穿戴式醫療監測裝置例如光體積描記波形生理感測器已被廣泛應用在臨床治療及居家照護。然而，對於利用穿戴式光體積描記波形生理感測器紀錄生理資訊而言，動作雜訊(motion artifact)的影響是一個嚴重的問題。因此在論文中，為了改善動作雜訊的問題，我們提出了一個基於光體積描記波形生理感測器之具動作雜訊偵測及去除的生理資訊萃取方法。
本論文提出的方法包含三個部分。第一個部分為光體積描記波形訊號之前處理，所選用的處理方法為離散小波轉換。利用離散小波轉換，萃取光體積描記波形訊號的直流及交流成分，並且偵測訊號的峰值，以便計算偵測動作雜訊之特徵值及生理資訊。第二個部分為特徵值萃取，目的是為了偵測光體積描記波形訊號中受動作雜訊干擾的區段。將接收到的訊號分段後，計算的四種特徵值，包含峰對峰值之標準差、峰對峰間距之標準差、峰對峰值之平均絕對值偏差及峰度(Kurtosis)。之後使用支持向量機器(SVM)偵測動作雜訊的偵測器，將四種特徵值當作支持向量機器的輸入資料，便可以藉由這些特徵值偵測動作雜訊的區段。為了驗證偵測的結果，論文中記錄11位健康受試者的光體積描記波形訊號進行分析，其中包含揮手所造成的動作雜訊。實驗結果顯示，偵測準確度(accuracy)達94.40%，敏感度(sensitivity)達90.35%，特異度(specificity)達99.36%。
第三部分為生理資訊萃取及動作雜訊消除。所萃取的生理資訊包含血氧飽和濃度(SpO2)和心率(HR)。由於動作雜訊會造成血氧飽和濃度(SpO2)和心率(HR)讀值的誤差，論文中採用卡爾曼濾波器(Kalman filter)修正讀值的偏差來達到動作雜訊消除的目的。藉由不同的動作雜訊偵測結果調整卡爾曼濾波器的參數，並以實驗來驗證結果。在左右揮手的實驗中，血氧飽和濃度及心率的平均絕對值偏差由1.34%及7.29 bpm修正為0.8%及4.29 bpm。而在上下揮手的實驗中，血氧飽和濃度及心率的平均絕對值偏差由1.31%及13.97 bpm修正為0.82%及6.87 bpm。

A wearable health-monitoring device such as photoplethysmography (PPG) is widespread
in clinical application and in-home care. However, the effect of motion artifacts is a big
problem. In this thesis, a method that is used to extract physiological information from PPG
signal with motion artifacts detect and removal is proposed.
The procedures of our method contain three parts. First, the PPG data which are recorded
from the wearable transmission type PPG sensor are preprocessed by discrete wavelet trans-
form (DWT) in order to remove some unwanted noise and extract AC and DC component of
PPG signal. Then, the characteristic points such as peaks and troughs are identified for the fea-
ture extraction and physiological information extraction. The second part is feature extraction
and motion artifact detection. The features include four time domain parameters, which are
standard deviation of peak-to-peak amplitudes, standard deviation of peak-to-peak intervals,
mean absolute deviation of peak-to-peak amplitudes, and the kurtosis of the signal segments.
To detect the motion artifact periods, we employ support vector machine (SVM) to classify.
The detection performance was verified on PPG signals recording by the 11 different healthy
subjects with waving hands. The detection method gives the best performance in 7 second
period with accuracy of 94.4%, the sensitivity of 90.35%, and the specificity of 99.36%.
The third part is arterial oxygen saturation (SpO 2 ), heart rate (HR) extraction and motion
artifact removal. The motion artifact removal part is accomplished by using Kalman filter to
track the SpO 2 and HR extracted from motion artifact-corrupted periods. The parameters of
Kalman filter are determined by the detection results. In the case of waving hand left-right,
the average mean absolute bias of artifact-corrupted SpO 2 and HR are 1.34% and 7.29 bpm,
respectively. After applying the algorithm, the bias become 0.8% and 4.29 bpm. In the case
of waving hand up and down, the errors of SpO 2 and HR reduce from 1.31% to 0.82% and 13.97 bpm to 6.87 bpm.

Abstract i
1 Introduction 1
1.1 Backgrounds . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.2 Motivation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
1.3 Photoplethysmography Overview . . . . . . . . . . . . . . . . . . . . . . . . 2
1.3.1 Operating Modes of Pulse Oximeter . . . . . . . . . . . . . . . . . . 3
1.3.2 Application of PPG in Physiological Measurement . . . . . . . . . . 4
1.4 Main Contributions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
1.5 Organization of the Thesis . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
2 Overview of PPG Signal Processing 7
2.1 Moving Average Filter . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
2.2 Fourier Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
2.3 Adaptive Filter . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
2.4 Independent Component Analysis . . . . . . . . . . . . . . . . . . . . . . . 10
2.5 Time-Frequency Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
2.6 Comparison of Different Signal Processing Methods . . . . . . . . . . . . . 11
3 Physiological Information Extraction with Proposed Method for Motion Artifact
Detection and Removal 13
3.1 Physiological Information Extraction . . . . . . . . . . . . . . . . . . . . . . 13
3.2 Wavelet Transform . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
3.2.1 Introduction to Wavelet Transform . . . . . . . . . . . . . . . . . . . 16
3.2.2 Continuous Wavelet Transform . . . . . . . . . . . . . . . . . . . . 18
3.2.3 Discrete Wavelet Transform . . . . . . . . . . . . . . . . . . . . . . 18
3.2.4 Noise Reduction Based on Wavelet Transform . . . . . . . . . . . . 20
3.2.5 Selecting Threshold . . . . . . . . . . . . . . . . . . . . . . . . . . 20
3.2.6 Thresholding Methods . . . . . . . . . . . . . . . . . . . . . . . . . 22
3.2.7 Peak Detection Method for PPG Signals . . . . . . . . . . . . . . . . 24
3.3 Overview of Proposed Method . . . . . . . . . . . . . . . . . . . . . . . . . 26
3.4 Motion Artifact Detection . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28
3.4.1 Feature Extraction . . . . . . . . . . . . . . . . . . . . . . . . . . . 28
3.4.2 Feature Normalization . . . . . . . . . . . . . . . . . . . . . . . . . 30
3.4.3 Support Vector Machine . . . . . . . . . . . . . . . . . . . . . . . . 32
3.4.4 Linearly Separable Binary Classication . . . . . . . . . . . . . . . . 32
3.4.5 Linearly Non-Separable Binary Classication . . . . . . . . . . . . . . 35
3.4.6 Non-Linear Support Vector Machine . . . . . . . . . . . . . . . . . . 37
3.4.7 Cross-validation . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39
3.5 Motion Artifact Removal . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41
3.5.1 Introduction to Kalman filter . . . . . . . . . . . . . . . . . . . . . . 41
3.5.2 Kalman Filtering Initialization and Operation . . . . . . . . . . . . . 43
4 Experimental Results and Comparison 47
4.1 Experimental Protocol . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47
4.1.1 Wearable PPG Sensor . . . . . . . . . . . . . . . . . . . . . . . . . 47
4.1.2 PPG Data Collection . . . . . . . . . . . . . . . . . . . . . . . . . . 48
4.1.3 Real-World Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48
4.1.4 Data from Pulse Oximeter Tester . . . . . . . . . . . . . . . . . . . . 50
4.2 Performance of SpO 2 and HR Extraction . . . . . . . . . . . . . . . . . . . . 52
4.3 Results of Motion Artifact Detection . . . . . . . . . . . . . . . . . . . . . . 54
4.3.1 Statistical Measures of the Performance for a Binary Classification . . 54
4.3.2 Optimization of Regularization Parameters of SVM . . . . . . . . . . 56
4.3.3 Results of Segment Length for Classification . . . . . . . . . . . . . 56
4.4 Results of Motion Artifact Removal . . . . . . . . . . . . . . . . . . . . . . 57
4.5 Comparisons . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59
5 Conclusion and Future Work 63
5.1 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63
5.2 Future Work . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64

[1] V. F. Blanc, M. Haig, M. Troli, and B. Sauv´e, “Computerized photo-plethysmography of the finger,” Canadian Journal of Anaesthesia, vol. 40, no. 3, pp. 271–278, 1993.
[2] A. A. Awad, M. A. M. Ghobashy, R. G. Stout, D. G. Silverman, and K. H. Shelley, “How does the plethysmogram derived from the pulse oximeter relate to arterial blood pressure in coronary artery bypass graft patients?” Anesthesia & Analgesia, vol. 93, no. 6, pp.
1466–1471, 2001.
[3] P. J. Strollo Jr and R. M. Rogers, “Obstructive sleep apnea,” New England Journal of Medicine, vol. 334, no. 2, pp. 99–104, 1996.
[4] P. A. Leonard, J. G. Douglas, N. R. Grubb, D. Clifton, P. S. Addison, and J. N. Watson,
“A fully automated algorithm for the determination of respiratory rate from the photoplethysmogram,” Journal of Clinical Monitoring and Computing, vol. 20, no. 1, pp.
33–36, 2006.
[5] J. Y. A. Foo and S. J. Wilson, “Estimation of breathing interval from the photoplethysmographic signals in children,” Physiological Measurement, vol. 26, no. 6, pp. 1049–1458,
2005.
[6] T. Tamura, Y. Maeda, M. Sekine, and M. Yoshida, “Wearable photoplethysmographic sensors–past and present,” Electronics, vol. 3, no. 2, pp. 282–302, 2014.
[7] K. Nakajima, T. Tamura, and H. Miike, “Monitoring of heart and respiratory rates by photoplethysmography using a digital filtering technique,” Medical Engineering &
Physics, vol. 18, no. 5, pp. 365–372, 1996.
[8] A. Johansson, P.A. Oberg, and G. Sedin, “Monitoring of heart and respiratory rates in newborn infants using a new photoplethysmographic technique,” Journal of Clinical
Monitoring and Computing, vol. 15, no. 7-8, pp. 461–467, 1999.
[9] C. Yu, Z. Liu, T. McKenna, A. T. Reisner, and J. Reifman, “A method for automatic identification of reliable heart rates calculated from ECG and PPG waveforms,” Journal of the American Medical Informatics Association, vol. 13, no. 3, pp. 309–320, 2006.
[10] J.-Y. Foo, C.-S. Lim, and P. Wang, “Evaluation of blood pressure changes using vascular
transit time,” Physiological Measurement, vol. 27, no. 8, pp. 685–694, 2006.
[11] D. Barschdorff and W. Zhang, “Respiratory rhythm detection with photoplethysmo-
graphic methods,” in IEEE International Conference on Engineering in Medicine and
Biology Society, 1994, pp. 912–913.
[12] E. Lee, N. Kim, N. Trang, J. Hong, E. Cha, and T. Lee, “Respiratory rate detection algorithms by photoplethysmography signal processing,” in IEEE International Conference on Engineering in Medicine and Biology Society, 2008, pp. 1140–1143.
[13] H. Lee, J. Lee, W.-G. Jung, and G.-K. Lee, “The periodic moving average filter for removing motion artifacts from ppg signals,” International Journal of Control Automation
and Systems, vol. 5, no. 6, pp. 701–706, 2007.
[14] K. A. Reddy, B. George, and V. J. Kumar, “Use of fourier series analysis for motion artifact reduction and data compression of photoplethysmographic signals,” IEEE Transactions on Instrumentation and Measurement, vol. 58, no. 5, pp. 1706–1711, 2009.
[15] H. H. Asada, H.-H. Jiang, and P. Gibbs, “Active noise cancellation using mems accelerometers for motion-tolerant wearable bio-sensors,” in IEEE International Conference on Engineering in Medicine and Biology Society, vol. 1, 2004, pp. 2157–2160.
[16] G. Comtois, Y. Mendelson, and P. Ramuka, “A comparative evaluation of adaptive noise
cancellation algorithms for minimizing motion artifacts in a forehead-mounted wearable pulse oximeter,” in IEEE International Conference on Engineering in Medicine and
Biology Society, 2007, pp. 1528–1531.
[17] R. Yousefi, M. Nourani, S. Ostadabbas, and I. Panahi, “A motion-tolerant adaptive algorithm for wearable photoplethysmographic biosensors,” IEEE Journal of Biomedical and Health Informatics, vol. 18, no. 2, pp. 670–681, 2014.
[18] M. Ram, K. V. Madhav, E. H. Krishna, N. R. Komalla, and K. A. Reddy, “A novel approach for motion artifact reduction in PPG signals based on AS-LMS adaptive filter,” IEEE Transactions on Instrumentation and Measurement, vol. 61, no. 5, pp. 1445–1457, 2012.
[19] P. Comon, “Independent component analysis, a new concept?” Signal Processing,
vol. 36, no. 3, pp. 287–314, 1994.
[20] B. S. Kim and S. K. Yoo, “Motion artifact reduction in photoplethysmography using independent component analysis,” IEEE Transactions on Biomedical Engineering, vol. 53, no. 3, pp. 566–568, 2006.
[21] J. Yao and S. Warren, “A short study to assess the potential of independent component analysis for motion artifact separation in wearable pulse oximeter signals,” in IEEE International Conference on Engineering in Medicine and Biology Society, 2005, pp.
3585–3588.
[22] C. Lee and Y. Zhang, “Reduction of motion artifacts from photoplethysmographic recordings using a wavelet denoising approach,” in IEEE International Conference on
Engineering in Medicine and Biology Society, 2003, pp. 194–195.
[23] Y.-S. Yan, C. C. Poon, and Y.-T. Zhang, “Reduction of motion artifact in pulse oximetry by smoothed pseudo wigner-ville distribution,” Journal of NeuroEngineering and
Rehabilitation, vol. 2, p. 3, 2005.
[24] J. A. Pologe, “Pulse oximetry: technical aspects of machine design,” International Anesthesiology Clinics, vol. 25, no. 3, pp. 137–153, 1987.
[25] T. Rusch, R. Sankar, and J. Scharf, “Signal processing methods for pulse oximetry,”
Computers in Biology and Medicine, vol. 26, no. 2, pp. 143–159, 1996.
[26] C. Cai and P. d. B. Harrington, “Different discrete wavelet transforms applied to denoising analytical data,” Journal of Chemical Information and Computer Sciences, vol. 38,
no. 6, pp. 1161–1170, 1998.
[27] M. Raghuram, K. Madhav, E. Krishna, and K. Reddy, “On the performance of wavelets
in reducing motion artifacts from photoplethysmographic signals,” in IEEE International
Conference on Bioinformatics and Biomedical Engineering (iCBBE), 2010, pp. 1–4.
[28] Q. Li, R. G. Mark, and G. D. Clifford, “Robust heart rate estimation from multiple asynchronous noisy sources using signal quality indices and a kalman filter,” Physiological
Measurement, vol. 29, no. 1, p. 15, 2008.
[29] Q. Li, R. G. Mark, G. D. Clifford et al., “Artificial arterial blood pressure artifact models and an evaluation of a robust blood pressure and heart rate estimator,” Biomedical
Engineering Online, vol. 8, no. 1, p. 13, 2009.
[30] R. Krishnan, B. Natarajan, and S. Warren, “Two-stage approach for detection and reduction of motion artifacts in photoplethysmographic data,” IEEE Transactions on Biomedical Engineering, vol. 57, no. 8, pp. 1867–1876, 2010.
[31] N. Selvaraj, Y. Mendelson, K. H. Shelley, D. G. Silverman, and K. H. Chon, “Statistical
approach for the detection of motion/noise artifacts in photoplethysmogram,” in IEEE
International Conference on Engineering in Medicine and Biology Society, 2011, pp.
4972–4975.
[32] Wikipedia Contributors. (2014, September) Kurtosis. [Online]. Available: http://en.wikipedia.org/wiki/Kurtosis
[33] A. Stolcke, S. Kajarekar, and L. Ferrer, “Nonparametric feature normalization for SVM-
based speaker verification,” in IEEE International Conference on Acoustics, Speech and
Signal Processing, 2008, pp. 1577–1580.
[34] S. Theodoridis and K. Koutroumbas, “Pattern recognition,” IEEE Transactions on Neural Networks, vol. 19, no. 2, p. 376, 2008.
[35] A. Ben-Hur and J. Weston, “A user’s guide to support vector machines,” in Data Mining
Techniques for the Life Sciences. Springer, 2010, pp. 223–239.
[36] C.-C. Chang and C.-J. Lin, “Libsvm: a library for support vector machines,” ACM Trans-
actions on Intelligent Systems and Technology (TIST), vol. 2, no. 3, p. 27, 2011.
[37] C.-W. Hsu, C.-C. Chang, and C.-J. Lin, “A practical guide to support vector classifica-
tion,” National Taiwan University, Department of Computer Science, 2003.
[38] J. A. Sukor, S. Redmond, and N. Lovell, “Signal quality measures for pulse oximetry
through waveform morphology analysis,” Physiological Measurement, vol. 32, no. 3,
pp. 369–384, 2011.
[39] S. Boyd and L. Vandenberghe, Convex optimization. Cambridge University Press, 2009.
[40] J. Shawe-Taylor and N. Cristianini, Kernel methods for pattern analysis. Cambridge
university press, 2004.
[41] S. Geisser, Predictive Inference. Chapman and Hall/CRC Press, 1993, vol. 55.
[42] R. R. Picard and R. D. Cook, “Cross-validation of regression models,” Journal of the
American Statistical Association, vol. 79, no. 387, pp. 575–583, 1984.
[43] G. McLachlan, K.-A. Do, and C. Ambroise, Analyzing microarray gene expression data.
John Wiley & Sons, 2005, vol. 422.
[44] R. G. Brown, Introduction to random signal analysis and Kalman filtering. Wiley New
York, 1983, vol. 8.
[45] G. Welch and G. Bishop, “An introduction to the Kalman filter,” University of North
Carolina; Chapel Hill, 2006.
[46] P. Andersen and B. Saltin, “Maximal perfusion of skeletal muscle in man.” The Journal
of Physiology, vol. 366, no. 1, pp. 233–249, 1985.
[47] J. Mietus, C. Peng, I. Henry, R. Goldsmith, and A. Goldberger, “The pnnx files: reexamining a widely used heart rate variability measure,” Heart, vol. 88, no. 4, pp. 378–380, 2002.
[48] SPOT Light SpO2 Functional Tester Users Manual, FLUKE Biomedical, 2012, rev. 1.
[49] J. W. Chong, D. K. Dao, S. Salehizadeh, D. D. McManus, C. E. Darling, K. H. Chon, and
Y. Mendelson, “Photoplethysmograph signal reconstruction based on a novel hybrid mo-
tion artifact detection–reduction approach. part I: Motion and noise artifact detection,”
Annals of Biomedical Engineering, pp. 1–13, 2014.
[50] A. Louw, C. Cracco, C. Cerf, A. Harf, P. Duvaldestin, F. Lemaire, and L. Brochard,
“Accuracy of pulse oximetry in the intensive care unit,” Intensive Care Medicine, vol. 27,
no. 10, pp. 1606–1613, 2001.
[51] J. W. Chong, D. K. Dao, S. Salehizadeh, D. D. McManus, C. E. Darling, K. H. Chon, and
Y. Mendelson, “Photoplethysmograph signal reconstruction based on a novel hybrid motion artifact detection–reduction approach. part II: Motion and noise artifact removal,” Annals of Biomedical Engineering, pp. 1–13, 2014.