Facial Video-Based Respiratory Rate Recognition Interpolating Pulsatile PPG Rise And Fall Times.

The respiratory rate is one of the most important vital parameters and is essential for initial medical assessment of a patient as well as for identification of a wide range of diseases. Besides, camera-based monitoring systems offer great potential for future health surveillance. In this work, we show that pulsatile PPG rise and fall times contain frequencies which are associated with respiration. By interpolating the differences between the rise and fall times of a pulsatile PPG signal derived from facial video images highly accurate estimates of the person’s respiratory rate can be determined. Combining this with another method, a new state-of-the-art algorithm for facial video-based respiratory rate recognition was developed. Thus, the respiratory rate is correctly recognized for over 90 % of windows on our database.