1 Multi-parameter Respiratory Rate Estimation from the Photoplethysmogram Walter Karlen, Member, IEEE, Srinivas Raman, J. Mark Ansermino, and Guy A. Dumont, Fellow, IEEE Abstract—We present a novel method for estimating respira- tory rate in real-time from the photoplethysmogram (PPG) ob- tained from pulse oximetry. Three respiratory induced variations (frequency, intensity, and amplitude) are extracted from the PPG using the Incremental-Merge Segmentation algorithm. Frequency content of each respiratory induced variation is analyzed using Fast Fourier Transforms. The proposed Smart Fusion method then combines the results of the three respiratory induced variations using a transparent mean calculation. It automatically eliminates estimations considered to be unreliable because of detected presence of artifacts in the PPG or disagreement between the different individual respiratory rate estimations. The algorithm has been tested on data obtained from 29 children and 13 adults. Results show that it is important to combine the three respiratory induced variations for robust estimation of respiratory rate. The Smart Fusion showed trends of improved estimation (mean root mean square error 3.0 breaths/min) compared to the individual estimation methods (5.8, 6.2 and 3.9 breaths/min). The Smart Fusion algorithm is being implemented in a mobile phone pulse oximeter device to facilitate the diagnosis of severe childhood pneumonia in remote areas. Keywords-respiratory rate, pulse oximeter, photoplethysmo- gram, data fusion. I. I NTRODUCTION P NEUMONIA kills more than 2 million children under five years old every year [1]. Almost all of these deaths occur in the developing world. Many of these deaths could be prevented by early detection and timely administration of simple treatments. While the integration of technology into health care has greatly improved the speed and accuracy of diagnosis and treatment of childhood pneumonia in the developed world, lack of access to clinical expertise and costly tests often delay diagnosis and treatment, and reduce survival rates across the developing world. An essential criteria integrated in many guidelines for diagnosis of pneumonia in ill children is the assessment of an elevated respiratory rate of ≥ 40 breaths/min (age 1-5 years) [2]. However, clinical measurement of respiratory rate has been shown to have poor reliability and repeatability [3]. Copyright c 2013 IEEE. Personal use of this material is permitted. How- ever, permission to use this material for any other purposes must be obtained from the IEEE by sending an email to pubs-permissions@ieee.org. This work was supported by the Swiss National Science Foundation (SNSF) and the Canadian Institutes for Health Research (CIHR). W. Karlen and G.A. Dumont are with the Department of Electrical and Computer Engineering, The University of British Columbia (UBC), 2332 Main Mall, Vancouver BC, V6T 1Z4, Canada (tel: +1 604 875 2000 x3646; e-mail: walter.karlen@ieee.org). S. Raman and J.M. Ansermino are with the Department of Anesthesiol- ogy, Pharmacology & Therapeutics, University of British Columbia (UBC), Vancouver BC, V6T 1Z4, Canada. An easy-to-use, inexpensive diagnostic device that can swiftly and accurately identify children with severe pneu- monia would enable the timely administration of appropriate treatments. Prompt and effective application of life-saving antibiotics or oxygen therapy would, in turn, minimize wastage of scarce and costly resources. Pulse oximeters offer the possibility of achieving an accurate diagnosis of pneumonia. Pulse oximeters apply the Beer-Lambert law to estimate hemoglobin oxygenation (SpO 2 ). The law describes that light intensity diminishes exponentially when traveling in an ab- sorbing medium and the absorption is dependent on the wave- length. Oxygenated hemoglobin preferentially absorbs infrared light and transmits red light and deoxygenated hemoglobin behaves in the opposite manner. In a pulse oximeter, two light emitting diodes (LEDs) emitting red (660 nm) and infra-red (940 nm) light are used to actively illuminate the patient’s tissue (usually at the finger tip) alternately. The intensity of the non-absorbed light from each LED is measured with a receiver photodiode. The ratio of the transmitted infra-red and red intensity is empirically related to the SpO 2 . In addition to blood composition, the light absorption and transmission depend on the traveled light path, optical density of the tissue and volume of blood present in the tissue [4]. This permits the display of the variation of blood volume in the finger over time with a photoplethysmogram (PPG), which has a pulsatile component and a constant component. From the pulsatile component of the PPG, heart rate (HR) can be easily deduced. In addition to SpO 2 and HR, respiratory rate (RR) is another important parameter that can potentially be estimated from the PPG [5]. All three parameters are strong predictors of critical illness in pneumonia that can differentiate it from mild respiratory tract infections [6]. However, commercial pulse oximeters are currently limited to measuring HR and SpO 2 and require bulky, expensive equipment to process and monitor the data, which restricts their widespread adoption in the developing world. The aim of our research is to develop a fully featured, low cost pulse oximeter that operates on mobile phones and can be used as a multi-functional screening tool to improve diagnosis and treatment of children with severe pneumonia. We have previously demonstrated the Phone Oximeter [7], a commer- cial and Federal Drug Administration (FDA) approved pulse oximeter connecting to a smart phone. The Phone Oximeter was specifically designed for continuous monitoring and not spot-check applications, and only provided experimental RR measurements. In this work we present an efficient processing algorithm for robust RR estimation with the aim of developing a com-