Fast timevarying linear filters for suppression of baseline drift in electrocardiographic signals
 Jiří Kozumplík^{1} and
 Ivo Provazník^{1}Email authorView ORCID ID profile
DOI: 10.1186/s1293801703160
© The Author(s) 2017
Received: 25 October 2016
Accepted: 30 January 2017
Published: 7 February 2017
Abstract
Background
The paper presents a method of linear timevarying filtering, with extremely low computational costs, for the suppression of baseline drift in electrocardiographic (ECG) signals. An ECG signal is not periodic as the length of its heart cycles vary. In order to optimally suppress baseline drift by the use of a linear filter, we need a highpass filter with timevarying cutoff frequency controlled by instant heart rate.
Methods
Realization of the highpass (HP) filter is based on a narrowband lowpass (LP) filter of which output is subtracted from the delayed input. The base of an LP filter is an extremely low computational cost Lynn’s filter with rectangular impulse response. The optimal cutoff frequency of an HP filter for baseline wander suppression is identical to an instantaneous heart rate. Instantaneous length of heart cycles (e.g. RR intervals) are interpolated between QRS complexes to smoothly control cutoff frequency of the HP filter that has been used.
Results and conclusions
We proved that a 0.5 dB decrease in transfer function, at a timevarying cutoff frequency of HP filter controlled by an instant heart rate, is acceptable when related to maximum error due to filtering. Presented in the article are the algorithms that enable the realization of timevariable filters with very low computational costs. We propose fast linear HP filters for the suppression of baseline wander with timevarying cutoff frequencies controlled by instant heart rate. The filters fulfil accepted professional standards and increase the efficiency of the noise suppression.
Keywords
Baseline drift ECG signal Timevarying linear filterBackground
Heart frequency in humans can vary between around 0.67 to 3 Hz (40–180 beats/min) depending on age, sex, stress, health state and a number of other factors. The lower limit of the range can be found in only a small number of physically trained persons in rest, usually in supine position. The upper limit is usually reached only in extreme physical stress. Heart frequency is usually denoted as heart rate (HR) measured by the number of contractions of the heart/min.
However, the ECG signal is not periodical but quasiperiodic (repetitive). Its heart frequency varies due to physiological or pathological reasons, thus it does not allow for the use of ideally set filters. Van Alsté et al. recommend attenuation of −0.5 dB at heart rate. In the case of online processing of longer signals, they recommend −0.5 dB at a fixed cutoff frequency 0.8 Hz [1]. The used filter may not introduce phase distortion. Cardiac electrophysiology societies recommend the use of a linear HP filter with cutoff frequency of 0.67 Hz and 3 dB attenuation. The AHA reports [2] and [3] recommend an amplitude response flat within <−0.5, 0.5> dB, within the range of 1.0–30 Hz. The reports recommend that lowfrequency cutoff be 0.05 Hz to avoid possible distortion of ST segments, but this frequency can be relaxed up to 0.67 Hz (−3 dB) for linear digital filters with zero phase distortion. Abacherli et al. refers in [4] to standards which recommend an HP filter without phase distortion with −3 dB at 0.67 Hz to suppress baseline drift during monitoring. In diagnostic devices, standards recommend attenuation of −0.9 dB, at the same cutoff frequency of 0.67 Hz. Luo et al. refers in [5] to the same values and recommends attenuation not more than 0.5 dB at 1 Hz for stresstest ECG.
All mentioned recommendations and standards only deal with baseline wander suppression by linear filters with the fixed cutoff frequency. However, the main disadvantage of such filtering is that it sets a universal cutoff frequency which causes a lower efficacy in filtering ECG signals with a higher HR. It is generally known that baseline drift spectrum can significantly overlay spectrum of the useful part of ECG signals. Thus, it is desirable to use the highest possible cutoff frequency of the highpass filter but acceptable regarding distortion of the useful part of ECG signals. This has been the reason for development of a number of alternative (nonlinear) filtering methods.
Meyer et al. approximated baseline drift by generating cubic splines from knots in PR intervals where we expect zero line of the ECG signal [6]. The main disadvantage of this method was the necessity of PR interval detection. The method became more efficient with increasing HRs when we obtained higher density of knots, while useful parts of the signal remained uncorrupted.
Thakor et al. used a simple adaptive filter with a constant reference signal and a single weight [7]. However, this filtering method was a source of certain ST segment distortion. Jane et al. [8] described a method based on a cascade of two adaptive filters. The first, simple, adaptive filter with a constant reference input and a single weight represented a simple HP filter with cutoff frequency of about 0.3 Hz. Its output fed a QRS complex detector that produced impulses derived from a rhythm of detected QRS complexes. The impulses entered the reference input of the second adaptive filter with a number of weights equal to a number of samples of the ECG cycle. The filter suppressed signals not correlated with the useful part of the ECG signal. ST segments were not distorted thanks to their direct relation to QRS complexes. A cascade adaptive filter was also used by Laguna et al. [9].
BlancoVelasco et al. exploited methods based on empirical mode decomposition (EMD) [10]. EMD decomposed the signal on a sum of intrinsic mode functions. These were derived directly from an analysed signal and represented a simple oscillatory mode as a counterpart to the simple harmonic function used in Fourier analysis.
Shusterman et al. developed a twostep procedure to correct baseline drift [11]. Firstly, two infinite impulse response filters were applied in a backward and forward direction to avoid phase distortion and obtained ECG signals free of large baseline wander. Secondly, QRS complexes were detected and the rest of the baseline drift was interpolated from determined PQ and TP intervals.
Shin et al. used modified nonlinear methods originally designed for the detrendization of heart rate variability signals to suppress baseline drift [12]. The resulting trend was derived from an estimation of overlapping shorttime trends and was based on a smoothness prior approach.
Fasano et al. applied an approach of baseline wander estimation and its removal in ECG signals based on the approximation of quadratic variation (measure of variability for discrete signals) reduction. Baseline wander was estimated by solving a constrained convex optimization problem where quadratic variation entered as a constraint [13].
Sharma et al. [14] described a method based on Hilbert vibration decomposition. The method considered the first component of the decomposition when applied to an ECG signal that corresponds to baseline wander of the signal.
Zivanovic et al. introduced a baseline wander modelling using loworder polynomials [15].
Hao et al. designed in [16] filtering based on an estimation of baseline wander using the mean–median filter and discrete wavelet transform.
This paper presents an application of a linear filter with a timevarying impulse response. This allows us to fulfil accepted professional standards and to increase the efficiency of the noise suppression. The main aim is to reach a maximum possible attenuation based on an instant HR.
Linear filters provide the correct filtering and it is widely accepted by the biomedical engineering community. At the same time, this filter cannot be considered as optimal due to its variable heart frequency. For more effective suppression of baseline drift, an HP filter with timevarying cutoff frequency related to instant heart frequency should be used.
Sörnmo proposed in [17] and [18] a timevarying filter. In [17], he used a bank of low pass filters with cutoff frequencies 0.5, 0.75, 1.0, 1.25 a 1.5 Hz (at −6 dB), the output of the filters were subtracted from the delayed input signal. Selection of a filter from the bank was based on the length of RR interval, or estimation of drift. Sampling frequency was decimated from 500 to 12.5 Hz to decrease computational cost of the filtering. However, decimation and interpolation caused a higher phase delay of the filter.
We propose a timevarying linear HP filter which does not introduce any phase distortion and excels with an extremely low computational load. The frequency response of the filter is adapted to an instant (interpolated) HR in each signal sample.
Methods
Filter design
Linear phase frequency characteristics beginning at the origin of axes of the phase frequency response are a strict requirement to prevent phase distortion that could decline the ST segment. This requirement can be fulfilled by using a finite impulse response (FIR) linear filter with symmetric impulse response.
Both the recursive and nonrecursive realizations of the cascade of two identical filters G _{ LP }, or H _{ LP } respectively, have a triangular impulse response.
We can easily derive that \(\frac{{f_{c} }}{{f_{0} }} = c = 1.253\).
Fixed filter realization
Presented above was the idea of an optimal HP filter with its impulse response length controlled by the instant length of an ECG cycle. Such a filter has a maximum possible attenuation in a frequency band below f _{ ECG } that can be reached by a linear system of this type. Further, the proposed filter is linear and it has linear phase frequency characteristics that are required for the processing of ECG signals.
Recursive realization of the Lynn’s filter is not an appropriate solution. Although the single pole on a unit circle counteracts with a zero at the same position, there are rounding errors due to division by a large number N ^{ 2 }. This negatively influences filtration.
Nonrecursive realization of the convolution leads to large impulse responses, thus it can be computationally expensive and slow. However, nonrecursive realization can be represented by a cascade of two nonrecursive (movingaverage) filters with a low number of necessary operations per sample interval. The idea is based on the use of a filter H with a rectangular impulse response where we add a new input sample to a sum, then we subtract the oldest input sample and finally divide by a constant N in each sampling interval. Two such filters in a series represent an LP filter with triangular impulse response. The needed HP filter requires one more subtraction.
The realized filter represents a fixed system based on Lynn’s filter with a low number of required operations. Its cutoff frequency can be chosen in advance. However, such a solution is the appropriate basis to design an elegant filter with a timevarying impulse response (and thus timevarying cutoff frequency).
Timevarying impulse response filter realization
An ECG signal is not periodic—the length of its heart cycle(s) vary. To suppress baseline drift optimally, we need an HP filter with timevarying cutoff frequency controlled by an instant HR. The heart frequency in each time instant can only be estimated as we usually measure heart cycles from detected QRS complexes. However, the instant length of heart cycles (e.g. RR intervals) can be interpolated to obtain a signal N _{ RR }(n) to smoothly control the cutoff frequency of the HP filter being used. We use simple 1^{st} order interpolation (by a line).
The impulse response is triangular; its values can be easily derived.
Direct realization of an LP filter with minimum delay
Interpolated instant values of RR intervals are stored in a circular buffer that contains N _{ max } samples corresponding to the longest possible impulse response of the Lynn’s filter.
It is obvious from (17) that the LP filter impulse response has always an odd number of samples.
Direct realization of the LP filter with the triangular impulse response with 2N − 1 samples (see Fig. 4) has no advantage of low computational complexity due to constantly changing all weights of the filter in time.
Realization of an LP filter by a cascade of two Lynn’s filters (knot inside QRS complexes)
We need to use a buffer of input signal samples (input for the first filter) and a buffer of first filters output samples (input for the second filter) besides a buffer of values of RR intervals.
Realization of an LP filter by a cascade of two Lynn’s filters (knots between QRS complexes)
Impulse responses of LP filters can vary in time differently based on how we interpolate RR intervals. Intuitively, we could place knots in the middle between neighbour QRS complexes, instead of placing them into QRS complexes as described in part “Realization of an LP filter by a cascade of two Lynn's filters (knot inside QRS complexes” section of methods.
Results
Computational complexity
The algorithm realizing the final filter provides interpolation of RR intervals and computation of the output sample that contribute to total computational load.
The advantage of the proposed algorithm lies in the extremely fast computation of its response due to simplicity of the used filter. As mentioned in the part Computational complexity in "Results" section, the filter requires 6 additions (or subtractions, respectively) and 2 divisions only to compute one output signal sample. Extremely low computational demands together with the highest possible efficiency of baseline wander suppression regarding to instant heart rate favour the proposed filter against the other timevarying systems presented in “Background” section. One of the most advanced adaptive filter to suppress baseline wander was presented in [17]. However, the used bank of low pass filters requires simultaneous computation of responses of many filters in order to deliver smooth output signal when switching between filters. Further, decimation and interpolation filters are never ideal and they are sources not only of higher phase delay but also of errors.
The algorithms were tested on MA1 set signals from The common standards for electrocardiography (CSE) database [20]. The signals were of 10 s length, sampled at f _{ s } = 500 Hz with quantization step 5 µV (4.8828125 µV). Artificial signals of CSE database were derived from real signals with common noise (without baseline wander) and periodized. The spectrum of each artificial signal is discrete, the first spectral line is located at the signal’s fundamental frequency f _{ ECG }. The signals do not contain any baseline drift. Thus, a linear HP filter with transfer = 1 at f _{ ECG } does not distort the signal. Hence, the MA1 signals were ideal for evaluation of signal distortion due to application of an HP filter with cutoff frequency equal to instant f _{ ECG }. The higher attenuation of the filter allows for more efficient suppression of the drift concerning its spectrum is usually partially overlapped with the lower spectrum of the useful signal.
Discussion
The method introduced for suppression of baseline drift in ECG signals using a linear timevarying HP filter represents optimal linear filtering with regard to setting its cutoff frequency. The cutoff frequency is controlled with instant (interpolated) heart frequency; thus the main disadvantage of a traditional linear filter in this application is the necessity of using a fixed cutoff frequency while the heart frequency physiologically varies. As it is well known, the fixed cutoff frequency is set to a certain value. This is in order to reach a maximum allowed distortion of the useful part of the signal under the worst conditions. Such an approach must be based on the lowest considered heart frequency. However, a more efficient baseline wander suppression requires a higher cutoff frequency in most cases. We proved that a 0.5 dB decrease in transfer function at cutoff frequency is acceptable when related to maximum error due to filtering.
The presented filter was evaluated by testing on a set of ECG signals of standard CSE database. The resulting mean error and standard deviation was low at the level of quantization step of the input signals.
The proposed method depends on reliable detection of QRS complexes. However, a QRS complex detector is a standard basic part of all ECG processing systems and its output is used for preprocessing and delineation of ECG signals. Impact of false positive or false negative detections of heart cycles on the filter efficacy is as follows. When any QRS complex is missed by the detector, only the length of the filter is effected and its cutoff frequency is decreased. Baseline wander removal may be less efficient, the useful part of the processed ECG signal is not distorted. When false QRS complex is detected (false extra heart beat “found”), cutoff frequency of the filter increases by shortening its length. Baseline wander removal is more efficient. However, the useful part of the processed ECG signal is not distorted if we prevent the situation by setting minimum length of the filter to highest expected heart rate. The highest expected rate has to be set according to clinical application: rest electrocardiography, stress test electrocardiography, etc.
Conclusion
A linear timevarying HP filter for optimal suppression of baseline drift was presented. The filter controls its cutoff frequency using an estimation of an instant HR. Such an approach allows us to reach the maximum possible attenuation of the filter while accepted professional standards on maximum allowed distortion are fulfilled. Further, there is no need to set a fixed cutoff frequency that would limit the highest possible frequency of a passband. The filter is suitable for standard ECG devices but also for smart/wearable solutions due to its simplicity and low computational demands.
Abbreviations
 LP:

lowpass
 HP:

highpass
 ECG:

electrocardiography
 HR:

heart rate
 CSE:

Common Standards for Electrocardiography
Declarations
Authors’ contributions
JK codesigned the methods and contributed to their evaluation and interpretation of results. He was a major contributor in writing the manuscript. IP codesigned the methods, wrote software code and lead evaluation and interpretation of results. Both authors read and approved the final manuscript.
Acknowledgements
Not applicable.
Competing interests
The authors declare that they have no competing interests.
Availability of data and materials
The data supporting the conclusions of this article are included within the article. Any queries regarding these data may be directed to the corresponding author.
Funding
This work has been supported by Grant project GACR P102/12/2034.
Open AccessThis article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.
Authors’ Affiliations
References
 Van Alsté JA, Schilder TS. Removal of baseline wander and powerline interference from the RCG by an efficient FIR filter with a reduced number of taps. IEEE T BioMed Eng. 1985;32(12):1052–60.View ArticleGoogle Scholar
 Bailey JJ, Berson AS, Garson A, Horan LG, Macfarlane PW, Mortara DW, Zywietz C. Recommendations for standardization and specifications in automated electrocardiography—bandwidth and digital signalprocessing—a report for healthprofessionals by an ad hoc writing group of the committee on electrocardiography and cardiac electrophysiology of the councilonclinicalcardiology. AmHeartAssoc Circ. 1990;81:730–9.Google Scholar
 Kligfield P, Gettes LS, Bailey JJ, Childers R, Deal BJ, Hancock EW, van Herpen G, Kors JA, Macfarlane P, Mirvis DM, Pahlm O, Rautaharju P, Wagner GS. Recommendations for the standardization and interpretation of the electrocardiogram—Part I: the electrocardiogram and its technology—a scientific statement from the American Heart Association Electrocardiography and Arrhythmias Committee, Council on Clinical Cardiology; the American College of Cardiology Foundation; and the Heart Rhythm Society—Endorsed by the International Society for Computerized Electrocardiology. Circulation. 2007;115:1306–24.Google Scholar
 Abacherli R, Schmid HJ. Meet the challenge of highpass filter and STsegment requirement with a DCcoupled digital electrocardiogram amplifier. J Electrocardiol. 2009;42:574–9.View ArticleGoogle Scholar
 Luo S, Johnston P. A review of electrocardiogram filtering. J Electrocardiol. 2010;43:486–96.View ArticleGoogle Scholar
 Meyer CR, Keiser HN. Electrocardiogram baseline noise estimation and removal using cubicsplines and Statespace computation techniques. Comput Biomed Res. 1977;10:459–70.View ArticleGoogle Scholar
 Thakor NV, Zhu YS. Applications of adaptive filtering to ECG analysis: noise cancellation and arrhythmia detection. IEEE T BioMed Eng. 1991;38(8):785–94.View ArticleGoogle Scholar
 Jane R, Laguna P, Thakor NV, Caminal P. Adaptive baseline wander removal in the ECG: comparative analysis with cubic spline technique. Comput Cardiol. 1992:143–6.
 Laguna P, Jané R, Caminal P. Adaptive filtering of ECG baseline wander. In: 14th Annual International Conference of the IEEE EMBS, conference proceedings. Engineering in Medicine and Biology Society; 1992, p. 509–10.
 BlancoVelasco M, Weng BW, Barner KE. ECG signal denoising and baseline wander correction based on the empirical mode decomposition. Comput Biol Med. 2008;38:1–13.View ArticleGoogle Scholar
 Shusterman V, Shah SI, Beigel A, Anderson KP. Enhancing the precision of ECG baseline correction: selective filtering and removal of residual error. Comput Biomed Res. 2000;33:144–60.View ArticleGoogle Scholar
 Shin SW, Kim KS, Song CG, Lee JW, Kim JH, Jeung GW. Removal of baseline wandering in ECG signal by improved detrending method. BioMed Mater Eng. 2015;26:S1087–93.View ArticleGoogle Scholar
 Fasano A, Villani V. Baseline wander removal for biological signals by qudratic variation reduction. Signal Process. 2014;99:48–57.View ArticleGoogle Scholar
 Sharma H, Sharma KK. Baseline wander removal of ECG signals using Hilbert vibration decomposition. Electron Lett. 2015;51(6):447–9.View ArticleGoogle Scholar
 Zivanovic M, GonzálezIzal M. Simultaneous powerline interference and baseline wander removal from ECG and EMG signals by sinusoidal modeling. Med Eng Phys. 2013;5:1431–41.View ArticleGoogle Scholar
 Hao W, Chen Y, Xin Y. ECG Baseline wander correction by meanmedian filter and discrete wavelet transform. In: 33rd Annual International Conference of the IEEE EMBS, conference proceedings. Engineering in Medicine and Biology Society, Boston, MA, USA; 2011. p. 2712–15.
 Sörnmo L. Timevarying digital filtering of ECG baseline wander. Med Biol Eng Comput. 1993;31:603–8.View ArticleGoogle Scholar
 Sörnmo L, Laguna P. Bioelectrical signal processing in cardiac and neurological applications. Cambridge: Elsevier Academic Press; 2005.Google Scholar
 Lynn PA, Fuerst W. Introductory digital signal processing with computer applications. Hoboken: Wiley; 1992.MATHGoogle Scholar
 Willems J, Arnaud P, van Bemmel JH, Bourdillon PJ, Degani R, Denis B, Harms FMA, Macfarlane PW, Mazzocca G, Meyer J, van Eck HJR, de Medina EOR, Zywietz C. Establishment of a reference library for evaluating computer ECG measurement programs. Comput Biomed Res. 1985;18(5):439–57.View ArticleGoogle Scholar