A Non-invasive Heart Rate Estimation Approach from Photoplethysmography

Abstract

Heart rate is a major internal psychological parameter that defines the function of human body along with its response. In conventional heart rate measurement, electrodes and chest straps are attached to the patient’s body, but this system is a bit uncomfortable. So, in this paper a non-invasive heart rate measurement technique is proposed from Photoplethysmography (PPG signal). PPG signal will be extracted from the facial video in realistic frame that involves three approaches, firstly when the face is in at standstill, secondly when human face is motion and finally at real time scenario. Region of interest is selected according to their skin color variation in response with the heart rate. This color variation is implemented on the RGB color space and PPG signal is obtained. Independent component analysis (ICA) and Fast Fourier Transform are applied on this signal and heart rate is estimated from FFT. This method is also compared with the conventional heart rate measurement technique with ECG signal. Among the three conditions, the results are more accurate for motionless face video. For real time heart rate measurement, a Graphical User Interface (GUI) is designed in this work which will play a significant role at instant critical condition in remote areas.