Respiratory Rate Estimation from Face Videos

TL;DR

This paper presents a contact-free method for estimating respiratory rate from face videos using remote photoplethysmography, employing motion compensation and filtering techniques to improve accuracy and simultaneously measure other vital signs.

Contribution

It introduces a novel framework that accurately estimates respiratory rate from face videos, overcoming limitations of prior methods based on chest motion.

Findings

Accurate remote respiratory rate measurement demonstrated.

Framework also provides heart rate and heart rate variability.

Method is robust to voluntary movements.

Abstract

Vital signs, such as heart rate (HR), heart rate variability (HRV), respiratory rate (RR), are important indicators for a person's health. Vital signs are traditionally measured with contact sensors, and may be inconvenient and cause discomfort during continuous monitoring. Commercial cameras are promising contact-free sensors, and remote photoplethysmography (rPPG) have been studied to remotely monitor heart rate from face videos. For remote RR measurement, most prior art was based on small periodical motions of chest regions caused by breathing cycles, which are vulnerable to subjects' voluntary movements. This paper explores remote RR measurement based on rPPG obtained from face videos. The paper employs motion compensation, two-phase temporal filtering, and signal pruning to capture signals with high quality. The experimental results demonstrate that the proposed framework can obtain accurate RR results and can provide HR, HRV and RR measurement synergistically in one framework.