Instrumentation & Measurement Magazine 25-5 - 44

Thus, the expected power of noise was:
10
P
segments that did not belong to class i and were correctly predicted
as other classes.
P

noise
10
signal
SNR
(3)
Data
The noise Sn extracted from the noise database was a discrete
matrix Snoise = [s1,s2,......,sn]. Thus, the power of Snoise
was:
PSnoise
 S *Snoise

The expected noise matrix Sn
length Snoise


'
noise
(4)
was calculated as:
P
n 
signal
SNR
SS
P
10
10

Snoise
noise
(5)
Thus, the ECG segments matrix Y with noise was:
Y
SSn signal
Evaluation Method
The model was trained on the wearable ECG segments with a
10-fold cross validation method. The collected ECG segments
were randomly divided into 10 parts, nine of which were used
as training data and the remaining part as validation data in
each fold. There was no overlap in the validation data of different
folds. To measure the model classification ability of each
class, the sensitivity (Sei
), precision (P+i), specificity (Spi
) of each
class and the accuracy (Acc) and F1-score (F1) of the classification
result were calculated as:
Sei 
TP FN
ii
TPi

P i
 TP FP

TPi

Spi 
Acc 
TN FP
TNi

1
3
()
TPi
*(

F 1 
 100%
 100%
ii
 100%
ii
  
ii i
2*TPi
2*TP FN FP
i ii

3
100%
(11)
where i= good, medium or poor. TPi represented the number of
ECG segments in class i that were correctly predicted as class
i; FNi
represented the number of ECG segments in class i that
were falsely predicted into other classes; FPi
represented the
number of ECG segments not from class i and that were falsely
predicted as class i; and TNi
represented the number of ECG
44
TP FP FN TN

100%
i )
(7)
(8)
(9)
(10)
(6)
Training and Validation Data: Wearable ECG
Database
The ECG data was collected by wearable ECG monitoring
devices, which collect standard lead-I and lead-II ECG signal
with a sampling frequency of 400 Hz and a resolution of
16 bits over a range of ±10 mV. Accurate signal quality classification
of ECG segments with special arrhythmia is one of
the difficulties in the signal quality classification of dynamic
ECGs. Due to the influence of individual differences in ECG
signals and the complexity of abnormal arrhythmia waveforms,
ECG signal quality evaluation algorithms based on
small datasets may regard special arrhythmia waveforms as
noise. This may affect automated analysis algorithms or doctors'
diagnoses of disease in patients. Therefore, to ensure that
important disease information is not be discarded as noise in
the dynamic monitoring, the evaluation of signal quality assessment
for the ECG signals with special arrhythmia is one
of the most important evaluation indicators for the signal
quality assessment algorithm. In this work, the ECG data was
collected from 20 patients who had current or previous cardiovascular
disease, and these collected recordings contained
five different abnormal heart rhythms, such as premature
ventricular beats (PVC), premature atrial beats (PAC), atrial
fibrillation (AF), right bundle-branch heart block (RBBB) and
tachycardia. A fixed time window of 10 s was used for segmenting
the ECG episodes.
The collected ECG segments were classified into three categories
according to the complexity of the noise in the segments
and its influence on signal feature points. The signal of good
quality indicated that the ECG segment was subject to a small
amount of noise interference and contained clear QRS waveform.
Feature points such as T and P wave were clearly visible
and easily recognized.
The signal of medium quality indicated that the ECG segment
was under the interference of various noises but without
large-scale pulse signal. The feature points such as P and T
wave could not be easily located. However, the QRS wave
in the segment could still be accurately discriminated which
means the noise in the ECG segment did not affect the detection
of diagnosis of rhythm information.
The signal of poor quality indicated that the ECG segment
contained little information such as the signal of lead detachment
and white noise, or the segment was interfered by large
noises. Some signal may contain large pulse interference
which could affect the detection of the QRS. Some segments
may contain severe white noise and large baseline drift, resulting
in an extreme change in overall amplitude and the inability
to locate its QRS position.
Finally, 3,000 ECG segments of each class were selected as
training data. The ECG segments were annotated manually
by three experienced employees and the doubtful cases were
IEEE Instrumentation & Measurement Magazine
August 2022

Instrumentation & Measurement Magazine 25-5

Table of Contents for the Digital Edition of Instrumentation & Measurement Magazine 25-5

Instrumentation & Measurement Magazine 25-5 - Cover1
Instrumentation & Measurement Magazine 25-5 - Cover2
Instrumentation & Measurement Magazine 25-5 - 1
Instrumentation & Measurement Magazine 25-5 - 2
Instrumentation & Measurement Magazine 25-5 - 3
Instrumentation & Measurement Magazine 25-5 - 4
Instrumentation & Measurement Magazine 25-5 - 5
Instrumentation & Measurement Magazine 25-5 - 6
Instrumentation & Measurement Magazine 25-5 - 7
Instrumentation & Measurement Magazine 25-5 - 8
Instrumentation & Measurement Magazine 25-5 - 9
Instrumentation & Measurement Magazine 25-5 - 10
Instrumentation & Measurement Magazine 25-5 - 11
Instrumentation & Measurement Magazine 25-5 - 12
Instrumentation & Measurement Magazine 25-5 - 13
Instrumentation & Measurement Magazine 25-5 - 14
Instrumentation & Measurement Magazine 25-5 - 15
Instrumentation & Measurement Magazine 25-5 - 16
Instrumentation & Measurement Magazine 25-5 - 17
Instrumentation & Measurement Magazine 25-5 - 18
Instrumentation & Measurement Magazine 25-5 - 19
Instrumentation & Measurement Magazine 25-5 - 20
Instrumentation & Measurement Magazine 25-5 - 21
Instrumentation & Measurement Magazine 25-5 - 22
Instrumentation & Measurement Magazine 25-5 - 23
Instrumentation & Measurement Magazine 25-5 - 24
Instrumentation & Measurement Magazine 25-5 - 25
Instrumentation & Measurement Magazine 25-5 - 26
Instrumentation & Measurement Magazine 25-5 - 27
Instrumentation & Measurement Magazine 25-5 - 28
Instrumentation & Measurement Magazine 25-5 - 29
Instrumentation & Measurement Magazine 25-5 - 30
Instrumentation & Measurement Magazine 25-5 - 31
Instrumentation & Measurement Magazine 25-5 - 32
Instrumentation & Measurement Magazine 25-5 - 33
Instrumentation & Measurement Magazine 25-5 - 34
Instrumentation & Measurement Magazine 25-5 - 35
Instrumentation & Measurement Magazine 25-5 - 36
Instrumentation & Measurement Magazine 25-5 - 37
Instrumentation & Measurement Magazine 25-5 - 38
Instrumentation & Measurement Magazine 25-5 - 39
Instrumentation & Measurement Magazine 25-5 - 40
Instrumentation & Measurement Magazine 25-5 - 41
Instrumentation & Measurement Magazine 25-5 - 42
Instrumentation & Measurement Magazine 25-5 - 43
Instrumentation & Measurement Magazine 25-5 - 44
Instrumentation & Measurement Magazine 25-5 - 45
Instrumentation & Measurement Magazine 25-5 - 46
Instrumentation & Measurement Magazine 25-5 - 47
Instrumentation & Measurement Magazine 25-5 - 48
Instrumentation & Measurement Magazine 25-5 - 49
Instrumentation & Measurement Magazine 25-5 - 50
Instrumentation & Measurement Magazine 25-5 - 51
Instrumentation & Measurement Magazine 25-5 - 52
Instrumentation & Measurement Magazine 25-5 - 53
Instrumentation & Measurement Magazine 25-5 - 54
Instrumentation & Measurement Magazine 25-5 - 55
Instrumentation & Measurement Magazine 25-5 - 56
Instrumentation & Measurement Magazine 25-5 - 57
Instrumentation & Measurement Magazine 25-5 - 58
Instrumentation & Measurement Magazine 25-5 - 59
Instrumentation & Measurement Magazine 25-5 - 60
Instrumentation & Measurement Magazine 25-5 - 61
Instrumentation & Measurement Magazine 25-5 - 62
Instrumentation & Measurement Magazine 25-5 - 63
Instrumentation & Measurement Magazine 25-5 - 64
Instrumentation & Measurement Magazine 25-5 - 65
Instrumentation & Measurement Magazine 25-5 - 66
Instrumentation & Measurement Magazine 25-5 - 67
Instrumentation & Measurement Magazine 25-5 - Cover3
Instrumentation & Measurement Magazine 25-5 - Cover4
https://www.nxtbook.com/allen/iamm/26-6
https://www.nxtbook.com/allen/iamm/26-5
https://www.nxtbook.com/allen/iamm/26-4
https://www.nxtbook.com/allen/iamm/26-3
https://www.nxtbook.com/allen/iamm/26-2
https://www.nxtbook.com/allen/iamm/26-1
https://www.nxtbook.com/allen/iamm/25-9
https://www.nxtbook.com/allen/iamm/25-8
https://www.nxtbook.com/allen/iamm/25-7
https://www.nxtbook.com/allen/iamm/25-6
https://www.nxtbook.com/allen/iamm/25-5
https://www.nxtbook.com/allen/iamm/25-4
https://www.nxtbook.com/allen/iamm/25-3
https://www.nxtbook.com/allen/iamm/instrumentation-measurement-magazine-25-2
https://www.nxtbook.com/allen/iamm/25-1
https://www.nxtbook.com/allen/iamm/24-9
https://www.nxtbook.com/allen/iamm/24-7
https://www.nxtbook.com/allen/iamm/24-8
https://www.nxtbook.com/allen/iamm/24-6
https://www.nxtbook.com/allen/iamm/24-5
https://www.nxtbook.com/allen/iamm/24-4
https://www.nxtbook.com/allen/iamm/24-3
https://www.nxtbook.com/allen/iamm/24-2
https://www.nxtbook.com/allen/iamm/24-1
https://www.nxtbook.com/allen/iamm/23-9
https://www.nxtbook.com/allen/iamm/23-8
https://www.nxtbook.com/allen/iamm/23-6
https://www.nxtbook.com/allen/iamm/23-5
https://www.nxtbook.com/allen/iamm/23-2
https://www.nxtbook.com/allen/iamm/23-3
https://www.nxtbook.com/allen/iamm/23-4
https://www.nxtbookmedia.com