A Facial Video based Framework to Estimate Physiological Parameters using Remote Photoplethysmography

Cardiovascular disease has become one of the greatest threats to the human body. Measurement of cardiovascular parameters in order to identify these diseases is very important. ECG, EEG, EOG etc. sensor based contact measurements are widely used. Recently, non-contact measurements have been widely studied. In this research, a framework to extract some physiological parameters such as heart rate, blood oxygen saturation and blood pressure using remote photoplethysmography (rPPG) has been proposed. From the facial video, after ROI selection and detrending, color channels (RGB) have been extracted and then the value of SpO2 was calculated. Filtering was performed on the raw signals. After that, FFT analysis was used to calculate the heart rate using 3 separate methods named as ICA, CHROM and using single Green channel. Meanwhile, SPA detrending was performed on the raw signal and systolic and diastolic blood pressure was calculated using the BMI value of the subject. Percentage of error for ICA, CHROM and green channel are 7.56, 9.70 and 15.87 respectively. For blood pressure the error percentage is 2.42 and 5.53 for systolic and diastolic BP respectively. For blood oxygen saturation, the average error is1.4%.