Implementation and Validation of a Non-Contact Measurement of Cardiac Activity

Abstract

This work aim to present a simple and reliable method of measuring the pulse rate directly from face images obtained using a webcam. Video sequences from a webcam are captured and are converted into frames. The frames are decomposed into their constituent colour channels. The red component signals from the RGB images of the face and the forehead region (region of interest) are extracted. The heart rate is obtained by further processing the red channel signal and finding the maximum of power spectral density. To validate whether the resultant pulse rate are related to the heart rate, the results thus obtained from the video sequences are compared with pulse rate measured from an Electrocardiogram (ECG) signal acquired for the same subject. The red channel component of the face image contains most information pertaining to the pulse rate. The maximum of power spectral density of the red channel component gives the heart rate. The ECG circuit implemented to validate the results obtained through this method shows that the deviation is about ± 4 bpm. Thus as a surrogate to the conventional ECG measurement technique the method implemented in this work could be used for continuous monitoring of patients especially the elderly persons at home.