Method Research And System Design Of Real-time Heart Rate Measurement Based On Face Video

Heart rate is an extremely important physiological parameter,which is widely used in emotion recognition and the diagnosis of heart arrhythmia,myocardial ischemia,coronary heart disease,hypertension,chronic heart failure and other cardiovascular diseases.At present,professional heart rate detection devices in the market must contact the skin to measure the heart rate parameters of the subjects,which causes certain discomfort to the subjects.In addition,it is not suitable for patients with skin burn and damage,newborn infants with delicate skin,or covert criminal lie detection.In view of the limitations of the application scenarios of traditional contact heart rate detection equipment,a non-contact real-time heart rate measurement system is designed in this paper,which can extract the real-time heart rate parameters of the subjects by collecting the face video from an ordinary camera as the input and combining with the corresponding video image processing algorithm and signal analysis algorithm.This method can continuously monitor the real-time heart rate parameters of the subject in a comfortable non-contact environment,and reduce the discomfort and anxiety of the subjects,which is of great significance for the development of telemedicine,especially for the prevention and diagnosis of cardiovascular diseases.It is also very suitable for newborn infants and other people who are not suitable to use traditional contact equipment to detect physiological parameters.The main work and research results of this paper are as follows:(1)Based on the principle of Imaging Photoplethysmography(IPPG)signal formation,a reasonable and effective IPPG signal extraction algorithm is designed.Firstly,the position of the face is accurately recognized and tracked,and the position of the skin pixels is extracted from it,so as to filter out the pixels that may be interfered by noise,such as eyes,hair and background.Then,according to the optical characteristics of oxygen and deoxyhemoglobin in red blood cells that are more sensitive to the 510-590nm spectral segment,the Cg color path which is close to the above spectrum is selected as the signal source,and the high-quality IPPG signal is obtained by extracting the face skin position's pixels in the Cg color path of each frame image.The effectiveness of extracting IPPG signal from Cg color pathway was proved by contrast experiment(2)CMOR5-3(Complex Morlet 5-3)wavelet is used to generate the energy spectrum of the IPPG signal,and the time-domain and frequency-domain components of the IPPG signal are analyzed.According to the physiological characteristics of heart rate parameters changes are continuous within a certain range,the pseudo noise is removed and the measurement accuracy is improved.By further analyzing the energy distribution characteristics of IPPG signals under strong noise interference,the energy proportional threshold detection method is set to effectively reduce the impact of noise interference on the measurement results.Compared with other non-contact heart rate measurement methods and typical biological signal analysis wavelets,it illustrates the advantages of CMOR5-3 wavelet analysis in heart rate parameter extraction.(3)The non-contact real-time heart rate measurement is realized by optimizing the algorithm,reducing the operation time and sampling frame rate,increasing the sampling interval,and designing a reasonable and efficient heart rate parameter extraction process,and so on.Based on this,the MATLAB and Python versions of the real-time heart rate parameter extraction system were developed to extract the real-time heart rate parameters of the subjects in a non-contact way.(4)In order to verify the validity of the non-contact real-time heart rate measurement system designed in this paper,the non-contact real-time heart rate measurement system and the finger clip pulse oximeter were used to measure the heart rate parameters of multiple subjects in the resting state,head movement state and horse stance state respectively,the mean absolute error(|M _e|),the standard deviation of error(SD),and the root mean square error(RMSE)are used as evaluation indicators to evaluate the accuracy of the measurement results of the non-contact system,and the consistency of the measurement results of the two methods was further evaluated by Bland-Altman analysis.The experimental results show that the extraction results of the non-contact real-time heart rate measurement system designed in this paper have good consistency with the extraction results of medical equipment.