Photoplethysmography Feature Extraction for the Evaluation of Hypertension

ABSTRACT

Photo-plethysmography (PPG) is an optical technique which measures the blood volume changes in the arterial blood using red and IR LED’s of wavelength 660nm and 940nm. This paper proposes the methodology to measure the pulse rate from the video signal obtained using a LETV-LE MAX 2 mobile phone’s camera and also to evaluate hypertension. The Android Smartphone records the intensity of light reflected from the index finger. The recorded video is separated into red, green, blue frames. Since the red video frames returned the useful plethysmographic information, they are filtered using Butterworth band pass filter and performed power spectral density analysis. The immediate peak gives the pulse rate of the respective subjects. The fifteen features of pulse waveform are extracted and by performing feature selection process, seven features are selected and underwent classification process using neural network. The accuracy of the prediction was calculated for both normal and hypertensive subjects.

Introduction

According to Global Health Observatory (GHO) data, raised blood pressure causes estimated deaths of 7.5 million from 12.8% of the total of all deaths. In India, about 33% of urban people and 25% of rural people are hypertensive. The highest occurrence of hypertension across the WHO region was in Africa at 46% for both sexes. The prevalence of hypertension was lowest in America and it was 35% for both sexes. Men have higher prevalence than women in all WHO regions. About 55% of 17 million annual deaths are due to cardiovascular diseases which are linked to hypertension. Early diagnosis of hypertension will prevent the number of death cases which is possible by using this method. Home care devices enhance our health emotionally, physically and daily living needs in a professional manner. Home care devices provide more comfort and care to elderly people in their own pace in familiar surroundings. The physical strength of elderly people deteriorates which make them more dependent on low-cost, easy handling and convenient home care devices. Home care devices improve the patient experience by improving the quality of care and satisfaction.

The application of mobile phone in healthcare brings about a drastic improvement in medical sector. Android Smartphone decreases the patient’s inconveniences in hospitals and saves their time without disturbing their daily activities. It maintains an effective communication between patients, their caregivers and healthcare providers. The use of mobile application to measure the physiological signals of human body helps to monitor the chronic and acute ailments of patient by reducing their doctor visit and helps to improve the diagnostic accuracy. It can be used in the emergency situation and can be accessed from everywhere.

Plethysmographs are the instruments which are used in the detection of the pulse pressure and arterial pulse waveforms in the extremities. Most of the plethysmograph techniques available, respond to a change in the blood volume as a blood pressure measurement. When blood is ejected from the ventricles during the contraction of the heart, a pulse with a pressure is transmitted to the extremities through the circulatory system. The pulse pressure on travelling through the heart vessels causes displacement in the vessel-wall. This pressure is measured at various points in the extremities.

Discussions

The pulse rate measured using this proposed method is compared with the one measured from the pulse oximeter, which shows nearly the similar readings of the pulse rate. The pulse rates are also measured both from the normal and hypertensive patients using the pulse oximeter and smart phone.

Yongbo Liang, et al (2019) shows the assessment of hypertension using photoplethysmography by using feature selection process of six types. The correlation between the BP levels and PPG features was established based on hypertension risk stratification approach. PPG feature of Peak to peak amplitude correlates to the physiological phenomenon of systolic blood pressure. Greater the peak to peak amplitude, greater the chance of hypertension due to greater systolic blood pressure.

 Zhencheng Chen, et al (2018) explored the physiological changes that take place in different blood pressure level depending on the arterial wave propagation theory. The features of PPG show the condition of changes in the vascular tissue and volume of blood. Based on the BP levels, PPG features provide information about the changes in the vascular stiffness and compliance and also vascular aging. The features of PPG signal that is varying depending on the BP levels gives necessary information about the heart and the arteries.

Enock Jonathan et al (2010) used the green channel of the visible light that are reflected from the index finger for photo-plethysmographic (PPG) imaging with a mobile phone that are operated using video mode. Compared to this study we performed the pulse rate estimation process over the red channel of the visible light which shows better morphological information of PPG signal that is obtained from the video recorded using the android smart phone. Sasivimon Sukaphat, et al (2016) used the camera of Android Smart phone that are equipped with the LED flash light for recording the light intensity and concluded that the red channel in the RGB signals in each frame recorded for about 40 seconds using the smart phone returned the information of plethysmographic signal. Compared to this study our method requires only 30 seconds for determining the pulse rate from the video recorded using Smartphone. The p value obtained on performing the student’s t-test gives 0.6 which shows that there is no much difference between the pulse rate readings measured using the conventional pulse oximeter and Android Smartphone.

Mohamed Elgendi (2012) described the potential information that are embedded in the waveform of PPG signal and further attention for its applications that are beyond pulse oximetry and the heart-rate calculation and performing the analysis of ppg waveform and thereby describing its features. Yuriy Kurylyak, et al (2015), states that the blood pressure can be estimated from the pulse duration of PPG signal using the artificial neural network and the analysis was done by using 15000 heartbeats that are extracted from the intensive care database in which 21 parameters are extracted that can be used as an input data in artificial neural network. Compared to these studies we extracted 15 features from the PPG signal and by performing principal component analysis method, these 15 features are limited to seven features depending on the eigen values of PCA method. Also our work shows improvement in the performance of the classifier when compared to the above studies and got the accuracy of 83.3% when classifying the selected features from the video recorded using Smartphone.

From the seven selected features, the peak interval feature that is extracted from the PPG signal of recoded video shows the similar accuracy for the hypertensive patients as that of mentioned in the work done by Mohamed Elgendi. The feature settling time gives higher value for normal patients than the hypertensive patients as mentioned in Mohamed Elgendi work.

Conclusion

This project has shown the possibility of measuring the pulse rate and the evaluation of hypertension using the android smart phone’s camera. It comprises the stages of data acquisition using the smart phone in video imaging mode and processing the recorded video in MATLAB and thereby extracting the features and classifying them to predict hypertension. The pulse rate readings measured using this technique is compared with that of the one measured with the help of pulse oximeter and student t-test analysis is performed between the values measured using pulse oximeter and smart phone. The feature extraction and feature selection process are performed on the pulse sensor data and mobile phone data. Therefore the accuracy of prediction is determined.

References

1. R.S.Khandpur, “Handbook Of Biomedical Instrumentation”, McGraw Hill Education Private Limited, Third Edition, 2014.
2. Esrat Jahan, Tilottoma Barua, Umme Salma, “An Overview On Heart Rate Monitoring And Pulse Oximeter System”, International Journal of Latest Research in Science and Technology , Vol- 3, pp.148-152, 2014. www.scholarworks.rit.edu/10054.html (2015).
3. Zahari Taha, Lum Shirley And Mohd Azraai Mohd Razman, “A Review On Non-Invasive Hypertension Monitoring System By Using Photoplethysmography Method”, Malaysian journal of movement health and exercise, Vol-6, pp.110-116, 2017.
4. www.news-medical.net/health/Photoplethysmography-(PPG).aspx (2018).
5. www.ttu.ee/public/Teaduskond/menetlemisel/Publication_V.pdf (2012).
6. David Houcque, “Introduction To Matlab For Engineering Students”, Northwestern University, second edition, 2005. www.cimss.ssec.wisc.edu/wxwise/class/aos340/spr00/whatismatlab.html (2014).
7. E. Jonathan and M. Leahy, “Investigating a smartphone imaging unit for photoplethysmography”, Physiological Measurement, Vol-31, pp.79-83, 2010.
8. D. Grimaldi, Y. Kurylyak, F. Lamonaca, and A. Nastro, “Photoplethysmography detection by smartphone’s videocamera”, IEEE International Conference Intelligent Data Acquisition and Advanced Computing Systems, Ukraine, pp. 88–491, 2016.
9. Y. Kurylyak, F. Lamonaca, and D. Grimaldi, “Digital image andsignal processing for measurement systems”, River Publishers, fifth edition, pp. 135-164, 2012.
10. H. Kobayashi, “Effect of measurement duration on accuracy of pulsecounting”, Ergonomics, Vol-56, pp. 1940-1944, 2013.
11. Noor Kamal Al-Qazzaz, Israa F. Abdulazez, Salma A. Ridha, “Simulation Recording of an ECG, PCG, and PPG for Feature Extractions”, Al-Khwarizmi Engineering Journal, Vol-10, pp.81-91, 2014.
12. Alair Dias Junior, Srinivasan Murali, Francisco Rincon and David Atienza, “Estimation of Blood Pressure and Pulse Transit Time Using Your Smartphone”, Euromicro conference on digital system design, Greece, 2015.
13. Enock Jonathan and Martin J. Leahy, “Cellular phone-based photoplethysmographic imaging”, Journal of Biophotonics, Vol-5, pp.293-296, 2010 .
14. Elgendi M, Jonkman M, De Boer F, “Heart Rate Variability and Acceleration Plethysmogram measured at rest”, International Joint Conference on Biomedical Engineering System and Technologies, Portugal, pp. 266-77, 2011.
15. C.G. Scully, J. Lee, J. Meyer, A.M. Gorbach, D. Granquist-Fraser, Y.Mendelson, and H. Chon Ki, “Physiological parameter monitoring from optical recordings with a mobile phone”, IEEE Transactions on Biomedical Engineering, Vol-59, pp.303–306, 2012.
16. Mohammed elgendi, “On the Analysis of Fingertip Photoplethysmogram Signals”, Current cardiology reviews, Vol-8, pp.14-25, 2012.
17. Sasivimon Sukaphat, Sukapon Nanthachaiporn, Kritsadayu Upphaccha, Pinta Tantipatrakul, “Heart Rate Measurement on Android Platform”, journal of biophotonics, Vol-19, pp.251-256, 2016.
18. Zahari Taha, Lum Shirley and Mohd Azraai Mohd Razman, “A review on non-invasive hypertension monitoring system using photoplethysmography method”, Malaysian journal of movement, health and exercise, Vol-6, pp.47-57, 2017.
19. Yuriy Kurylyak, Francesco Lamonaca, Domenico Grimaldi, “A neural network-based method for continuous blood pressure estimation from a PPG signal”, International instrumentation and measurement technology conference, Minneapolis, pp.210-218, 2013.
20. www.who.int/gho/ncd/risk_factors/blood_pressure_prevalence_text/en.html (2013).
21. Yongbo Liang, Zhencheng Chen, Rabab Ward and Mohamed Elgendi, “ Hypertension assessment using photoplethysmography – a risk stratification approach”, Journal of clinical medicine, Vol-8(1), pp. 12, 2019.
22. Yongbo Liang, Zhencheng Chen, Rabab Ward and Mohamed Elgendi, “ Hypertension assessment via ECG and PPG signals: An evaluation using MIMIC database”, Diagnostics, Vol-8(3), pp.65, 2018.