Baseline Wander and Power-Line Interference Removal from ECG Signals Using Fourier Decomposition Method

Singh, Pushpendra; Srivastava, Ishita; Singhal, Amit; Gupta, Anubha

doi:10.1007/978-981-13-0923-6_3

Cited by 16 publications

(10 citation statements)

References 23 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The new method intro-220 duces periodic median subtraction (PMS) and presents a number of improvements with respect to existing subtraction methods. The validation showed that the filter also performs better than frequency spectrum-based filter, i.e., the current "gold standard", the FDM filter (Singh et al, 2019).…”

Section: Discussionmentioning

confidence: 99%

“…Many filters perform badly at the beginning and end of the signal (edge effects), and are thus not well suited for relatively short recordings. 45 Frequency-based methods involve moving averages technique and IIR notch (kaur Aneja and Singh, 2009), spectrum interpolation (Leske and Dalal, 2019), multi-taper decomposition (Mitra and Bokil, 2007, for multichannel techniques such as EEG) and Fourier decomposition (Singh et al, 2019).…”

Section: Introductionmentioning

confidence: 99%

“…This new, periodic median subtraction (PMS), filter is tested on artificial 70 and experimental data. As a reference the current best performing frequencybased hum filter is used (Singh et al, 2019). Figure 1 illustrates the creation of the filter schematically.…”

Section: Introductionmentioning

confidence: 99%

“…The results of the simulations are summarized in Figure 2 (Singh et al, 2019) and the new PMS hum-filter for simulated data (1000 Hz white noise of amplitude 1, with hum of amplitude 10). At 10 s, there was a step in the data; at 20-22s the noise amplitude was reduced to 8.…”

mentioning

confidence: 99%

“…In the example of Figure 3A As a reference, the error of the FDM filter (Singh et al, 2019) is shown in 195 blue. A comparison between panels G and H reveals a correlation between the amplitude of the signal and the error of the filter.…”

mentioning

confidence: 99%

See 4 more Smart Citations

Robust, general purpose, digital power line hum filter which is free of deformations and which can be applied to large transients

Lussanet

2019

Preprint

View full text Add to dashboard Cite

Power line interference ("hum noise") is a common source of noise in recorded biological data. It has a highly constant frequency with harmonics and little variation in amplitude and wave shape. In contrast to stochastic noise it is phasic, i.e., the phase relation remains constant across long intervals. In digital recordings, the measurement frequency is typically very close to a 5 multiple of the hum frequency (50 or 60 Hz). Common filters introduce various kinds of distortions and errors. The here proposed subtraction method makes use of the specific properties of power line hum. For this, it computes a moving estimate of the hum noise by taking the periodic median of the high-pass filtered signal. The resulting periodic median subtraction (PMS) filter reliably 10 removes hum of any harmonic composition, even if the ground frequency is lacking. The filter is completely free of border artifacts. It does not introduce distortions, even if around sharp transients such as in force plate recordings of jumping. The filter is validated on recorded and artificial data. The errors are quantified. The results also show that the errors of hum filters generally 15 increase with the high-frequency content of the data. Thus, the removal of hum from a force plate recording generally gives better results than from an EMG recoding. Compared to other filters, the errors of the PMS filter are generally lower than the best hum filters currently known.

show abstract

Section: Discussionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

mentioning

confidence: 99%

mentioning

confidence: 99%

See 3 more Smart Citations