Using video stream for continuous monitoring of breathing rate for general setting

Abstract

Measuring a breathing rate has multiple clinical applications from early detection of diseases to monitoring of patients in a critical condition. Several techniques have been introduced to measure the breathing rate. The conventional breathing rate monitoring systems are contact-based and expensive, and sometimes their presence will interfere with the normal breathing system and thus produce an incorrect reading. Furthermore, they are not used in case a patient skin is sensitive (or burned skin). This paper presents a contact-free breathing rate measurement using video analysis. To accomplish this, we introduced two algorithms. The first one is approximating chest regions from face detection algorithm, and the second one is using peak-to-peak measurement to calculate a breathing rate of a person on the incoming signal. We tested the proposed algorithm on 10 subjects, and the result shows an overall 11.61% increase in accuracy (with 84.66% average accuracy) compared with FFT-based implementation. Furthermore, the time complexity of the proposed breathing rate measurement method is O(logn), whereas FFT implementation is O(nlogn). Thus, the proposed system is promising that it can be used as a standalone system or as a component of another application.