Non-invasive cuff-less blood pressure estimation using a hybrid deep learning model

Abstract

Conventional blood pressure (BP) measurement methods have different drawbacks such as being invasive, cuff-based or requiring manual operations. There is significant interest in the development of non-invasive, cuff-less and continual BP measurement based on physiological measurement. However, in these methods, extracting features from signals is challenging in the presence of noise or signal distortion. When using machine learning, errors in feature extraction result in errors in BP estimation, therefore, this study explores the use of raw signals as a direct input to a deep learning model. To enable comparison with the traditional machine learning models which use features from the photoplethysmogram and electrocardiogram, a hybrid deep learning model that utilises both raw signals and physical characteristics (age, height, weight and gender) is developed. This hybrid model performs best in terms of both diastolic BP (DBP) and systolic BP (SBP) with the mean absolute error being 3.23 ± 4.75 mmHg and 4.43 ± 6.09 mmHg respectively. DBP and SBP meet the Grade A and Grade B performance requirements of the British Hypertension Society respectively.

Full text

Vol.:(0123456789)
Optical and Quantum Electronics (2021) 53:93
https://doi.org/10.1007/s11082-020-02667-0
1 3
Non‑invasive cuff‑less blood pressure estimation using
ahybrid deep learning model
SenYang1· YapingZhang1· Siu‑YeungCho1· RicardoCorreia2· StephenP.Morgan2
Received: 8 July 2020 / Accepted: 14 December 2020 / Published online: 25 January 2021
© The Author(s) 2021
Abstract
Conventional blood pressure (BP) measurement methods have different drawbacks such as
being invasive, cuff-based or requiring manual operations. There is significant interest in
the development of non-invasive, cuff-less and continual BP measurement based on physi-
ological measurement. However, in these methods, extracting features from signals is chal-
lenging in the presence of noise or signal distortion. When using machine learning, errors
in feature extraction result in errors in BP estimation, therefore, this study explores the use
of raw signals as a direct input to a deep learning model. To enable comparison with the
traditional machine learning models which use features from the photoplethysmogram and
electrocardiogram, a hybrid deep learning model that utilises both raw signals and physical
characteristics (age, height, weight and gender) is developed. This hybrid model performs
best in terms of both diastolic BP (DBP) and systolic BP (SBP) with the mean absolute
error being 3.23 ± 4.75 mmHg and 4.43 ± 6.09 mmHg respectively. DBP and SBP meet
the Grade A and Grade B performance requirements of the British Hypertension Society
respectively.
Keywords Blood pressure (BP)· Cuff-less· Photoplethysmogram (PPG)·
Electrocardiogram (ECG)· Deep learning
1 Introduction
Blood pressure (BP) is one of the most important and commonly measured clinical param-
eters and accurate measurement is crucial for therapeutic decisions. The World Health
Organization (WHO) estimates that 1.13 billion people worldwide have hypertension
which is a major cause of premature death. However, fewer than 1 in 5 people with hyper-
tension have the problem under control (World Health Organisation 2019). One of the
global targets for noncommunicable diseases is to reduce the prevalence of hypertension by
* Stephen P. Morgan
steve.morgan@nottingham.ac.uk
1 International Doctoral Innovation Centre, University ofNottingham Ningbo China, 199 Taikang
East Road, Ningbo, China
2 Optics andPhotonics Research Group, University ofNottingham, Nottingham, UK
Content courtesy of Springer Nature, terms of use apply. Rights reserved.

S.Yang et al.
1 3
93 Page 2 of 20
25% by 2050 (baseline 2010). Regular BP monitoring is thus essential for prevention and
control in the general population and for hypertensive patients.
Despite its importance, the existing non-invasive regular BP measure methods have
downsides that can be ascribed to their measuring devices. The most popular cuff-based
BP measurement requires user to follow protocols to obtain accurate BP values. Some cuff-
based devices, e.g., mercury sphygmomanometer, require frequent calibration. Other dis-
advantages include movement artefacts during physical activity (Ogedegbe and Pickering
2010) and discomfort during cuff inflation. In addition, for some people, the act of going
to the doctor triggers a response making their BP soar which clinicians recognize as white-
coat syndrome (Stergiou etal. 2018).
Due to the aforementioned factors, developing a cuff-less, continual or near real-time
periodic, robust, comfortable, and wearable BP measurements system is desirable. Using
physiological signals to conduct non-invasive and cuff-less BP measurement emerged in
the past decade (Li etal. 2018; Liang et al. 2018a; Sharifi etal. 2019; Yoon etal. 2009).
Two typical physiological signals used in BP estimation are the photoplethysmogram
(PPG) and electrocardiogram (ECG). Pulse arrival time (PAT) is the time interval between
the R-wave peak of the ECG and the systolic peak of the PPG. When the PAT is longer, it
indicates a lower BP, while a shorter PAT indicates a higher BP, but the precise relation-
ship is uncertain due to the complexity of the cardiovascular system. This method requires
a calibration protocol for stepwise increases in BP and several simultaneous measurements
of ECG, PPG and a reference method (e.g. a mercury sphygmomanometer). Furthermore,
individual calibrations are often needed to increase accuracy. Therefore, it is a challenge to
use PAT for BP measurement under clinical conditions (Hennig and Patzak 2013).
Recently, there has been growing interest in cuff-less and non-invasive BP estimation
using machine learning algorithms with the PPG and ECG (Chen et al. 2019; Kachuee
etal. 2017; Mousavi etal. 2019; Ribas Ripoll and Vellido 2019; Rundo etal. 2018). Most
of the studies extracted specific features in the time domain or frequency domain and their
results reveal the high correlation of these features with BP (Elgendi etal. 2019; Kachuee
etal. 2017; Tanveer and Hasan 2019; Wang etal. 2018). The two main challenges with
these approaches are the need for considerable signal processing and extraction of features
associated with physiological signals.
These drawbacks, alongside the emerging methods of using raw signals as inputs into
deep learning for different purposes (Gotlibovych etal. 2018; Slapničar etal. 2019), have
motivated us to investigate this approach for non-invasive BP measurement. To ensure the
high quality of the data, this research conducted a series of measurements on 45 partici-
pants to obtain a database of 315 records, each containing PPG, ECG, BP values and cor-
responding participant’s physical characteristics (i.e., age, height, weight and gender). This
is suitable for the investigation of the use of deep learning with raw signals and physical
characteristics for BP measurement for the first time to the authors’ knowledge. Moreover,
another objective is to compare the accuracy of predictions between traditional machine
learning methods and the novel hybrid deep learning model.
The novelty of this study is threefold. Firstly, although there have been attempts pre-
dicting BP values using deep learning methods, they rely on the use of physiological sig-
nals. To date, no one has tried to use both physiological signals and physical characteristics
as inputs in a deep learning structure. This study presents the first attempt in this regard
by devising a novel hybrid deep learning model. Secondly, this study provides a compre-
hensive comparison not only between traditional machine learning methods and hybrid
deep learning models, but also between hybrid deep learning models with different struc-
tures. Thirdly, the methods used to collect data to predict BP are simple and replicable.
Content courtesy of Springer Nature, terms of use apply. Rights reserved.

Non‑invasive cuff‑less blood pressure estimation using ahybrid…
1 3
Page 3 of 20 93
Combined with the automatic nature of the hybrid deep learning model, it largely reduces
the complexity for the end user and has the potential of large-scale implementation.
The paper is organized as follows: Sect. 2 explains the experimental data processing
procedures and all the algorithms used in this work. Section3 demonstrate the experimen-
tal design for this study. Section4 presents the results and finally, Sect. 5 concludes and
discusses the research.
2 Methods
Before application of machine learning algorithms, the ECG and PPG are pre-processed
and features are extracted. Several popular machine learning algorithms are then applied to
estimate BP from signal features and physical characteristics. Unlike traditional methods,
the proposed deep learning method does not require feature extraction and key information
contained in the raw data are automatically extracted by the deep learning network by self-
learning. Data acquisition will be described in Sect.3.
2.1 Data pre‑processing
The acquired PPG signal is processed by a Chebyshev II bandpass filter with the lower
and upper cut off frequencies of 0.5 and 10Hz respectively in order to reduce noise within
the raw PPG signal (Liang etal. 2018b). For the ECG signal, baseline drift and high fre-
quency noise are removed using a Butterworth bandpass filter with lower and upper cut
off frequencies of 0.5 and 40Hz respectively (Shin etal. 2010). Afterwards, the PPG and
ECG signals are normalized and their peaks in each period are obtained. The most stable
segments are chosen from both signals by a calculation of the highest cross-correlation
coefficient between periods which is defined by neighbouring peaks (Kachuee etal. 2015).
2.2 Feature extraction
2.2.1 PPG
In the literature, morphological features from PPG and complexity features from ECG are
often used to predict BP (Elgendi 2012; Kachuee etal. 2017; Simjanoska etal. 2018; Yang
etal. 2020). There are more than twenty features that can be extracted from a PPG signal
and its first and second derivatives (Elgendi 2012). Twelve of them are selected and used
for further estimation in this research. A PPG signal with labelled features is displayed in
Fig.1a and a PPG and its second derivative signals are shown in Fig.1b. A summary of
used features is listed in Table1.
2.2.2 ECG
The extracted and used features in this research are listed in Table2. Most of these features
from ECG signals are obtained from complexity analysis, except heart rate which is calcu-
lated from the measurement of the peak-to-peak time interval of the ECG signals.
Content courtesy of Springer Nature, terms of use apply. Rights reserved.

S.Yang et al.
1 3
93 Page 4 of 20
2.2.3 Pulse arrival time (PAT)
PAT is extracted and applied as one of the features in this study. PAT is defined as the time
interval between the electrical activation of the heart and arrival of the pulse pressure at a
distal point measured as the time between the peaks of PPG and ECG (Chan etal. 2019). It
Fig. 1 a A PPG signal with labelled features, adapted from (Kachuee etal. 2017). b Measured PPG sig-
nal (upper) and its second derivative (lower), indicating systolic and diastolic peaks, a-wave and b-wave.
Adapted from (Elgendi 2012)
Table 1 Extracted PPG features used in the study
Feature no. Feature name Descriptions Figure
1 Systolic amplitude (Chua and
Heneghan 2006) (Chua etal. 2010)
Systolic peak Figure1a
2 Pulse width Width
3 Peak to peak interval Time difference two successive systolic
peaks
4 Inflection point (Millasseau etal. 2002) Used to replace diastolic point
5 Augmentation index (Elgendi 2012)
AuI
=
x
y
6 Large arterial stiffness index Inversely related to the time interval ΔT
7 S1 Areas under the PPG signal
8 S2
9 S3
10 S4
11 Crest time (Alty etal. 2007) CT
12 Ratio of b/a (Baek etal. 2007) From 2nd derivative Figure1b
Content courtesy of Springer Nature, terms of use apply. Rights reserved.

Non‑invasive cuff‑less blood pressure estimation using ahybrid…
1 3
Page 5 of 20 93
includes the pre-ejection period, which is the time it takes for blood to leave the heart after
the heart’s electrical impulse.
2.3 Traditional machine learning methods
Several commonly used machine learning methods are used in this study to evaluate the
effectiveness of different methods in predicting BP using features extracted from PPG and
ECG signals and physical characteristics.
LASSO (least absolute shrinkage and selection operator) is a linear model with L1 prior
as a regularizer (Friedman etal. 2010). As a large number of features are used to predict
BP, it is important to add a regularization term in linear models to help with the variable
selection. LASSO is able to perform both variable selection and regularization, leading to
increase of prediction accuracy. The amount of regularization is controlled by α, the coeffi-
cient of the L1 term, and it can be determined experimentally using cross-validation during
the training process. In this study, fivefold cross-validation is used to select α.
Support Vector Regression (SVR) is a popular machine learning model and has been
proven to be an effective tool in real-value function estimation (Drucker etal. 1996). SVR
uses a symmetrical loss function and errors with absolute values that are smaller than a
certain threshold are ignored. As a result, the model produced by SVR depends only on a
subset of the training data. A fivefold cross-validated grid-search is used to search for the
optimal values for several important parameters, including kernel type (linear, polynomial,
radial basis function), kernel coefficient (0.1, 0.01, 0.001, 0.0001), regularization param-
eter (1, 0.1, 0.01, 0.001, 0.0001) and epsilon-tube (0.1, 1, 5, 10, 20) which specifies the
tolerance level.
AdaBoost, which is short for Adaptive Boosting, is an ensemble method and can be used
to fit a sequence of weak learners (other types of learning algorithms) to improve perfor-
mance (Drucker 1997). The final output is a combination of a weighted sum of predictions
generated by these weak learners. A commonly used weak learner, a decision tree regres-
sor is adopted in this study. A fivefold cross-validated grid-search is further used to search
for the optimal values of the number of iterations (5, 50, 500), learning rate (1, 0.1, 0.01,
0.001, 0.0001) and loss function (linear, square, exponential).
Random forest (RF) is another ensemble method that constructs a number of deci-
sion trees built from samples drawn with replacement (Breiman 2001). With the added
Table 2 Extracted features from
ECG (Yang etal. 2020)Feature name Number of
features
Autoregressive (AR) model coefficients of order 8 8
Multifractal wavelet leader
Second cumulant of scaling exponents 1
Holder exponents 1
Shannon Entropy (SE) values for the maximal overlap
discrete wavelet packet transform at level 5
32
Hjorth parameters
Signal mobility 1
Signal complexity 1
Heart rate 1
Content courtesy of Springer Nature, terms of use apply. Rights reserved.

S.Yang et al.
1 3
93 Page 6 of 20
randomness, random forest can decrease the variance of the forest estimator. A fivefold
cross-validated grid-search is used to search for the optimal values of several important
parameters, namely the number of trees (100, 150, 200, 500, 1000), the criterion to meas-
ure the quality of a split (mean squared error, mean absolute error) and the minimum num-
ber of samples required to split an internal node (2, 3, 4, 5, 10).
K-Nearest Neighbours (KNN) is a non-parametric method that calculates the predicted
value by taking weighted average values of k nearest neighbours. K is an integer value that
needs to be specified, as well as weighting scheme and distance metric. In this study, a
fivefold cross-validated grid-search is used to search for the optimal values of k (1, 5, 10,
15, 20), weighting scheme (uniform, distance) and distance metric (Euclidean, Manhattan).
Multi-layer Perceptron (MLP) is a typical class of feedforward neural network and it has
the capability to learn non-linear models. It consists of at least three layers, including input,
hidden and output layers. A fivefold cross-validated grid-search is used to search for the
optimal values of several important parameters, namely number of hidden layer (1, 2, 3),
number of nodes in the hidden layers (5, 10, 20, 50), activation function in the hidden layer
(logistic sigmoid, hyperbolic tangent, ReLU), coefficient for the L2 regularization term (1,
0.1, 0.01, 0.001, 0.0001) and maximum number of iterations (100, 200, 500, 1000).
2.4 Proposed deep learning model
This study proposes a novel deep learning model to utilize the information contained in the
PPG and ECG along with physical characteristics to predict BP. In contrast to the methods
mentioned earlier, which require pre-processing and feature extraction from the PPG and
ECG, deep learning models can take directly the raw signal data as input and the feature
learning is essentially embedded in the modelling process. This novel hybrid deep learning
model consists of various types of neural network models, such as Convolutional neural
network (CNN), Long short-term memory (LSTM) and fully connected layer (Dense). The
Dense layer is essentially a hidden layer in the MLP.
CNN was initially developed for image classification problems, where it receives two-
dimensional image pixels as input and generates output after a series of operations that
involve pattern learning. Multiple CNN layers are often applied in problems like this so
that simple patterns can first be identified in the lower layers and be used to form more
complex patterns within higher layers (Krizhevsky etal. 2012). The same process can be
applied to one-dimensional time series data, such as the PPG and ECG in this study. One-
dimensional CNN (1D CNN) can automatically learn to extract useful features from these
signals and how to construct appropriate models to predict BP.
1D CNN applies the convolution operation on the input data with a number of fil-
ters (also called feature detector) (LeCun and Bengio 1995). The length of these filters
can be specified and it is often referred to as kernel size. These filters are then moved
along the signals and the shift size is referred to as strides, which is often chosen to be
1. Different types of padding can be applied to determine the size of the output. Zero-
padding is often found to perform well in practice (Krizhevsky etal. 2012), and it is
also adopted in this study. An activation function is often applied to the results gener-
ated from the convolution operation. ReLU is very popular and found to perform well in
practice (Jarrett etal. 2009). Convolutional layers are often followed by dropout layers
for regularization, and then pooling layers, such as max pooling and average pooling
(Krizhevsky etal. 2012; Srivastava etal. 2014). CNN models tend to learn very quickly
and the dropout layer can help slow down the learning process and result in a potentially
Content courtesy of Springer Nature, terms of use apply. Rights reserved.

Non‑invasive cuff‑less blood pressure estimation using ahybrid…
1 3
Page 7 of 20 93
better final model. The pooling layers can help reduce the dimension and consolidate
learned features to the most essential elements. Pool length of 2 is often used in practice
and it is also adopted in this study. Several convolutional layers can be stacked together
to extract more complicated features. Hyperparameters that need to be determined for
1D CNN layers include kernel size (3, 5, 7, 9), number of filters (64, 128, 256, 512) and
number of epochs (20, 50, 100). In this study, the range of kernel sizes, number of fil-
ters and epochs is investigated using a cross-validation process in which an optimum is
selected based on accuracy and convergence time.
LSTM network model is a special type of recurrent neural network (RNN) that is able
to learn long-term dependencies (Hochreiter and Schmidhuber 1997). It has been proven
to be effective for sequence prediction tasks such as speech recognition, natural language
processing and machine translation (Chen etal. 2017; Cui etal. 2016; Tian etal. 2017).
A typical memory block in LSTM contains a memory cell and three gates, namely,
input, output and forget gates. The activation functions associated with the gates are
often logistic sigmoid function. LSTM can support multiple parallel sequences of input
data, such as the PPG and ECG signals in this study. LSTM can be used to automati-
cally learn temporal dependencies in raw PPG and ECG signals and use them to predict
BP values (Su etal. 2018). The parameter needs to be chosen for LSTM is the length of
state vector (10, 50, 100).
CNN and LSTM are two types of deep learning structures that can be used separately
to automatically learn from raw PPG and ECG signals to predict BP. They can also
be stacked together in a way that the output from CNN is fed to the following LSTM
layer. This stacked structure can be used to extract useful features and then learn the
long-term temporal dependencies from the raw signals. This type of structure has been
used for tasks such as detection of diabetes (Goutham etal. 2018), human activity rec-
ognition (Ordóñez and Roggen 2016), continuous cardiac monitoring (Saadatnejad etal.
2020), atrial fibrillation detection (Gotlibovych etal. 2018) and classification of myo-
cardial infarction (Baloglu etal. 2019), and it is often found to perform well in practice.
In addition to the raw signals, this study investigates a novel deep learning structure
that can also utilize useful information contained in physical characteristics to predict
BP. This novel model consists of various types of models, including CNN, LSTM and
Dense. This new structure can directly take raw signals and physical characteristics as
input at the same time. It can learn to automatically pick up useful information con-
tained in different types of input data and find an optimal way to link to BP.
3 Experimental design
Two streams of experiments are conducted in this study. The first stream involves the
use of physical characteristics and features extracted from PPG and ECG signals, which
are then used as input in traditional machine learning methods, namely LASSO, SVR,
AdaBoost, RR, KNN and MLP in this study. The second stream is the construction of
novel hybrid models that consists of various deep learning methods such as CNN and
LSTM and utilises physical characteristics and the raw PPG and ECG directly as inputs.
Several different architectures of hybrid models are investigated which are comprised of
different numbers of layers of CNN and LSTM. Experimental data is gathered using the
set-up and protocol described in the next section.
Content courtesy of Springer Nature, terms of use apply. Rights reserved.

S.Yang et al.
1 3
93 Page 8 of 20
3.1 Data acquisition system
The proposed cuff-less BP estimation system is illustrated as Fig.2 where BP, ECG
and PPG are measured simultaneously. Acquisition of data with this system has been
reviewed and approved according to the ethical review process in place.
The measurement system shown in Fig.2 comprises three sections for measurement
of BP, ECG and PPG. BP values are measured as a reference standard by a commercial
device (Lloyds Pharmacy Fully Automatic Blood Pressure Monitor LBPK1) with meas-
urement accuracy of ± 3mmHg (Lloyds 2021). The PPG signal is measured by infrared
transmission through the finger via a finger clip sensor (HRM-2511E, Kyoto Electronic
Co., China) with data transferred to a data acquisition board (Easy Pulse Sensor Ver-
sion 1.1, Elecrow, China) (Raj 2013). The ECG is measured with 3 disposable solid gel
electrodes based on the lead I configuration placed on 2 wrists and an ankle connected
to a data acquisition board (Analog devices, AD8232) (Lu etal. 2014). Due to availabil-
ity and convenience, the power for both circuit boards is supplied by an Arduino UNO
board (Arduino Co., Italy).
The measured PPG and ECG are then transmitted to a data acquisition device (USB-
6211, National Instruments). The sample frequency for the data acquisition is 1k sam-
ples/second in order to achieve a high-quality signal. The collected signals are sent to
the processing unit which is a battery powered laptop for the benefit of minimum noise
and to isolate the subject from mains power lines. All data were monitored and recorded
through LabView (National Instruments).
Fig. 2 System for measurement of BP, ECG and PPG for cuff-less BP estimation
Table 3 Physical characteristics
of participants in the experiment Mean Max Min
Age (years) 23.24 61 21
Height (cm) 168.78 190.5 150
Weight (kg) 66.84 110 45
Gender 23 Males 22 Females
Content courtesy of Springer Nature, terms of use apply. Rights reserved.

Non‑invasive cuff‑less blood pressure estimation using ahybrid…
1 3
Page 9 of 20 93
3.2 Data collection protocol
Data collection is performed on 45 participants. A detailed description is listed in
Table3.
All participants are healthy adults with no apparent arterial disease or physiological
abnormality. Informed consent is obtained from all participants and they are requested to
not take drinks that contained caffeine or a heavy meal 4h before the experiment to prevent
a large variability in BP. For each participant, the data collection includes two measure-
ment sections occurring on the same day using the same data acquisition protocol. Each
experiment takes less than 40min to collect all relevant signals with the following protocol:
1. The participant stays still for 10min, during which the consent forms are signed; the
individual physical characteristics including age, height, weight and gender are recorded;
the cuff of commercial BP measurement device is worn on the upper right arm; elec-
trodes are pasted on the limbs for ECG signal; and the clip is fixed on the index finger
of the left hand for PPG signal acquisition. Participants are requested to keep still during
the measurement because the PPG signal is sensitive to movement.
2. The PPG and ECG are recorded continuously for a period of 3min. At the same time,
BP is also measured. This procedure is repeated 3 times.
3. To induce a change in BP, the participant is asked to go downstairs from the 4th floor
to the 1st floor and then return as rapidly and safely as possible.
4. Once the participant returns, the same procedure in step (2) is repeated, but for 4 times.
Hence, there are 7 sets of data collected from each participant within around 40 min.
Accordingly, there is a total number of 315 records of data obtained. For each record
of the data, it includes PPG, ECG, BP and the corresponding participant’s physical
characteristics.
While raw PPG and ECG signals are fed directly into the hybrid deep learning model,
pre-processed and extracted features from PPG and ECG signals are used as inputs for tra-
ditional machine learning methods. As detailed in Sect.2.2, 12 features are extracted from
PPG and 45 features are extracted from ECG. In addition, PAT is also extracted, which
involves the use of both PPG and ECG. As a result, there are 58 features extracted from
PPG and ECG in total. Combined with four physical characteristics, the input dimension
for each observation in traditional machine learning models is 62. As models for DBP and
SBP are separately built, the output for each observation in traditional machine learning
models is 1, which is the corresponding DBP or SBP value.
3.3 Cross validation experiments
To generate a model with good generalization ability, this study conducts fivefold cross-
validation (CV) experiments where the training and testing samples are from different sets
of subjects. Since there are 45 participants in this study, data samples from 9 random par-
ticipants are used as testing samples and the rest are used as training samples in each CV
experiment. CV experiments are repeated 20 times and the evaluation results are averaged
over these 20 experiments. Separate models are built for systolic blood pressure (SBP)
and diastolic blood pressure (DBP). Such an experimental design provides robust results
as it involves multiple experiments to tackle the potential instability in a particular CV
Content courtesy of Springer Nature, terms of use apply. Rights reserved.

S.Yang et al.
1 3
93 Page 10 of 20
experiment. In addition, CV can also help avoid dependence of results on the choice of the
split in each experiment.
During the training process of each CV experiment, the hyperparameters of traditional
machine learning methods are determined using a fivefold CV on the training data set.
For hybrid deep learning models, as they take a lot more time to train, their hyperparam-
eters are determined in the first CV experiment where there are 5 different train-test splits.
The best hyperparameters are decided to be the ones that are chosen most times in these 5
splits.
3.4 Hybrid model architectures
A general representation of the architecture of the hybrid model is shown in Fig.3.
In contrast to traditional machine learning methods, this newly proposed hybrid
model can take raw PPG and ECG signals and physical characteristics as simultaneous
inputs by combining different deep learning structures. This hybrid model does not need
any feature extraction from the raw signals and can learn to extract the optimal fea-
tures itself. The input consists of two main parts, namely raw PPG and ECG signals and
physical characteristics. The dimension of the signal part for each sample observation
1D CNN
Dropout
Maxpooling
CNN
Block 1
1D CNN
Dropout
Maxpooling
CNN
Block n
Raw PPG and ECG signals
…...
Physical characteriscs
Dense
LSTM
Dropout
Concaten ate
Dense
BP
Fig. 3 The architecture of the hybrid model
Content courtesy of Springer Nature, terms of use apply. Rights reserved.

Non‑invasive cuff‑less blood pressure estimation using ahybrid…
1 3
Page 11 of 20 93
is (5000, 2), which means it is 5s of data (sampling rate 1000/s) and 2 channels (PPG
and ECG), while the dimension of physical characteristics is 4, including age, height,
weight and gender. Again, as the models for DBP and SBP are separately built, the out-
put dimension is 1.
CNN blocks and LSTM are used to extract features from raw signals while dense is
used to extract features from physical characteristics. The features learnt are then con-
catenated and fed to another Dense layer. Finally, this dense layer is followed by output
layer with no (linear) activation function as the target variable BP is continuous. Mean
absolute error (MAE) is used as the loss function.
CNN layers can be stacked together to extract more complicated features, different
numbers of CNN blocks are used to form several different architectures. The number
of CNN blocks is set to vary from 1 to 5, which leads to 5 different architectures. We
denote hybrid models with 1–5 CNN blocks as Hybrid Model 1–5 respectively.
Each CNN block is comprised of 1D CNN, dropout and maxpooling layers. Dropout
layer is also used following LSTM because it can impose regularization and prevent
overfitting. The dropout rate defines the probability of a randomly selected neuron being
dropped out. The dropout is only implemented during the training and not used in the
testing. The dropout rate is chosen from 0.1, 0.2 and 0.5 during the training, when other
hyperparameters are being chosen for the hybrid model, including the number of hidden
nodes for the Dense layers (10, 50, 100).
4 Results
4.1 Measured BP
Histograms of the BP data obtained from sphygmomanometer are presented in Fig.4.
The measured DBP and SBP ranged from 56–106mmHg to 84–170mmHg respectively.
The relatively large range of BP values is driven by interval measurement after physical
exercise in order to test the robustness of the prediction of BP values.
Fig. 4 Histogram of the BP values measured of: a SBP; b DBP
Content courtesy of Springer Nature, terms of use apply. Rights reserved.

S.Yang et al.
1 3
93 Page 12 of 20
4.2 Training andprediction results
After the first CV experiment, the hyperparameters are all chosen for hybrid models. Then
after 20 repetitions of CV experiments, the model with the best prediction performance is
Hybrid Model 3 with the following configuration details:
• CNN parameters: kernel size 7, number of filters 128 and number of epochs 20.
• Dropout rate: 0.5
• Pool length of Maxpooling layer: 2
• Length of state vector of LSTM: 50
• Number of hidden nodes for the Dense layers: 50
The BP prediction results of traditional machine learning methods and newly proposed
hybrid models are shown in Table4. Criteria for performance evaluation are MAE and
standard deviation (STD) of estimation.
The MAE and STD are calculated over 20 repetitions of CV experiments. According
to Table 4, some comparisons can be made. For instance, what stands out in the table
is that Hybrid Model 3 performs best in terms of both DBP and SBP with the results of
3.23 ± 4.75 mmHg and 4.43 ± 6.09 mmHg respectively. It is closely followed by Hybrid
Model 4 and Hybrid Model 5. It suggests that 3 CNN blocks are sufficient to extract useful
features from the raw signals.
It is clear that in all models, hybrid models achieved lower SBP and DBP errors than
traditional machine learning methods. It indicates that this newly proposed hybrid model
architecture can extract more information from the raw signals than manually extracted fea-
tures, which leads to a more accurate prediction of BP when combined with physical char-
acteristics. This also alludes to the possible misrepresentation of information by manually
extracted features due to challenges encountered with distorted waveforms. Deep learning
models seem to be more robust in this regard.
In addition to the comparisons within traditional machine learning methods, SVR is
found to perform best in terms of DBP with the result of 5.05 ± 7.26mmHg and followed
Table 4 Mean absolute error
(MAE) and standard deviation
(STD) achieved by traditional
machine learning methods and
novel hybrid models. Hybrid
Model n denotes n CNN blocks
in the model
Diastolic blood pressure
(mmHg)
Systolic blood pres-
sure (mmHg)
MAE STD MAE STD
LASSO 8.27 10.47 10.28 13.49
SVR 5.05 7.26 7.66 9.87
AdaBoost 7.82 8.08 8.92 10.95
RF 6.99 8.38 8.01 9.82
KNN 6.18 7.93 8.74 10.37
MLP 5.82 7.29 6.92 9.11
Hybrid model 1 4.98 5.85 6.73 8.02
Hybrid model 2 4.12 5.76 5.35 7.72
Hybrid model 3 3.23 4.75 4.43 6.09
Hybrid model 4 3.94 4.97 4.89 6.73
Hybrid model 5 3.83 5.01 5.22 7.28
Content courtesy of Springer Nature, terms of use apply. Rights reserved.

Non‑invasive cuff‑less blood pressure estimation using ahybrid…
1 3
Page 13 of 20 93
by MLP and KNN. In terms of SBP, MLP, with 6.92 ± 9.11mmHg, performs best among
traditional machine learning methods, followed by SVR and RF. LASSO is found to per-
form worst regarding both DBP and SBP, and it can be inferred that there exist strong non-
linear relationships between features used and BP which can lead to inferior performance
of LASSO. In addition, the inherent complexity in this problem necessitates using power-
ful regression algorithms, like hybrid models.
In order to further understand the origins of the improvements provided by the proposed
hybrid model, the performance of models without physical characteristics is investigated.
This investigates whether the superior performance of the hybrid models is mainly driven
by automatic feature extraction of raw signals or due to inclusion of the physical character-
istics of the subject.
The number of CNN blocks of these deep learning models are again set to vary from 1
to 5, and we denote these models by DP 1–5 respectively. The same procedure applied for
hybrid models is used for training.
The results of DP are shown in Table5. Regarding DBP prediction, the accuracy of DP
models improves quickly when the number of CNN blocks increases from 1 to 3. However,
as the number further increases from 3 to 5, the performance does not improve. A similar
pattern can be observed for the case of SBP prediction by DP model, except that the lowest
MAE and STD are obtained when the number of CNN blocks is equal to 4. When compar-
ing DP and hybrid model with the same number of CNN blocks, the results indicate that
the hybrid model always perform better than DP. This indicates that the inclusion of physi-
cal characteristics increases the prediction accuracy.
The results from the 3 best performing models, which are Hybrid Models with 3, 4 and
5 CNN blocks are further compared with British Hypertension Society (BHS) standard as
shown in Table6. This standard requires that the cumulative percentage of error is under
5mmHg, 10mmHg and 15mmHg (O’Brien etal. 2001). In this work, the predicted value
of DBP obtained from the Hybrid Model with 3 CNN blocks is consistent with Grade A
and the other two models meet Grade C. In addition, the hybrid model with 3 CNN blocks
is in congruence with Grade B and that with 4 CNN blocks meet Grade C in the estimation
of SBP values. However, the estimation of SBP from the Hybrid model with 5 CNN blocks
is not consistent with the BHS standard.
The Association for the Advancement of Medical Instrumentation (AAMI) standard
requires BP measurement devices to have MAE and STD values lower than 5mmHg and
8mmHg, respectively. According to Table6, all hybrid models achieve the requirements
when estimating DBP. However, only Hybrid Model 3 and 4 is consistent with the stand-
ard in SBP estimation. Also, the MAE and STD values of all traditional machine learning
models are outside the stipulated limits.
Table 5 Mean absolute error
(MAE) and standard deviation
(STD) of DP, Dense and new
hybrid models
Diastolic blood pressure
(mmHg)
Systolic blood pres-
sure (mmHg)
MAE STD MAE STD
DP 1 6.53 7.32 8.03 9.78
DP 2 5.29 6.13 6.98 8.15
DP 3 4.32 5.11 5.51 7.89
DP 4 4.33 5.18 5.37 7.60
DP 5 4.40 5.24 5.63 7.97
Content courtesy of Springer Nature, terms of use apply. Rights reserved.

S.Yang et al.
1 3
93 Page 14 of 20
Table 6 Comparison of the 3 best performing models with the BHS standard (O’Brien etal. 2001). Hybrid Model n denotes n CNN blocks in the model
Cumulative Error Percentage
DBP SBP
≤ 5mmHg (%) ≤ 10mmHg (%) ≤ 15mmHg (%) ≤ 5mmHg (%) ≤ 10mmHg (%) ≤ 15mmHg (%)
3 best performance models
Hybrid model 3 69.53 87.29 97.97 51.65 85.95 90.02
Hybrid model 4 59.83 84.24 96.53 43.77 74.78 91.24
Hybrid model 5 63.42 85.36 98.21 39.35 68.57 86.29
BHS standard
Grade A 60 85 95
Grade B 50 75 90
Grade C 40 65 85
Content courtesy of Springer Nature, terms of use apply. Rights reserved.

Non‑invasive cuff‑less blood pressure estimation using ahybrid…
1 3
Page 15 of 20 93
Table 7 A summary of the comparison of BP estimation results with other work
*MIMIC II stands for Multi-parameter Intelligent Monitoring in Intensive Care (MIMIC) II
Wor k Subjects Database Features DBP
(mmHg)
SBP (mmHg) Models Notes
MAE STD MAE STD
Dey etal. (2018) 205 Collected by authors PPG 5.0 6.1 6.9 9.0 Combined model Demographic and
physiological parti-
tioning
Kachuee etal. (2017) 5599 MIMIC II* PPG and PTT 5.32 6.14 11.17 10.09 AdaBoost Calibration free
He etal. (2016)> 2000 MIMIC II* PPG and ECG 4.44 3.72 8.29 5.84 Random forest Calibration free
Gao etal. (2016) 65 Collected by authors PPG 4.6 – 5.1 – Discrete wavelet transform Phone-obtained PPG
Kachuee etal. (2015) 850 MIMIC II* PPG and PTT 6.34 8.45 12.38 16.17 SVM Calibration free
This work 45 315 records and col-
lected by authors
Physical characteris-
tics, raw PPG and
ECG
3.23 4.75 4.43 6.09 Hybrid Model with 3 CNN blocks Calibration free
Content courtesy of Springer Nature, terms of use apply. Rights reserved.

S.Yang et al.
1 3
93 Page 16 of 20
Table7 compares the proposed approach in this paper and other works in the literature,
which use PPG, ECG and machine learning algorithms for BP estimation. In general, it is
difficult to compare related work in this field because of different and inadequately speci-
fied databases, different signal pre-processing procedures, and different evaluation methods
with different machine learning algorithms. In addition, some authors mixed all data and
split randomly for training and testing, but they did not explicitly report whether any data
of test subjects were included in the training data. Thus far, it is difficult to perform an
objective comparison between different research.
However, from a general perspective, most previous research applies traditional
machine learning algorithms, and the best results of DBP estimation ranges from about 4
to 6mmHg. The best SBP estimation results vary from 5 to 12mmHg. Although the BP
estimation results of traditional algorithms in this work have no obvious advantages com-
pared with the results from previous research, it is evident that hybrid models provide more
accurate prediction of BP.
5 Discussion andconclusions
In this paper, a novel hybrid deep learning model is proposed to predict BP using raw PPG,
ECG signals and some physical characteristics. Traditional machine learning methods used
in predicting BP involve extracting features from signals and it often presents challenges
when the quality of the signal is not good. This novel hybrid deep learning model consists
of several different types of deep learning layers which enable the automatic feature extrac-
tion and can learn to extract optimal features in the modelling process. The hybrid models
are tested on the data set collected and provide superior prediction results compared with
traditional machine learning models. Deep learning models have shown high performance
in many research areas and this study has shown its enormous potential in its application
in predicting BP. Because of its flexible structure, deep learning models can receive vari-
ous combinations of different types of inputs. This is a very useful feature as incorporating
more physiological data that can be relevant to BP is likely to increase the prediction accu-
racy. The best performance of hybrid model achieves 3.23 ± 4.75mmHg for DBP estima-
tion and 4.43 ± 6.09mmHg for SBP estimation. This result is consistent with Grade A and
Grade B in the estimation of DBP and SBP respectively. In line with this, this model also
achieves the requirements of the AAMI standard. It indicates that hybrid models with raw
PPG and ECG signals have high potential in cuff-less BP estimation.
Different number of CNN blocks are used in this study and three CNN blocks are found
to provide the best prediction results. Compared with its application in other areas such as
image processing, which often benefits from many more CNN layers, the useful features
contained in the physiological signals are not as complex. Therefore, the hybrid model
does not have to be very deep. Indeed, hybrid models with four and five CNN blocks are
outperformed by the hybrid model with three CNN blocks.
LSTM is included after CNN blocks as it is very useful in finding the important tem-
poral features in time series data and is suitable in processing signals. After LSTM the
features extracted from signals are combined with physical characteristics, which are age,
height, weight and gender in this study.
With the automatic learning of optimal features in the training stage, this hybrid model
minimizes the risk of omitting important features contained in the signals. Traditional
Content courtesy of Springer Nature, terms of use apply. Rights reserved.

Non‑invasive cuff‑less blood pressure estimation using ahybrid…
1 3
Page 17 of 20 93
feature extraction entails professional knowledge of specific signals and it is not often pos-
sible to extract all features that are potentially useful.
Despite the promising results found in this study, many important research questions
remain. The focus was on developing a general hybrid model whose hyperparameters are
determined using a pool of data from a number of individuals. A more accurate model can
be estimated by tuning these hyperparameters based on the data of each individual. After
this process, these models can be further calibrated to provide potentially more accurate
prediction for different people as their associated optimal structure and hyperparameters
may vary. The data used in this study is collected by the authors and there are 315 samples
in total. As deep learning models often require a big data set, more data is likely to further
improve the prediction accuracy. Therefore, in the next stage, we intend to further test the
novel hybrid model on bigger data sets, including those that are publicly available. In addi-
tion, this paper focuses on improving the prediction accuracy of BP, however, before deep
learning approaches are widely adopted it is important to consider the causality and rela-
tive importance of various features in predicting BP values (Holzinger etal. 2019). Due to
deep learning models’ multilayer and nonlinear structure, the relationship between input
and output is not transparent and predictions are often not traceable. This causes problems
in the interpretability of the deep learning models and make them of limited use in cases
where causalities are of great importance in the study. Although beyond the scope of this
work, this is a future direction that should be investigated.
Acknowledgements The authors acknowledge the financial support from the International Doctoral Inno-
vation Centre, Ningbo Education Bureau, Ningbo Science and Technology Bureau, and the University of
Nottingham. This work was also supported by the UK Engineering and Physical Sciences Research Council
[grant numbers EP/G037345/1 and EP/L016362/1].
Compliance with ethical standards
Conflict of interest The authors declare that they have no conflict of interest.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License,
which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long
as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Com-
mons licence, and indicate if changes were made. The images or other third party material in this article
are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the
material. If material is not included in the article’s Creative Commons licence and your intended use is not
permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly
from the copyright holder. To view a copy of this licence, visit http://creat iveco mmons .org/licen ses/by/4.0/.
References
Alty, S.R., Angarita-Jaimes, N., Millasseau, S.C., Chowienczyk, P.J.: Predicting arterial stiffness from the
digital volume pulse waveform. IEEE Trans. Biomed. Eng. 54, 2268–2275 (2007)
Baek, H.J., Kim, J.S., Kim, Y.S., Lee, H.B., Park, K.S.: Second derivative of photoplethysmography for esti-
mating vascular aging. In: 2007 6th International Special Topic Conference on Information Technol-
ogy Applications in Biomedicine, pp. 70–72 (2007)
Baloglu, U.B., Talo, M., Yildirim, O., Tan, R.S., Acharya, U.R.: Classification of myocardial infarction
with multi-lead ECG signals and deep CNN. Pattern Recognit. Lett. 122, 23–30 (2019). https ://doi.
org/10.1016/j.patre c.2019.02.016
Breiman, L.: Random forests. Mach. Learn. 45, 5–32 (2001). https ://doi.org/10.1023/A:10109 33404 324
Content courtesy of Springer Nature, terms of use apply. Rights reserved.

S.Yang et al.
1 3
93 Page 18 of 20
Chan, C., Hosanee, W., Kyriacou, Z., Allen, A., Lovell, F., Elgendi, K.: Multi-site photoplethysmography
technology for blood pressure assessment: challenges and recommendations. J. Clin. Med. 8, 1827
(2019). https ://doi.org/10.3390/jcm81 11827
Chen, Q., Zhu, X., Ling, Z., Wei, S., Jiang, H., Inkpen, D.: Enhanced LSTM for natural language inference.
In: Proceedings of the 55th Annual Meeting Assocication Computing Linguistics, Vol. 1, Long Papers,
pp. 1657–1668 (2017). https ://doi.org/10.18653 /v1/P17-1152
Chen, S., Ji, Z., Wu, H., Xu, Y.: A non-invasive continuous blood pressure estimation approach based on
machine learning. Sensors 19, 2585 (2019). https ://doi.org/10.3390/s1911 2585
Chua, C.P., Heneghan, C.: Continuous blood pressure monitoring using ECG and finger photoplethysmo-
gram. In: 2006 International Conference of the IEEE Engineering in Medicine and Biology Society,
pp. 5117–5120. IEEE, New York (2006)
Chua, E.C.-P., Redmond, S.J., McDarby, G., Heneghan, C.: Towards using photo-plethysmogram ampli-
tude to measure blood pressure during sleep. Ann. Biomed. Eng. 38, 945–954 (2010). https ://doi.
org/10.1007/s1043 9-009-9882-z
Cui, Y., Wang, S., Li, J.: LSTM neural reordering feature for statistical machine translation. In: Proceed-
ings of the 2016 Conference on North America Chapter Association Computing Linguistics Human
Language Technology, pp. 977–982 (2016). https ://doi.org/10.18653 /v1/N16-1112
Dey, J., Gaurav, A., Tiwari, V.N.: InstaBP: cuff-less blood pressure monitoring on Smartphone using sin-
gle PPG sensor. In: 2018 40th Annual International Conference of the IEEE Engineering in Medi-
cine and Biology Society (EMBC), pp. 5002–5005. IEEE, Honolulu (2018)
Drucker, H.: Improving regressors using boosting techniques. In: ICML, pp. 107–115 (1997)
Drucker, H., Burges, C.J.C., Kaufman, L., Smola, A., Vapnik, V.: Support Vector Regression Machines
(1996)
Elgendi, M.: On the analysis of fingertip photoplethysmogram signals. Curr. Cardiol. Rev. 8, 14–25
(2012). https ://doi.org/10.2174/15734 03128 01215 782
Elgendi, M., Fletcher, R., Liang, Y., Howard, N., Lovell, N.H., Abbott, D., Lim, K., Ward, R.: The use
of photoplethysmography for assessing hypertension. Npj Digit. Med. 2, 1–11 (2019). https ://doi.
org/10.1038/s4174 6-019-0136-7
Friedman, J., Hastie, T., Tibshirani, R.: Regularization paths for generalized linear models via coordi-
nate descent. J. Stat. Softw. 33, 1–22 (2010)
Gao, S.C., Wittek, P., Zhao, L., Jiang, W.J.: Data-driven estimation of blood pressure using photop-
lethysmographic signals. In: 2016 38th Annual International Conference of the IEEE Engineering
in Medicine and Biology Society (EMBC), pp. 766–769. IEEE, Orlando (2016)
Gotlibovych, I., Crawford, S., Goyal, D., Liu, J., Kerem, Y., Benaron, D., Yilmaz, D., Marcus, G., Li,
Y.: End-to-end deep learning from raw sensor data: Atrial fibrillation detection using wearables.
ArXiv180710707 Cs Statistics (2018)
Goutham, S., Kpas, S., Vinayakumar, R.: Automated detection of diabetes using CNN and CNN-
LSTM network and heart rate signals. Procedia Comput. Sci. 132, 1253–1262 (2018). https ://doi.
org/10.1016/j.procs .2018.05.041
He, R., Huang, Z.-P., Ji, L.-Y., Wu, J.-K., Li, H., Zhang, Z.-Q.: Beat-to-beat ambulatory blood pressure
estimation based on random forest. In: 2016 IEEE 13th International Conference on Wearable and
Implantable Body Sensor Networks (BSN), pp. 194–198. IEEE, San Francisco (2016)
Hennig, A., Patzak, A.: Continuous blood pressure measurement using pulse transit time. Somnologie
Schlafforschung Schlafmed. 17, 104–110 (2013). https ://doi.org/10.1007/s1181 8-013-0617-x
Hochreiter, S., Schmidhuber, J.: Long short-term memory. Neural Comput. 9, 1735–1780 (1997)
Holzinger, A., Langs, G., Denk, H., Zatloukal, K., Müller, H.: Causability and explainability of artificial
intelligence in medicine. WIREs Data Min. Knowl. Discov. 9, 666 (2019). https ://doi.org/10.1002/
widm.1312
Jarrett, K., Kavukcuoglu, K., Ranzato, M., LeCun, Y.: What is the best multi-stage architecture for
object recognition? In: 2009 IEEE 12th International Conference on Computer Vision, pp. 2146–
2153 (2009)
Kachuee, M., Kiani, M.M., Mohammadzade, H., Shabany, M. (2015): Cuff-less high-accuracy calibra-
tion-free blood pressure estimation using pulse transit time. In: 2015 IEEE International Sympo-
sium on Circuits and Systems (ISCAS), pp. 1006–1009 (2015)
Kachuee, M., Kiani, M.M., Mohammadzade, H., Shabany, M.: Cuffless blood pressure estimation algo-
rithms for continuous health-care monitoring. IEEE Trans. Biomed. Eng. 64, 859–869 (2017). https
://doi.org/10.1109/TBME.2016.25809 04
Krizhevsky, A., Sutskever, I., Hinton, G.E.: Imagenet classification with deep convolutional neural net-
works. In: Advances in Neural Information Processing Systems, pp. 1097–1105 (2012)
Content courtesy of Springer Nature, terms of use apply. Rights reserved.

Non‑invasive cuff‑less blood pressure estimation using ahybrid…
1 3
Page 19 of 20 93
LeCun, Y., Bengio, Y.: Convolutional networks for images, speech, and time series. In M. A. Arbib
(Ed.), The handbook of brain theory and neural networks, MIT Press (1995)
Li, J.-Q., Li, R., Chen, Z.-Z., Deng, G.-Q., Wang, H., Mavromoustakis, C.X., Song, H., Ming, Z.: Design
of a continuous blood pressure measurement system based on pulse wave and ECG signals. IEEE J.
Transl. Eng. Health Med. 6, 1–14 (2018). https ://doi.org/10.1109/JTEHM .2017.27888 85
Liang, Y., Chen, Z., Ward, R., Elgendi, M.: Hypertension assessment via ECG and PPG Signals: An
evaluation using MIMIC database. Diagnostics 8, 65 (2018a). https ://doi.org/10.3390/diagn ostic
s8030 065
Liang, Y., Elgendi, M., Chen, Z., Ward, R.: An optimal filter for short photoplethysmogram signals. Sci.
Data 5, 180076 (2018b). https ://doi.org/10.1038/sdata .2018.76
Lloyds, pharmacy: LLOYDSPHARMACY LBPK1 USER MANUAL (2021). https ://www.manua lslib
.com/manua l/63501 8/Lloyd sphar macy-Lbpk1 .html. Accessed 12 Jan 2021
Lu, T.C., Liu, P., Gao, X., Lu, Q.Y.: A portable ECG monitor with low power consumption and small size
based on AD8232 chip. Appl. Mech. Mater. 513–517, 2884–2887 (2014). https ://doi.org/10.4028/
www.scien tific .net/AMM.513-517.2884
Millasseau, S.C., Kelly, R.P., Ritter, J.M., Chowienczyk, P.J.: Determination of age-related increases in large
artery stiffness by digital pulse contour analysis. Clin. Sci. Lond. Engl. 1979(103), 371–377 (2002).
https ://doi.org/10.1042/cs103 0371
Mousavi, S.S., Firouzmand, M., Charmi, M., Hemmati, M., Moghadam, M., Ghorbani, Y.: Blood pressure
estimation from appropriate and inappropriate PPG signals using a whole-based method. Biomed. Sig-
nal Process. Control 47, 196–206 (2019). https ://doi.org/10.1016/j.bspc.2018.08.022
O’Brien, E., Parati, G., Myers, M.: Blood pressure measuring devices: recommendations of the European
Society of Hypertension. BMJ 322, 531–536 (2001). https ://doi.org/10.1136/bmj.322.7285.531
Ogedegbe, G., Pickering, T.: Principles and techniques of blood pressure measurement. Cardiol. Clin. 28,
571–586 (2010). https ://doi.org/10.1016/j.ccl.2010.07.006
Ordóñez, F.J., Roggen, D.: Deep convolutional and LSTM recurrent neural networks for multimodal wear-
able activity recognition. Sensors 16, 115 (2016). https ://doi.org/10.3390/s1601 0115
Raj, B.: Easy Pulse Sensor (Version 1.1) Overview (Part 1) | Embedded Lab. (2013). http://embed ded-lab.
com/blog/easy-pulse -versi on-1-1-senso r-overv iew-part-1/. Accessed 12 Jan 2021
Ribas Ripoll, V., Vellido, A.: Blood pressure assessment with differential pulse transit time and deep learn-
ing: a proof of concept. Kidney Dis. 5, 23–27 (2019). https ://doi.org/10.1159/00049 3478
Rundo, F., Ortis, A., Battiato, S., Conoci, S.: Advanced bio-inspired system for noninvasive cuff-less
blood pressure estimation from physiological signal analysis. Computation 6, 46 (2018). https ://doi.
org/10.3390/compu tatio n6030 046
Saadatnejad, S., Oveisi, M., Hashemi, M.: LSTM-based ECG classification for continuous monitor-
ing on personal wearable devices. IEEE J. Biomed. Health Inform. 24, 515–523 (2020). https ://doi.
org/10.1109/JBHI.2019.29113 67
Sharifi, I., Goudarzi, S., Khodabakhshi, M.B.: A novel dynamical approach in continuous cuffless blood
pressure estimation based on ECG and PPG signals. Artif. Intell. Med. 97, 143–151 (2019). https ://doi.
org/10.1016/j.artme d.2018.12.005
Shin, W., Cha, Y.D., Yoon, G.: ECG/PPG integer signal processing for a ubiquitous health monitoring sys-
tem. J. Med. Syst. 34, 891–898 (2010). https ://doi.org/10.1007/s1091 6-009-9304-7
Simjanoska, M., Gjoreski, M., Gams, M., Madevska Bogdanova, A.: Non-invasive blood pressure estima-
tion from ECG using machine learning techniques. Sensors 18, 1160 (2018). https ://doi.org/10.3390/
s1804 1160
Slapničar, G., Mlakar, N., Luštrek, M.: Blood pressure estimation from photoplethysmogram using a spec-
tro-temporal deep neural network. Sensors 19, 3420 (2019). https ://doi.org/10.3390/s1915 3420
Srivastava, N., Hinton, G., Krizhevsky, A., Sutskever, I., Salakhutdinov, R.: Dropout: a simple way to pre-
vent neural networks from overfitting. J. Mach. Learn. Res. 15, 1929–1958 (2014)
Stergiou, G., Palatini, P., Asmar, R., de la Sierra, A., Myers, M., Shennan, A., Wang, J., O’Brien, E., Parati,
G.: Blood pressure measurement and hypertension diagnosis in the 2017 US guidelines: first things
first. Hypertension 71, 963–965 (2018). https ://doi.org/10.1161/HYPER TENSI ONAHA .118.10853
Su, P., Ding, X., Zhang, Y., Liu, J., Miao, F., Zhao, N.: Long-term blood pressure prediction with deep
recurrent neural networks. In: 2018 IEEE EMBS International Conference on Biomedical Health
Informatics (BHI), pp. 323–328 (2018)
Tanveer, Md.S., Hasan, Md.K.: Cuffless blood pressure estimation from electrocardiogram and photop-
lethysmogram using waveform based ANN-LSTM network. Biomed. Signal Process. Control 51, 382–
392 (2019). https ://doi.org/10.1016/j.bspc.2019.02.028
Tian, X., Zhang, J., Ma, Z., He, Y., Wei, J., Wu, P., Situ, W., Li, S., Zhang, Y.: Deep LSTM for Large
Vocabulary Continuous Speech Recognition. ArXiv:170307090 Cs (2017)
Content courtesy of Springer Nature, terms of use apply. Rights reserved.

S.Yang et al.
1 3
93 Page 20 of 20
Wang, Q., Xu, Y., Zeng, G., Sun, M.: Continuous blood pressure estimation based on two-domain fusion
model. Comput. Math. Methods Med. 2018, 1–10 (2018). https ://doi.org/10.1155/2018/19816 27
World Health Organisation: Hypertension. (2019). https ://www.who.int/news-room/fact-sheet s/detai l/hyper
tensi on. Accessed 12 Jan 2021
Yang, S., Zaki, W.S.W., Morgan, S.P., Cho, S.-Y., Cor reia, R., Zhang, Y.: Blood pressure estimation with
complexity features from electrocardiogram and photoplethysmogram signals. Opt. Quantum Electron.
52, 135 (2020). https ://doi.org/10.1007/s1108 2-020-2260-7
Yoon, Y., Cho, J.H., Yoon, G.: Non-constrained blood pressure monitoring using ECG and PPG for personal
healthcare. J. Med. Syst. 33, 261–266 (2009). https ://doi.org/10.1007/s1091 6-008-9186-0
Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and
institutional affiliations.
Content courtesy of Springer Nature, terms of use apply. Rights reserved.

1.
2.
3.
4.
5.
6.
Terms and Conditions
Springer Nature journal content, brought to you courtesy of Springer Nature Customer Service Center
GmbH (“Springer Nature”).
Springer Nature supports a reasonable amount of sharing of research papers by authors, subscribers
and authorised users (“Users”), for small-scale personal, non-commercial use provided that all
copyright, trade and service marks and other proprietary notices are maintained. By accessing,
sharing, receiving or otherwise using the Springer Nature journal content you agree to these terms of
use (“Terms”). For these purposes, Springer Nature considers academic use (by researchers and
students) to be non-commercial.
These Terms are supplementary and will apply in addition to any applicable website terms and
conditions, a relevant site licence or a personal subscription. These Terms will prevail over any
conflict or ambiguity with regards to the relevant terms, a site licence or a personal subscription (to
the extent of the conflict or ambiguity only). For Creative Commons-licensed articles, the terms of
the Creative Commons license used will apply.
We collect and use personal data to provide access to the Springer Nature journal content. We may
also use these personal data internally within ResearchGate and Springer Nature and as agreed share
it, in an anonymised way, for purposes of tracking, analysis and reporting. We will not otherwise
disclose your personal data outside the ResearchGate or the Springer Nature group of companies
unless we have your permission as detailed in the Privacy Policy.
While Users may use the Springer Nature journal content for small scale, personal non-commercial
use, it is important to note that Users may not:
use such content for the purpose of providing other users with access on a regular or large scale
basis or as a means to circumvent access control;
use such content where to do so would be considered a criminal or statutory offence in any
jurisdiction, or gives rise to civil liability, or is otherwise unlawful;
falsely or misleadingly imply or suggest endorsement, approval , sponsorship, or association
unless explicitly agreed to by Springer Nature in writing;
use bots or other automated methods to access the content or redirect messages
override any security feature or exclusionary protocol; or
share the content in order to create substitute for Springer Nature products or services or a
systematic database of Springer Nature journal content.
In line with the restriction against commercial use, Springer Nature does not permit the creation of a
product or service that creates revenue, royalties, rent or income from our content or its inclusion as
part of a paid for service or for other commercial gain. Springer Nature journal content cannot be
used for inter-library loans and librarians may not upload Springer Nature journal content on a large
scale into their, or any other, institutional repository.
These terms of use are reviewed regularly and may be amended at any time. Springer Nature is not
obligated to publish any information or content on this website and may remove it or features or
functionality at our sole discretion, at any time with or without notice. Springer Nature may revoke
this licence to you at any time and remove access to any copies of the Springer Nature journal content
which have been saved.
To the fullest extent permitted by law, Springer Nature makes no warranties, representations or
guarantees to Users, either express or implied with respect to the Springer nature journal content and
all parties disclaim and waive any implied warranties or warranties imposed by law, including
merchantability or fitness for any particular purpose.
Please note that these rights do not automatically extend to content, data or other material published
by Springer Nature that may be licensed from third parties.
If you would like to use or distribute our Springer Nature journal content to a wider audience or on a
regular basis or in any other manner not expressly permitted by these Terms, please contact Springer
Nature at
onlineservice@springernature.com