Cardiointervalogram Determination from Video Sequence of Human Face

The process of obtaining human pulse signals in the remote photoplethysmography system has been developed. The procedure is based on the use of a video camera that remotely and contactlessly records video of a human body area, in particular, its face, analysis of the dynamics of signals reflecting red, green, and blue light from the skin, and subsequent synthesis of these signals to obtain photoplethysmographic pulse signals. The main attention is paid to determining cardiointervals of the photoplethysmogram to implement variational pulsometry systems. The stages of implementing the remote photoplethysmography method are considered. An algorithm for processing color traces from a face video has been developed, based on forming a “mask” of a pulse and constructing a cross-correlation function. The results of calculations and experiments on restoring photoplethysmographic signals are presented. The methods were tested on our videos, where three areas of the human body were selected as regions of interest: forehead, cheek, and neck. The results indicate the possibility of determining cardiointervalograms using the correlation method of analyzing processed color traces.