Monitoring of fetal heart rate using sharp transition FIR filter. (July 2018)

• Record Type: Journal Article
• Authors: Marchon, Niyan; Naik, Gourish; Pai, Radhakrishna
• Is Part Of: Biomedical signal processing and control. Volume 44(2018)
• Page Start: 191
• Page End: 199
• Publication Date: 2018-07
• Subjects: Finite impulse response -- Linear phase filter -- Fetal heart rate -- QRS detector; Signal processing -- Periodicals; Biomedical engineering -- Periodicals; Signal Processing, Computer-Assisted -- Periodicals; Image Processing, Computer-Assisted -- Periodicals; Biomedical Engineering -- Periodicals; 610.28
• DOI: 10.1016/j.bspc.2018.04.017
• Languages: English
• ISSNs: 1746-8094

Abstract

Highlights: Sharp transition FIR BPF to detect precise fetal and maternal heart rate. The fetal and maternal frequency fiduciary edges form the band edges of the filter. The performance parameters of the FQRS detector improved with increasing filter order. The Gibb's phenomenon was mitigated with proposed slope matching technique. Algorithm generated fetal R-peaks agreed with the reference annotations. Abstract: We propose a novel technique which uses a linear phase sharp transition finite impulse response filter with reduction in band edge ripples due to Gibb's phenomenon. This is accomplished by a slope matching technique. This design can be used for processing a wide variety of signals irrespective of their bandwidth. This paper includes the mathematical design analysis of the band pass filter with slope matching which used fetal frequency fiduciary edges to filter the raw abdominal ECG signals. Our filter design displayed a fairly good magnitude response in all passband, stopband and transition regions with very little computation time and thus is suitable for real time processing of numerous records at a time. The magnitude responses of our finite impulse response filter design was compared to Parks-McClellan algorithm. It was observed that our algorithm showed better results in the passband and stopband at lower filter orders. The peak passband in conventional finite impulse response designs is about 18% which is reduced to around 1.5% with the use of trigonometric functions in the proposed filter model combined with slope equalization technique. The QRS detector effectively extracts the fetal R-peaks to display the fetal heart rate variability which diagnoses the mother-fetus health status. The accuracy of the noninvasive fetal heart rate measurement using our technique is found to be impressive and matches closely with the fetal scalp records. The obtained results show that the sensitivity ranges from 88.36% to 100%, the positive predictive value ranging from 92.31% to 100% and an overall average accuracy recorded at 95.19%.