| As the most important vital sign of the human body,the heart rate has attracted more and more attention with the improvement of the quality of life.Traditional contact heart rate detection devices,such as the common electrocardiogram and photoplethysmography sensor,are widely used in hospitals,nursing homes and other public health places.However,these contact devices need to be attached to the subject’s body,which hinders the normal action and brings discomfort.At the same time,there are problems such as low efficiency and expensive cost,and they cannot play the role of initially screening a large number of patients in these public places.Since the outbreak of COVID-19,it has become an urgent need to realize a contactless,multi-person,efficient and reliable heart rate detection.Therefore,this paper implements a video-based contactless multi-person heart rate detection method.The main research work is as follows:(1)Research on key feature points of human face based on PFLD algorithm.Because the algorithm is easy to transplant and the operation rate on CPU is high,the PFLD algorithm is used to train the feature key point model of positive face and side face on different data sets,and puts forward the positive face judgment algorithm to realize the key point detection of multiple faces in a large motion state.(2)The ROI selection framework based on PFLD algorithm was proposed.Firstly,the ROI selection area of positive face and lateral face was proposed respectively,and the ROI selection area was improved.Finally,the experimental equipment and environment were analyzed,and the results of the experimental data were analyzed.(3)A multi-person heart rate detection system based on video analysis was designed.First,the design framework of the system is proposed,and then the implementation process of the three functional modules of the system is introduced in detail,and the realtime heart rate measurement and visualization function is realized with the camera. |
PDF Full Text Request