Facial Video Heart Rate Detection Based on FaceBoxes and ResNet34
The non-contact heart rate detection based on facial videos has problems such as motion artifacts and noise interference.To overcome the impact of motion artifacts on heart rate detection,a facial video heart rate detection method based on FaceBoxes and improved ResNet34 is proposed in this paper.Performing face detection and feature point detection on facial video frames can accurately locate the ROI region of each frame,overcome the influence of small movements,extract RGB three channel signals within the ROI region,perform spatial averaging preprocessing and signal denoising,obtain pulse wave signals,and calculate heart rate.The experimental results show that the feature point detection based on improved ResNet34 performs well in facial video heart rate detection,overcomes the influence of motion artifacts to a certain extent,and improves the inference speed of the original facial video heart rate detection method.
facial video heart rate detectionfeature detectionFaceBoxesResNet34
国家科技期刊平台
NSTL
万方数据
