Motion Artifact Reduction from Finger

Photoplethysmogram Using Discrete

Wavelet Transform

Anita Biswas, Monalisa Singha Roy and Rajarshi Gupta

Abstract In this work, discrete wavelet transform was used to remove the effect
of motion artifact on the Photoplethysmogram (PPG) signal obtained at fingertip.
Clean PPG signal and motion data (one direction) were collected from 40 healthy
volunteers at 14-bit resolution using NI 6009 DAQ card, and synthetic noisy signal
was generated by addition. The noisy signal was first decomposed into a specific
number of levels to obtain different frequency bands. Then, soft thresholding method
was used to remove the noisy components. Different wavelet functions (Daubechies,
Symlet, Coiflet) and soft thresholding methods (‘rigrsure,’ ‘heursure,’ ‘sqtwolog,’
etc.) were used to denoise the corrupted PPG signal. A comparative study was made
between all of these methods by calculating performance measures such as signal-
to-noise ratio improvement, mean square error, and percentage noise retention. The
mother wavelet ‘Db6’ and ‘rigrsure’ soft thresholding method showed the best result.

Keywords Photoplethysmogram · Motion artifact · Discrete wavelet transform

Thresholding · Denoising

1 Introduction

During the last two decades, the Photoplethysmography (PPG) technique has become
an emerging area of research for its wide application in clinical physiological mon-
itoring [1]. PPG is a simple, noninvasive optoelectronic method to detect the blood
volume changes on the peripheral body parts. The PPG sensor detects change in
light intensity due to blood volume changes as a result of cardiac systole and diastole

A. Biswas (B) · M. Singha Roy · R. Gupta

Department of Applied Physics, University of Calcutta, Kolkata, India
e-mail: [email protected]
M. Singha Roy
e-mail: [email protected]
R. Gupta
e-mail: [email protected]
© Springer Nature Singapore Pte Ltd. 2019 89
S. Bhattacharyya et al. (eds.), Recent Trends in Signal and Image Processing,
Advances in Intelligent Systems and Computing 727,
https://doi.org/10.1007/978-981-10-8863-6_10
90 A. Biswas et al.

[2, 3] using near red and infrared wavelength light source and detector. It is found
that PPG waveform can provide a lot of valuable clinical information such as blood
pressure, heart rate, respiratory rate, and cardiac output [4–6].
However, for ubiquitous and real-time measurement system, PPG signals are eas-
ily prone to motion artifacts (MA) which could lead to inaccurate interpretation of
the PPG waveform. The common sources of errors are ambient light at the photode-
tector, poor contact of the PPG sensor and the skin, patient’s movement, respiration,
etc. [7, 8]. The motion artifacts which mix the raw PPG signal that resides in the same
frequency range with the actual PPG data (0–2 Hz). In particular, the motion artifact
reduction is the most challenging issue. Over the last few years, many researchers
focused on the area of motion artifacts removal technique from PPG signal.
The adaptive filtering technique used in real-time application is a well-known
method for motion artifacts removal. But it requires an additional sensor to provide
reference signal which is the major drawback of this method [9, 10]. AS-LMS adap-
tive filter eliminates this drawback by generating the reference signal internally from
the MA-corrupted PPG itself. Principal Component Analysis (PCA) is a popular
signal processing tool that can be effectively used for noise elimination, signal sepa-
ration, and feature extraction. By selecting suitable number of principal components,
the dimension of the actual data can be reduced but also retaining maximum possible
information of the actual data [11]. In [12], the frequency characteristics of the PPG
signal are analyzed to characterize the clean PPG and motion artifact. It is observed
that the raw PPG has fundamental, second and third harmonic structure caused by
arterial blood flow and component caused by motion artifact. The raw signal is divided
into different groups depending on the relative location of these frequency compo-
nents. The motion artifact is composed of single frequency component. Another
frequency domain analysis method named Empirical Mode Decomposition (EMD),
that empirically extracts the oscillatory behavior from the signal and decompose a
time series data into a number of intrinsic mode function (IMFs). In [13], Hilbert
Hung Transform is used to generate instantaneous frequency for each IMF. The PPG
signal is divided into six scales of different frequency bands. According to mean
frequency, the scales 1–3 match with clean PPG signal and its harmonics and scales
4–6 correlate with the motion artifacts. An improved algorithm is presented in [14]
which combine PCA with the EMD method to generate a smaller group of orthogonal
variables. A fewer number of principal components are able to represent the clinical
information of the signal, and this number is always less than the number of IMFs.
In all of these frequency domain methods, the Fast Fourier Transform (FFT) is used
to convert the signal from time domain to frequency domain. Wavelet transform is
a powerful tool that has been widely used in ECG signal denoising. However, its
application in PPG preprocessing is comparatively fewer [15, 16]. In [16], appli-
cation of wavelet transform for MA reduction is described. However, the authors
only evaluated the systolic peak amplitude variations and spectrum analysis and no
detailed study on error parameters was done.
In this work, we used synthetic noisy PPG by imposing three different noise
levels and decompose the signal using DWT. The performance of different mother
wavelets and different thresholding rules on signal reconstruction quality was also
Motion Artifact Reduction from Finger Photoplethysmogram … 91

investigated. Additionally, a detailed study on error parameters was done to evaluate

the effectiveness of the proposed method.

2 Methodology

2.1 Data Collection Protocol

The PPG signal was recorded from forty male and female volunteers in the age
group 20–45 years. Prior to data collection, they were requested to sit in a relaxed
condition. A transmission type PPG sensor was placed on the left hand index finger,
kept in stationary condition, while an accelerometer was attached on the right hand
index finger. For our application, we considered only horizontal movement of the
hand (X-axis). The right hand was moved from elbow at three predefined levels of
motion (level, i.e., 10, 15, and 20 cm), each non-periodic. For every subject, data
collection was done for 30 s. at 60 Hz using NI6009 DAQ device. The motion data
was added with the clean PPG signal in different proportions to generate synthetic
noisy signal. The PPG data was stored in calibrated voltage format against sampling
time in <*.txt> formatted file.

2.2 Wavelet Denoising Approach

Because the PPG signal is non-stationary in nature, the discrete wavelet transform
(DWT) is an appropriate signal processing tool for analyzing the time–amplitude
information from a signal. It analyzes the signal at different resolutions by decom-
posing it into various successive frequency bands. The signal is passed through a
series of high-pass filter and low-pass filter.
Since in DWT the signal is a discrete time function, it can be represented as x[n],
where n is an integer. The signal is passed through a half band digital low-pass filter
with impulse response h[n]. The convolution operation in discrete time is defined as
follows:
∞

x[n] ∗ h[n]  x[k] · h[n − k] (1)
k−∞

The half band low-pass filter removes all the frequencies that are above half of
the highest frequency in the signal.
The denoising algorithm is given below:
a. Decomposition—The input signal was decomposed up to level 5 (to get a fre-
quency level around 1.7 Hz, that of PPG signal) by choosing appropriate wavelet
92 A. Biswas et al.

Fig. 1 Wavelet 60 Hz sampling PPG data array

decomposition structure of
the PPG data
Level 1 AC: 0-30 Hz DC: 30-60 Hz

Level 2 AC: 0-15 Hz DC: 15-30 Hz

Level 3 AC: 0-7.5 Hz DC: 7.5-15 Hz

Level 4 AC: 0-3.75 Hz DC: 3.75-7.5 Hz

Level 5 AC: 0-1.575 Hz DC: 1.875-3.75 Hz

function using Daubechies6 (Db6) as the basis wavelet. The decomposition struc-
ture is shown in Fig. 1.
b. Thresholding—In ‘soft’ thresholding, the coefficients below a threshold value
are set to zero, while the magnitude of others is truncated by the same value.
Mathematically, it is expressed as

d̂ j, k  sign(d j, k )|d j, k − t|, i f |d j, k |> t

 0, other wise (2)

where d j, k represents the kth wavelet coefficient in jth level, d̂ j, k is the thresh-
olded value, and t is the threshold. There are various soft thresholding rules, like
fixed threshold (‘sqtwolog’), ‘rigrsure’ based on Stein’s unbiased risk estimate,
‘heursure’ a combination of first two rules, and ‘minimax’ based on mean square
error [17].
The soft thresholding rules were selected for our application, and it is applied in
each level (A5, D5, …, D1) for selective truncation of the wavelet coefficients.
The thresholded coefficients at different levels are shown in Fig. 2.
c. Reconstruction—After addition of the selected coefficients at each level, the
reconstruction was done by the inverse discrete wavelet transform (IDWT).

3 Results and Discussions

For evaluating the denoising performance, the objective assessment was done by
computing the following metrics which are common for any denoising operation.
Motion Artifact Reduction from Finger Photoplethysmogram … 93

Fig. 2 Soft thresholding applied at different levels using Db6


M S E (Mean Squar e Err or )  1
N
[x(i) − x̃(i)]2
i

M AE (Maximum Absolute Err or )  max[x(i) − x̃(i)]

Pds −Pcs (1)
P N R (Per centage N oise r etention)  Pcs
× 100
 x 2 (i)

S N Ri (Signal to noise ratio impr ovement)  10 × log

i n 2 (i)
 x 2 (i)
i [x(i)−x̃(i)]2

where x = clean signal, x̃ = reconstructed signal, Pcs = power of clean signal, N =

number of samples, Pds = power of reconstructed signal, n = noisy signal, and P =
10 × log |x(i)|2 . Among these, MSE provides the global error figure, since errors
can be positive or negative. PNR provides the residual noise that is retained in the
preprocessed signal. SNRi actually denotes the ratio of input and output SNR (to the
preprocessing algorithm), where x(i)−x̃(i) is the residual noise. MAE is the maxi-
mum absolute sample-to-sample error after reconstruction and local residual noise.
Table 1 shows these parameters for six arbitrarily chosen healthy volunteers’ data for
three levels of motion using ‘rigrsure’ thresholding rule. Among the parameters, we
observed that the MAE, averaged over the six volunteers, remains almost unchanged,
while MSE and PNR increase slightly with enhanced noise, as expected. Computed
over all volunteers’ data, for level 1 noise, the average MSE, PNR, MAE, and SNR
are 0.104, 2.05, 0.89 mV, and 14.12, respectively. When the noise level increases to
level 2, these values are 0.143, 2.44, 0.983 mV, and 11.45, respectively. For highest
level 3, these values are 0.230, 2.855, 0.910 mV, and 12.54, respectively. In sum-
mary, with noise-level increase, the change in MSE, PNR, MAE, and SNR are 37%,
19%, 10%, and 18%, respectively, from level 1 to 2 and 60%, 16%, 7%, and 9.51%,
respectively, for level 2 to level 3. Among these, the MAE is least changed, which
means the local distortion could be kept lower in the denoising process.
94 A. Biswas et al.

Table 1 Denoising parameters for db6 mother wavelet function

Volunteers’ detail (age, Level of MSE PNR MAE (mV) SNR
M/F, weight in Kg, noise
height in cm)
#1 (27, M, 59, 183) 1st 0.09 1.16 0.92 11.26
2nd 0.19 2.76 1.12 8.29
3rd 0.21 3.09 1.01 5.08
#2 (43, M, 72, 173) 1st 0.18 2.85 1.05 7.84
2nd 0.20 2.98 1.16 7.26
3rd 0.16 2.24 1.17 7.43
#3 (29, F, 55, 167) 1st 0.23 3.40 1.02 5.06
2nd 0.30 3.77 1.47 4.33
3rd 0.43 6.89 0 2.43
#4 (24, M, 65, 167) 1st 0.06 1.83 0.65 15.92
2nd 0.06 1.71 0.38 14.27
3rd 0.11 2.77 0.77 10.39
#5 (24, M, 68, 170) 1st 0.14 2.58 1.01 33.07
2nd 0.10 3.45 0.60 20.54
3rd 0.11 0.33 1.07 19.33
#6 (24, F, 60, 157) 1st 0.07 0.79 0.97 11.27
2nd 0.09 0.27 1.19 15.03
3rd 0.18 1.68 1.44 36.59

The denoising performance is also pictorially represented using Fig. 3 for volun-
teer 1. The upper panel shows the original and reconstructed signal superimposed,
with the residual, i.e., the sample-to-sample error in lowest panel.
Table 2 shows the denoising parameter values for different noise levels using
various thresholding rules for volunteer #1 using Db6 wavelet. Here, we studied
volunteer 1 data for different levels of noise. It is clearly observable from Table 1
and 2 that MSE value is minimum for ‘rigrsure’ thresholding rule, while ‘heursure’
provides the least performance. For local distortion assessment, the fixed thresholding
provides worst performance, but its SNR improvement is very good (Fig. 4).
In the next phase, four clinical features: systolic peak height, diastolic peak height,
peak-to-peak time, and crest time of the reconstructed data were compared with the
original signal to assess the percentage deviation for five volunteers as in Table 1.
The used noise levels are level 1 at a proportion of 100 and 200%. The deviations
are shown in Fig. 5. It is observed that for volunteer #1, there is no variation of peak-
to-peak time event for 200% of noise, while the other features slightly increase.
For volunteer #2, the crest time changes by 25% in denoised data. As a general
observation, using level 1 noise at 200% proportion, most of the features could be
retained within 15% of their original values.
Motion Artifact Reduction from Finger Photoplethysmogram … 95

Fig. 3 Subjective assessment of denoising performance

Table 2 Denoising parameters for different thresholding rules (Volunteer #1, Db6)
Level of noise Threshold MSE PNR MAE (mV) SNRi
rule
1st sqtwolog 0.77 2.78 2.21 2.21
heursure 0.92 2.09 0.92 0.92
minimax 0.39 1.53 1.53 1.53
2nd sqtwolog 0.45 0.62 1.77 1.77
heursure 0.19 2.33 1.12 1.12
minimax 0.24 0.86 1.19 1.19
3rd sqtwolog 0.48 1.07 1.50 1.50
heursure 0.21 2.60 1.02 1.02
minimax 0.24 1.02 1.21 1.21

Table 3 shows the different denoising parameters for various mother wavelet
functions for volunteer #1 and noise level 1. It is observed from Tables 1 and 3 that
the denoising parameters show better result for Db6 mother wavelet. In terms of MSE,
the Symlets 2, 3, and 4 provide the worst performance. The denoising performance
is also graphically shown in Fig. 5. In the published literature using DWT-based PPG
denoising [15, 16], no quantitative performance is provided. Only [9] reports SNR of
1.732 dB with horizontal motion of PPG sensor and [11] reports SNR improvement
from 0.078 to 0.318 dB. The proposed method achieved an average SNRi of 7.40 dB.
96 A. Biswas et al.

Fig. 4 Deviation of clinical signatures in the reconstructed signal for five volunteers

Table 3 Denoising parameters for different mother wavelet functions (volunteer #1)
Mother wavelets MSE PNR MAE (mV) SNRi
Db5 1.27 4.62 2.54 0.25
sym1 3.50 2.10 5.40 7.72
sym2 4.42 0.09 4.59 25.12
sym3 6.10 9.57 4.89 7.58
sym4 2.05 4.76 6.15 11.16
sym5 5.01 4.69 4.60 0.52
coif1 1.17 2.84 3.11 8.82
coif2 2.19 6.50 5.18 11.16
coif3 1.76 4.24 3.43 4.02
coif4 2.37 5.75 4.15 1.98
coif5 2.38 3.41 4.13 3.09

4 Conclusion

In this paper, we present the application of wavelet transform for reduction of motion
artifact from PPG signal. Among all wavelets used in the study, the Daubechies
wavelet provided best result using ‘rigrsure’ soft thresholding rule. Using three
Motion Artifact Reduction from Finger Photoplethysmogram … 97

Fig. 5 Denoising performance of different mother wavelet functions

different intensity levels of noise, the reconstructed waveform could retain the most
of clinical signatures within 15% of their respective values. We have also made a
comparative study between different denoising parameters by using other mother
wavelet functions.

References

1. Allen JJ (2007) Photoplethysmography and its application in clinical physiological measure-

ment. Physiol Meas 28(3):1–39
2. Kyriacou PA (2006) Pulse oximetry in the oesophagus. Physiol Meas 27(1):1–35
3. Hertzman AB (1938) The blood supply of various skin areas as estimated by the photoelectric
plethysmograph. Am J Physiol 124:328–340
4. Hee X, Goubran RA, Liu XP (2014) Secondary peak detection of PPG signal for continuous
cuffless arterial blood pressure measurement. IEEE Trans Instrum Meas 63(6):1431–1439
5. Gil E, Orini M, Bailon R (2010) Photoplethysmography pulse rate variability as a surro-
gate measurement of heart rate variability during non-stationary conditions. Physiol Meas
31(9):1271–1290
6. Nakajima K, Tamura T, Miike H (1996) Monitoring of heart and respiratory rates by photo-
plethysmography using a digital filtering technique. Med Eng Phys 18(5):365–372 Elsevier
7. Sweeney KT, Ward TE, MacLoone SF (2012) Artifact removal in physiological signals-
practices and possibilities. IEEE Trans Inf Technol Biomed 16(3):488–500
98 A. Biswas et al.

8. Han H, Kim MJ, Kim J (2007) Development of real time motion artifact reduction algorithm
for a wearable photoplethysmography. In: IEEE EMBS conference France, pp 1538–1541
9. Raguram M, Madhav KV, Krishna EH, Reddy KN, Reddy KA (2011) On the performance of
time varying step–size least mean squares (TVS-LMS) adaptive filter for MA reduction from
PPG signals. In: ICCSP Conference, pp 431–435
10. Raguram M, Madhav KV, Krishna EH (2012) A novel approach for motion artifact reduction
in PPG signals based on AS-LMS adaptive filter. IEEE Trans Instrum Meas 61(5):1445–1457
11. Castells F, Laguna P, Sornmo L (2007) Principal component analysis in ECG signal processing.
Eurasip J Adv Signal Process 2007(1):98–98
12. Cho JM, Sung YK, Shin KW (2012) A preliminary study on photoplethysmogram (PPG) signal
analysis for reduction of motion artifact in frequency domain. In: IEEE EMBS conference on
IECBES, pp 28–33
13. Wang Q, Yang P (2010) Artifacts reduction based on empirical mode decomposition (EMD)
in photoplethysmography for pulse rate detection. In: IEEE EMBC conference, pp 959–962
14. Pinheiro E, Postolache O, Girao P (2012) Empirical mode decomposition and principal compo-
nent analysis implementation in processing non-invasive cardiovascular signals. Measurement
45(2):175–181. Elsevier
15. Lee CM, Zhang YT (2003) Reduction of motion artifacts from photoplethysmographic record-
ings using a wavelet denoising approach. In: IEEE EMBS Asian-Pacific conference on biomed-
ical engineering, pp 194–195
16. Raghuram M, Madhav KV, Krishna EH (2010) On the performance of wavelets in reducing
motion artifacts from photoplethysmographic signals. In: IEEE international conference on
ICBBE, China, pp 121–128
17. Wavelet toolbox user guide MATLAB® . www.mathworks.com