An Automatic Peak Identification Method for Photoplethysmography Signals
LI Su-yi1，XU Zhuang1，XIONG Wen-ji2，JIANG Shan-qing1，WU Jiang1*
1. College of Instrumentation and Electrical Engineering，Jilin University，Changchun 130026，China
2. The First Hospital of Jilin University，Changchun 130021，China
Abstract：The automatic recognition of the peak point of the PPG signal is directly related to the accuracy of non-invasive measurement of blood oxygen saturation and the extraction of PP intervals. In this paper, a combined wavelet processing method was proposed. Based on the principle of wavelet multi-resolution analysis, the proposed method corrected the baseline wander which would influence the peak amplitude, and then the peak was identified automatically by using the quadratic spline wavelet modulus maximum algorithm. Using the signals acquired by a self-developed pulse oxygen saturation detecting device to evaluate the effectiveness, the method could correct the baseline wander and identified the peak points of the signal, and to evaluate the stability and reliability of the method, we used a noisy signal. Furthermore, by using ten segments of the measured PPG signals, we compared the peak recognition accuracy of the proposed method with that of a traditional differential threshold method to validate the effectiveness. The results showed that the method not only eliminated the baseline wander, but also could accurately detect the peak of the noisy signal, which had a good anti-jamming capability, and was beneficial to improve the detection of blood oxygen saturation and the accuracy of PP interval extraction. Furthermore, it was helpful to the evaluation of respiratory function and heart rate variability analysis.
Key words：PPG signal；Automatic recognition of the peak point；Wavelet multi-resolution analysis；Quadratic spline wavelet modulus maximum algorithm
[1] Khandoker A H，Karmakar C K，Palaniswami M. Medical Engineering and Physics，2011，33：204.
[2] Bulte C S E，Keet S W M，Boer C，et al. European Journal of Anaesthesiology，2013，68(7)：775.
[3] Schfer A，Vagedes J. International Journal of Cardiology，2013，166(1)：15.
[4] Kasambe P V，Rathod S S. Procedia Computer Science，2015，49(1)：282.
[5] Krishnan R，Natarajan B，Warren S. IEEE Transactions on Biomedical Engineering，2010，57(8)：1867.
[6] Murthy N K L，Madhusudana P C，Suresha P，et al. IEEE Signal Processing Letters，2015，22(12)：2391.
[7] Kavsaolu A R，Polat K，Bozkurt M R. Computers in Biology and Medicine，2014，49：1.
[8] Galeotti L，Scully C G，Vicente J，et al. Physiological Measurement，2015，36(8)：1.
[9] Mallat S G. IEEE Transactions on Pattern Analysis & Machine Intelligence，2010，11(7)：674.
[10] ZHAO Xue-zhi，XIANG Ke，YE Bang-yan，et al(赵学智，向 可，叶邦彦，等). Journal of Vibration and Shock(振动与冲击)，2010，29(11)：6.
