieved up to 0.93% higher accuracy and 89.32% lower detection error rate compared to existing filters, with significantly lower power usage on FPGAs.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: