Novel method for extracting spatio-temporal structure of blood volume pulse for accurate heart rate estimation (IMAGE)
Caption
Researchers from TUS propose a method that involves constructing a chrominance matrix using time-series signals from facial patches in an RGB facial video, followed by spatio-temporal structure analysis of blood volume pulse (BVP) using physics-informed DMD in a time-delay coordinate system. Further, heart rate estimation in the time domain is done using beat-to-beat peak period analysis of the extracted BVP signal.
Credit
Dr. Yoshihiro Maeda and Kosuke Kurihara from Tokyo University of Science, Japan
Usage Restrictions
Cannot be used without permission
License
Original content