Novel Algorithms to Monitor Continuous Cardiac Activity with a Video Camera

Recent advances in computer vision methods have made physiological signal extraction from imaging sensors feasible. There is a demand to translate current post-hoc methods into real-time physiological monitoring techniques. Algorithms that function on a single frame of data meet the requirements for continuous, real-time measurement. If these algorithms are computationally efficient they may serve as the basis for an embedded system design that can be integrated within the vision hardware, turning the camera into a physiological monitor. Compelling results are presented derived from an appropriate algorithm for extracting cardiac pulse from sequential, single frames of a color video camera. Results are discussed with respect to physiologically relevant features of variability in beat-to-beat heart rate.