Heart Rate Estimation from Facial Videos by Adaptive Region Selection Utilizing Machine Learning

Details

19:30 - 20:30 | Tue 6 Mar | Caribbean ABC | TuPO.12

Session: Poster Session # 2 and BSN Innovative Health Technology Demonstrations

Abstract

In the HR estimation using cameras, choice of the image regions, at which the heart rate (HR) is calculated, is critically important as it greatly affects the estimation accuracy. In this work, a novel algorithm for HR estimation that uses adaptive region selection is proposed. The image regions that clearly contain pulse waveforms are quickly found by a region selector using a machine learning technique. The experimental results show that the proposed method achieves the absolute average error less than 1.1BPM with the processing time less than 0.6 s for a single HR estimation.