Extraction And Analysis Of The Heart Rate Of The Video Signal

Video-based heart rate measurement with a low-cost, non-contact, no electrodes, comfortable advantage can play a big role in remote medical monitoring, military, and other fields. The heart rate continuous measurement in the cases of neonatal intensive care, long-term epilepsy detection, burns or testing, sleep studies and trauma patients and others can avoid contact with the cable testers and other messy things, reducing the discomfort of the testers.The heart rate is extracted from the human face video based on the video-based heart rate measurement’s current research and related learning algorithms. The main work is as follows:(1) Human face is captured by a video camera. The face region is located in the image with human face which captured by a video camera and divided into three regions. Principal component analysis is used to extract the principal signal of three signals which is the average of three human face region’s pixel value of green channel. The peaks of the principal signal are detected after detrending and filtering the signal. Overlapping histogram is proposed to statistic the intervals between the peaks and calculate the average of the result which can be used to calculate the heart rate.(2) Due to the influence of ambient light and other factors, the measurement results will be subject to different degrees of influence, such as the detection of peaks, there may no peak can be detected or more peaks is detected, this paper uses the peak fitting and other methods to make more accurate peak detection, and estimate the peak location better. While lengthening the length of the video to increasing the measurement data can also improve the method.(3) Real-time measurement software is complied in the Visual Studio 2010 platform, using the C ++ language and Open CV.