Design And Research Of Non-contact Heart Rate Detection System Based On Facial Video

Heart rate is a key physiological parameter of the human body.The detection of heart rate is of great significance to the prevention,diagnosis and treatment of diseases and health monitoring.The current mainstream contact heart rate monitoring equipment needs to be in direct contact with the skin,which limits its usage scenarios and easily runs the risk of disease transmission.In view of these problems,this paper aims to propose a non-contact heart rate detection algorithm and provide a mobile terminal-based non-contact heart rate detection scheme.Based on r PPG technology,this paper analyzes the research status of non-contact heart rate detection algorithms,summarizes the problems to be solved,determines the basic scheme of heart rate detection,and designs the overall process of the heart rate detection system.Algorithm(RVG),which generates expression change video by building a 3D facial model,simulates real facial video data by adjusting facial pixel values,and obtains synthetic data,which can be used as a supplement and extension of real r PPG data to train heart rate detection algorithms to improve its detection performance;propose an encoder-decoder(ED-RPPG)-based heart rate signal amplification model,train the model with synthetic data,and improve the signal-to-noise ratio of the heart rate signal by amplifying the skin color change in the facial video,thereby achieving the accuracy of the human heart rate estimate.The experimental results show that the average absolute heart rate error of the Deep Phys model trained with synthetic video data is 18.65%,lower than that of the real data,and the average absolute error of heart rate detected across datasets is reduced by 23.38%.The synthetic data improves the heart rate detection accuracy of the model;using the ED-RPPG model The average absolute percentage error of the heart rate tested can reach 5.48%,the Pearson correlation can reach 0.8,and the heart rate detection is effective.Employing the heart rate detection algorithm to develop a mobile heart rate detection application based on the Android system,the application can display the heart rate detection results and achieve the self-monitoring function of the user’s heart rate.The experimental test of the heart rate detection application shows that the heart rate detection system has the advantages of simple operation interface,high heart rate detection accuracy and low development cost,which can meet the heart rate detection needs in most environments and provide people with a convenient heart rate detection solution.It reduces the burden of the subjects and provides a good technical foundation for the development of remote smart medical care in the future.