Smartphone-based respiratory rate estimation using photoplethysmographic imaging and discrete wavelet transform

Abstract

Respiratory rate is a key vital sign that needs daily monitoring for hospital patients in general and those with respiratory conditions in particular. Moreover, it is a predictor of major heart conditions. Yet, studies have shown that it is widely neglected in hospital care due partially to the discomfort caused by the required equipment. In this paper, we propose a smartphone-based method for accurate measurement of the respiratory rate using the video of the skin surface as recorded by the smartphone built-in camera in the presence of the flash light. From this input, we use frame averaging to extract a photoplethysmographic signal of the red, green, and blue channels. Next, we apply discrete wavelet transform on the best representative photoplethysmographic signal for respiratory signal extraction and estimate the rate. Fifteen subjects participated in the testing and evaluation. The maximum absolute error was 0.67 breaths/min, whereas the root mean square error was 0.366 breaths/min. The average percentage error and average percentage accuracy using our approach were 2.2%, 97.8% respectively. A comparison with other works in the literature reveal a superior performance in terms of accuracy, ease of use, and cost.