Professional Documents
Culture Documents
7, JULY 2015
Abstract— Wrist pulse has long been recognized as an impor- blood parameters, etc. [3]. According to this principle, wrist
tant physiological symbol for judging the health status of patients. pulse is usually regarded to give more body information
It is regarded as one of the four diagnosis techniques in tradi- than the electrocardiogram (ECG), which results in broader
tional Chinese medicine and plays an important role in disease
analysis. However, the pulse diagnosis skill is subjective and needs applications in health status analysis [10], [18]–[24].
decades of experience to master. To make the pulse diagnosis In TCM, a Chinese medicine practitioner took pulse by
more objective and easily understood by practitioners, many putting fingers on the patients’ wrist, certain position that
pulse-taking platforms were invented, but several disadvantages, is regarded as the position of pulse [1]. However, the pulse
including the simplicity and insufficiency, limit the applications. diagnosis skill requires several years experience to practice
This paper proposes a novel wrist pulse system for the pulse diag-
nosis to overcome the shortcomings in previous pulse platforms. and master. Sometimes it may be not accordant among
We integrated a pressure sensor with a photoelectric sensor to different practitioners because of the subjective judgment [26].
make a fusion sensor to feel pulse in different approaches and Compare with the finger-feeling diagnosis, a scientific way of
designed the multichannel sensor arrays structure to acquire pulse diagnosis is using the sensing elements to simulate the
more pulse information. Finally, pulse analysis algorithm and functions of fingers and transforming the physical signals into
classification methods are introduced. Experiments on disease
classification are carried out to test the system performance with digital signals [27]. The quantification of the pulse diagnosis
multichannel and different sensor arrays. The results show that solves the problem of obtaining the objective pulse signals [4].
the novel system is not only able to distinguish between healthy The pulse information which includes strengths, amplitudes,
pulse samples and subjects suffering from diabetes, but also good fluency, shapes, widths, variations of the rhythm, and so
at obtaining more information than the conventional pulse system on are obtained from the digital pulse signals for further
with single channel or simplex-type sensor.
processing [3].
Index Terms— Different sensor arrays, multichannel array Recently, there are two kinds of pulse devices that have been
design, pulse waveform analysis, wrist pulse system. reported. One is called pulse oximeter, applied with a probe
attached to the patient’s finger [8], [9]. It is a noninvasive
I. I NTRODUCTION
method for measuring the oxygen saturation of arterial blood.
Fig. 6. Physic model of the pressure sensor. 1: fixed mount. 2: strain gauge. TABLE II
3: cantilever beam. 4: contactor. 5: skin. 6: blood vessel.
PARAMETERS OF S EMICONDUCTOR G AUGE
TABLE III
P RESSURE S ENSOR S TANDARDIZATION
Fig. 10. Top view of photoelectric sensor array and the real object.
Fig. 15. Pulse waveforms taken by the proposed pulse system. (a) Part of
typical oringnal pulse waveform collected by pressure sensor, the horizontal
axis represents the time (s) and the ordinate axis represents the amptitude
of the pulse waveform (mV). (b) Denoised pulse signal. (c) Baseline wander
Fig. 14. Overview of sensor probe framework. removed pulse signal.
Fig. 17. Original photoelectric pulse waveforms from the sensor arrays of the three pulse-taking positions.
Fig. 18. Pressure pulse waveforms collected from the three channels.
in amplitudes, which reflect the distance from the center of nine signals in the array. It can be seen that the waveforms
pulse waveforms. On the basis of fitting, the width of the blood are stable and almost the same in shape and frequency. Fig. 18
vessel is obtained by setting a proper threshold. The radial shows the pressure pulse waveforms from the three channels
arterial varies in width as it associated with the conditions of with independent contact static pressures. The details and
vessel. Different health states usually lead to the change of amplitudes of each period have slight difference among the
vessel form and it is an important symbol for judging the whole time series. Generally, the pulse waveforms from the
fullness or tension produced by the fluid content of blood Chi channel have the least power energies. And the pulsations
vessels [15]. On the other hand, the sensor array provides a even disappear at this channel under certain physiological
criterion for the pulse location. It guarantees that the pulse condition such as the weakness. It is an important symbol
waveforms are acquired under the optimal collecting position. for the pulse diagnosis.
Additionally, the proposed system is able to measure
30 channels of wrist pulse simultaneously at Cun, Guan,
and Chi three pulse-taking positions. The photoelectric pulse B. Disease Diagnosis
waveforms of the three channels are shown in Fig. 17, and In the second stage, we test the system ability in disease
each column presents a channel and each channel includes diagnosis. The classification between healthy persons and
WANG et al.: NOVEL MULTICHANNEL WRIST PULSE SYSTEM 2029
Fig. 19. Statistical distribution of pulse features between healthy subjects (blue point) and diabetes (red point). (a) Pulse feature distribution. (b) Pulse width
distribution.
TABLE V TABLE VI
D ATA O RGANIZATION S TATISTICAL R ESULTS OF THE F EATURE D ISTRIBUTION
patients known to have suffered from diabetes is selected TSSA algorithm to select the representive of pulse array in
to evaluate the performance. Modern clinical studies demon- each channel. Then the features of pulse shape, pulse width,
strate that premature loss of arterial elasticity and endothe- frequency, and static contact pressure are extracted. We also
lial function exists in patients with diabetes. Such loss will extract the wavelet coefficients as energy features, because the
eventually decrease the flexibility of vasculature and changes energy distribution in different frequency bands are different
the shape, amplitude, and rhythm of patient wrist pulses in from that of a patient and can be easily obtained by WT.
correspondence with the hemodynamic characteristics of blood Finally, we combine all the features as a vector in serial
flow [48]. Computerized pulse diagnosis for diabetes has been strategy.
studied by many researches [52]–[54]. Therefore, pulse is Table VI gives statistical results of pulse feature distribution
considered as an indicator for diabetes. between healthy subjects and diabetes. Although there is dif-
In our experiments, we select the diabetes diagnosis ference in average value of energy and frequency feature, the
as a case study and constitute a pulse dataset containing large standard deviation value leads to the overlap. Fig. 19(a)
125 healthy subjects and 125 diabetes patients. All the patients shows the distribution of pulse shape features between these
were inpatient volunteers from the Hong Kong Prince of Wales two patterns. The middle line stands for average value of
Hospital. All the healthy volunteers were from the physical pulse shape, and the interval at each point is the standard
examination center of the hospital. The labels of the healthy deviation. It can be seen that the main difference appears
subjects were determined according to the recent health check, in the descent stage of the pulse waveforms. Especially, in
whereas the diabetic patients were confirmed and correlated by the period between pulse tidal wave and pulse dicrotic wave
comparing their levels with standard clinical blood markers. pulse signals have significant difference in average value and
The distribution of the pulse database is ensured to avoid the deviation interval. The width distribution shown in Fig. 19(b)
influence caused by biological interpatient variability. Table V indicates that the width of the healthy subjects is more like
lists the summary of the dataset. The age, body–mass index, the uniform distribution, while the pulse width of the diabetes
and gender distribution is similar between healthy and diabetes is concentrated in the lower bands.
patients. The statistical distribution of pulse features shows that the
Three pressure pulse signals and 27 photoelectric array classification between health and diabetes is complicated. The
signals are collected for each subject from Cun, Guan, and Chi linear classification algorithm or statistical algorithms are not
three positions. The pulse preprocessing methods composed of able to distinguish the healthy subjects from the diabetes
denoising, baseline wanders removal, and period segmentation patients at a high accuracy. A comparison between linear-SVM
are adopted. For the feature extraction, we first choose the classifier and RBF–SVM classifier is implemented.
2030 IEEE TRANSACTIONS ON INSTRUMENTATION AND MEASUREMENT, VOL. 64, NO. 7, JULY 2015
TABLE VII
C LASSIFICATION R ESULTS
Table VII shows the classification rates between healthy Fig. 20. PCA 2-D plot of the multichannel fusion features for diabetes
diagnosis.
subjects and diabetes by using independent single-pulse
features and fusion features, respectively. It can be seen that
any of these independent pulse features is enough for the pulse
diagnosis between the two patterns. Meanwhile, the relation-
ship between healthy and diabetes is complicated. The linear
classification algorithm used to distinguish the healthy subjects
from the diabetes patients is not able to achieve satisfactory
results as kernel method.
With the fusion features, we select the SVM with RBF Fig. 21. Loadings of PCA1.
kernel for the classification between health and diabetes
through the comparison. To construct the classifiers, we adopt is the pulse principle shape features of Guan channel, the
the fivefold cross-validation method. The healthy and diabetes beginning features interval is the pulse principle shape features
samples are randomly divided into five equal folds, respec- of Cun channel, and the end part is the pulse features of
tively. Each fold contains 25 healthy samples and 25 diabetes Chi channel. The loadings of PCA1 are concentrated on pulse
subjects. Then, we select four of them each time to form the shape features of pressure Guan pulse, pressure Cun pulse,
training set and leave one as the testing set for verification. and photoelectric Guan pulse. The Chi channel contributes
The specifications of the kernel are determined by the grid- least to the PCA1. Meanwhile, the wavelet energy features
searching technique. We obtain the final accuracy by averaging also have effect on the PCA1. To sum up, the pulse shape
the classification accuracy of the fivefold cross validation as features and wavelet energies extracted from Cun and Guan
the indicator of system performance. channel is clearly the most important separation factor between
To test the performance of fusion sensors, we compare healthy and diabetes patients. Because modern clinical studies
and list the classification accuracy between single-type pulse demonstrate that there is premature loss of arterial elasticity
waveforms and the combination pulse waveforms, respectively. and endothelial function for patients with diabetes [48].
Moreover, the classification results between the traditional The selected features of PCA1 mainly reflect the global
single channel pulse waveforms and the multichannel pulse shape information, and the detail information such as
waveforms are also presented to test the superiority of the secondary shape features are not considered. Therefore, using
novel design. Fig. 20 shows a 2-D PCA analysis of the the only single PCA1 is not enough to well distinguish the
fusion feature distribution with the first principal component healthy patients from the diabetes group. The secondary com-
(PC1) plotted against the second (PC2). The ∗ represents ponents also play an important part in pulse pattern recognition
the diabetes samples and + the healthy samples. The two and improve the classification rates.
dimensions explained 23.7% of the variation in the data, 14.6% These experiments have proved the significance of the
for PC1, and 9.1% for PC2 component. The two classes are TCPD and verified the theory of Three Parts with Nine
discriminative in spite of some feature overlaps. Palpation Sites, which demands that pulse from Cun, Guan,
To see more clearly which features contribute to PCA1, and Chi should be considered as a whole effect for physiolog-
we provide insight into the loadings of the PCA1 as shown ical status diagnosis.
in Fig. 21. As mentioned previously, the feature vector is
composed of 1608 elements for one subject from three V. C OMPARISON AND A NALYSIS
channels. At each channel, the feature dimension is 536,
including 268 pressure features and 268 photoelectric features, A. Advantages of the Proposed Multichannel Pulse System
respectively. The feature is fusion is fused in a serial strategy. We develop a novel wrist pulse system which solves the
It can be seen that six parts have obvious influence and problems and disadvantages in the previous pulse-taking plat-
weights for PCA1. The obvious feature interval at the middle forms. For the sensor selection stage, the traditional strain
WANG et al.: NOVEL MULTICHANNEL WRIST PULSE SYSTEM 2031
[15] C.-S. Hu, Y.-F. Chung, C.-C. Yeh, and C.-H. Luo, “Temporal and spatial [37] K. Q. Wang, L. S. Xu, L. Wang, Z. G. Li, and Y. Z. Li, “Pulse base-
properties of arterial pulsation measurement using pressure sensor array,” line wander removal using wavelet approximation,” in Proc. Comput.
Evidence-Based Complementary Alternative Med., vol. 2012, pp. 1–9, Cardiol., Sep. 2003, pp. 605–608.
May 2011, Art. ID 745127. [38] L. Xu, D. Zhang, and K. Wang, “Wavelet-based cascaded adaptive filter
[16] S. Charbonnier, S. Galichet, G. Mauris, and J. P. Siche, “Statistical and for removing baseline drift in pulse waveforms,” IEEE Trans. Biomed.
fuzzy models of ambulatory systolic blood pressure for hypertension Eng., vol. 52, no. 11, pp. 1973–1975, Nov. 2005.
diagnosis,” in Proc. 16th IEEE Instrum. Meas. Technol. Conf. (IMTC), [39] L. Xu, K. Wang, and D. Zhang, “Modern researches on pulse waveform
May 1999, pp. 103–108. of TCPD,” in Proc. IEEE Int. Conf. Commun., Circuits, Syst. West Sino
[17] S. Ahmad, M. Bolic, H. Dajani, V. Groza, I. Batkin, and S. Rajan, “Mea- Expo., Jun./Jul. 2002, pp. 1073–1077.
surement of heart rate variability using an oscillometric blood pressure [40] L. S. Xu, K. Q. Wang, and L. Wang, “Pulse waveforms classification
monitor,” IEEE Trans. Instrum. Meas., vol. 59, no. 10, pp. 2575–2590, based on wavelet network,” in Proc. IEEE-EMBS 27th Annu. Int. Conf.
Oct. 2010. Eng. Med. Biol. Soc., Jan. 2006, pp. 4596–4599.
[18] M. Saito, M. Matsukawa, T. Asada, and Y. Watanabe, “Noninvasive [41] L. Wang, K.-Q. Wang, and L.-S. Xu, “Recognizing wrist pulse wave-
assessment of arterial stiffness by pulse wave analysis,” IEEE Trans. forms with improved dynamic time warping algorithm,” in Proc. Int.
Ultrason., Ferroelectr., Freq. Control, vol. 59, no. 11, pp. 2411–2419, Conf. Mach. Learn. Cybern., Aug. 2004, pp. 3644–3649.
Nov. 2012. [42] M. F. O’Rourke, A. Pauca, and X.-J. Jiang, “Pulse wave analysis,” Brit.
[19] P. Dupuis and C. Eugène, “Combined detection of respiratory and J. Clin. Pharmacol., vol. 51, no. 6, pp. 507–522, 2001.
cardiac rhythm disorders by high-resolution differential cuff pres- [43] L. Wang, K. Q. Wang, and L. S. Xu, “Lempel–Ziv decomposition based
sure measurement,” IEEE Trans. Instrum. Meas., vol. 49, no. 3, arrhythmic pulses recognition,” in Proc. 27th Int. Conf. IEEE Eng. Med.
pp. 498–502, Jun. 2000. Biol. Soc., Shanghai, China, Sep. 2005, pp. 4606–4609.
[20] J. M. Kang, T. Yoo, and H. C. Kim, “A wrist-worn integrated health [44] L. Wang, K. Q. Wang, and L. S. Xu, “Lempel–Ziv decomposition based
monitoring instrument with a tele-reporting device for telemedicine and arrhythmic pulses recognition,” in Proc. IEEE-EMBS 27th Annu. Int.
telecare,” IEEE Trans. Instrum. Meas., vol. 55, no. 5, pp. 1655–1661, Conf. Eng. Med. Biol. Soc., Jan. 2006, pp. 4606–4609.
Oct. 2006. [45] M. R. Ram, K. V. Madhav, E. H. Krishna, N. R. Komalla, K. Sivani,
[21] H. M. Haqqani, J. B. Morton, and J. M. Kalman, “Using the and K. A. Reddy, “ICA-based improved DTCWT technique for MA
12-lead ECG to localize the origin of atrial and ventricular tachycardias: reduction in PPG signals with restored respiratory information,” IEEE
Part 2—Ventricular tachycardia,” J. Cardiovascular Electrophysiol., Trans. Instrum. Meas., vol. 62, no. 10, pp. 2639–2651, Oct. 2013.
vol. 20, no. 7, pp. 825–832, 2009. [46] K. A. Reddy, B. George, and V. J. Kumar, “Use of Fourier series
[22] K. Humphreys, T. Ward, and C. Markham, “Noncontact simultane- analysis for motion artifact reduction and data compression of photo-
ous dual wavelength photoplethysmography: A further step toward plethysmographic signals,” IEEE Trans. Instrum. Meas., vol. 58, no. 5,
noncontact pulse oximetry,” Rev. Sci. Instrum., vol. 78, no. 4, pp. 1706–1711, May 2009.
pp. 044304-1–044304-6, 2007. [47] P.-Y. Zhang and H.-Y. Wang, “A framework for automatic time-domain
[23] C.-C. Tyan, S.-H. Liu, J.-Y. Chen, J.-J. Chen, and W.-M. Liang, characteristic parameters extraction of human pulse signals,” EURASIP
“A novel noninvasive measurement technique for analyzing the pressure J. Adv. Signal Process., vol. 2008, p. 468390, Nov. 2008.
pulse waveform of the radial artery,” IEEE Trans. Biomed. Eng., vol. 55, [48] Y. Chen, L. Zhang, D. Zhang, and D. Zhang, “Wrist pulse signal diagno-
no. 1, pp. 288–297, Jan. 2008. sis using modified Gaussian models and fuzzy C-means classification,”
[24] L. Xu, D. Zhang, K. Wang, and L. Wang, “Arrhythmic pulses detec- Med. Eng. Phys., vol. 31, no. 10, pp. 1283–1289, 2009.
tion using Lempel–Ziv complexity analysis,” EURASIP J. Adv. Signal [49] M. R. Neuman, “Measurement of blood pressure [tutorial],” IEEE Pulse,
Process., vol. 2006, p. 018268, Mar. 2006. vol. 2, no. 2, pp. 39–44, Mar./Apr. 2011.
[25] K. Wang, C. Xu, D. Zhang, and N. Li, “Analysis of Doppler ultrasound [50] T. Kageyama, M. Kabuto, T. Kaneko, and N. Nishikido, “Accuracy of
blood flow signals of wrist radial artery for discriminating healthy people pulse rate variability parameters obtained from finger plethysmogram:
from 3 kinds of patients,” in Proc. 21st IEEE Int. Symp. Comput.-Based A comparison with heart rate variability parameters obtained from
Med. Syst. (CBMS), Jun. 2008, pp. 87–89. ECG,” J. Occupat. Health, vol. 39, no. 2, pp. 154–155, 1997.
[26] W. C. Tang and H. J. Sun, “The detected method of multipath of pulse [51] S. Lukman, Y. He, and S.-C. Hui, “Computational methods for tra-
conditions and research of transducer,” Chin. J. Traditional Med. Sci. ditional Chinese medicine: A survey,” Comput. Methods Programs
Technol., vol. 7, no. 5, pp. 319–320, 2000. Biomed., vol. 88, no. 3, pp. 283–294, Dec. 2007.
[52] H.-T. Wu, C.-H. Lee, C.-K. Sun, J.-T. Hsu, R.-M. Huang, and
[27] L. Xu, K. Wang, and D. Zhang, “Modern research on traditional Chinese
pulse diagnosis,” Eur. J. Oriental Med., vol. 4, no. 6, pp. 46–54, 2004. C.-J. Tang, “Arterial waveforms measured at the wrist as indicators of
diabetic endothelial dysfunction in the elderly,” IEEE Trans. Instrum.
[28] L. Xu, K. Wang, D. Zhang, Y. Li, Z. Wan, and J. Wang, “Objecti-
Meas., vol. 61, no. 1, pp. 162–169, Jan. 2012.
fying researches on traditional Chinese pulse diagnosis,” Inform. Med.
[53] H.-T. Wu et al., “Arterial stiffness using radial arterial waveforms
Slovenica, vol. 8, no. 1, pp. 56–63, 2003.
measured at the wrist as an indicator of diabetic control in the elderly,”
[29] X. Lisheng, W. Kuanquan, D. Zhang, and S. Cheng, “Adaptive baseline
IEEE Trans. Biomed. Eng., vol. 58, no. 2, pp. 243–252, Feb. 2011.
wander removal in the pulse waveform,” in Proc. 15th IEEE Symp.
[54] I. B. Wilkinson et al., “Increased augmentation index and systolic stress
Comput.-Based Med. Syst. (CBMS), Jun. 2002, pp. 143–148.
in type 1 diabetes mellitus,” QJM, vol. 93, no. 7, pp. 441–448, 2000.
[30] A. J. Joshi, S. Chandran, V. K. Jayaraman, and B. D. Kulkarni,
“Multifractality in arterial pulse,” in Proc. 19th Int. Conf. Pattern
Recognit. (ICPR), Dec. 2008, pp. 1–4.
[31] K.-Q. Wang, L.-S. Xu, D. Zhang, and C. Shi, “TCPD based pulse
monitoring and analyzing,” in Proc. Int. Conf. Mach. Learn. Cybern.,
2002, pp. 1366–1370.
[32] M. Nitzan, “Automatic noninvasive measurement of arterial blood pres-
sure,” IEEE Instrum. Meas. Mag., vol. 14, no. 1, pp. 32–37, Feb. 2011.
[33] Y. Hata, Y. Kamozaki, T. Sawayama, K. Taniguchi, and H. Nakajima,
“A heart pulse monitoring system by air pressure and ultrasonic sensor
systems,” in Proc. IEEE Int. Conf. Syst. Syst. Eng. (SoSE), Apr. 2007,
pp. 1–5. Dimin Wang received the B.S. degree in
[34] K. Wang, L. Xu, Z. Li, D. Zhang, N. Li, and S. Wang, “Approximate electronic engineering from Tsinghua University,
entropy based pulse variability analysis,” in Proc. 16th IEEE Symp. Beijing, China, in 2009, where he is currently
Comput.-Based Med. Syst., Jun. 2003, pp. 236–241. pursuing the Ph.D. degree with the Department of
[35] G. C. Jin, M. Yu, and N. K. Bao, “Research of multi-point pulse wave Electronic Engineering.
computer measurement system using PVDF,” J. Tsinghua Univ. Sci. He was a Research Assistant with the Department
Technol., vol. 39, no. 8, pp. 117–120, 1999. of Computing, Hong Kong Polytechnic University,
[36] M. R. Ram, K. V. Madhav, E. H. Krishna, N. R. Komalla, and Hong Kong, from 2010 to 2011. His current
K. A. Reddy, “A novel approach for motion artifact reduction in PPG research interests include signal processing, pattern
signals based on AS-LMS adaptive filter,” IEEE Trans. Instrum. Meas., recognition, machine learning, and computerized
vol. 61, no. 5, pp. 1445–1457, May 2012. medical diagnosis.
2034 IEEE TRANSACTIONS ON INSTRUMENTATION AND MEASUREMENT, VOL. 64, NO. 7, JULY 2015
David Zhang (F’09) received the M.Sc. degree in Guangming Lu received the B.S. Degree in electri-
computer science from Peking University, Beijing, cal engineering, the master’s degree in control theory
China, in 1982, the Ph.D. degree in computer sci- and control engineering, and the Ph.D. degree in
ence from the Harbin Institute of Technology (HIT), computer science and engineering from the Harbin
Harbin, in 1985, and the Ph.D. degree in electrical Institute of Technology (HIT), Harbin, China, in
and computer engineering from the University of 1998, 2000, and 2005, respectively.
Waterloo, Waterloo, ON, Canada, in 1994. He was a Post-Doctoral Fellow with Tsinghua Uni-
He was a Post-Doctoral Fellow with Tsinghua versity, Beijing, China, from 2005 to 2007. He is cur-
University, Beijing, and then an Associate Professor rently a Professor with the Biocomputing Research
with Academia Sinica, Beijing, from 1986 to 1988. Center, Shenzhen Graduate School, HIT. His current
He has been the Chair Professor with the Hong Kong research interests include pattern recognition, image
Polytechnic University, Hong Kong, since 2005, where he was the Founding processing, and automated biometric technologies and applications.
Director of the Biometrics Research Centre supported by the Hong Kong
Special Administrative Region Government in 1998. He also serves as the
Visiting Chair Professor with Tsinghua University, and an Adjunct Professor
with Peking University, Shanghai Jiao Tong University, Shanghai, China,
HIT, and the University of Waterloo. He has selected as a 2014 Highly
Cited Researcher by Thomson Reuters. He has authored over 10 books,
over 300 international journal papers, and holds around 30 patents from the
U.S./Japan/Honk Kong/China.
Prof. Zhang is a Croucher Senior Research Fellow, Distinguished Speaker
of the IEEE Computer Society, and fellow of the International Association
for Pattern Recognition. He is the Founder and Editor-in-Chief of the Inter-
national Journal of Image and Graphics, a Book Editor of the International
Series on Biometrics (Springer), an Organizer of the International Conference
on Biometrics Authentication, and an Associate Editor of over 10 international
journals, including the IEEE transactions.