Non-contact Blood Pressure Measurement Research Based On Video Magnification

Blood pressure(BP)is not only an important human physiological parameter but also an important measure that may be used to diagnose whether individuals have cardiovascular diseases.As hypertension becomes increasingly widely recognized as a risk factor for cardiovascular,cerebrovascular,kidney,and other diseases,as well as the world’s top cause of death,more people are focusing on the diagnosis,prevention,and treatment of hypertension and its effects.It is vital to explore a practical method of continuously monitoring human BP.Traditional contact BP measuring methods and equipment cannot overcome the limitations of human interaction.Cuff inflation and deflation may cause discomfort to the wearer and will not allow for continuous BP monitoring.Exploring noninvasive,non-contact,and continuous BP detection is therefore critical for early hypertension diagnosis,treatment,and prevention.Based on video motion amplification technology,this thesis presents a new non-contact BP measurement method.Using a camera to collect video of the human face and neck,and implementing video motion amplification to amplify the weak pulse wave signals of the common carotid artery and facial artery,calculating the pulse wave transit time(PTT)of the two-way pulse wave,and estimating the heart rate(HR)by the pulse wave.Noncontact continuous BP measurement without feeling the human body is realized by combining PTT,HR,and the suggested multi-mode BP measurement model.This thesis’ s major research project is as follows:1)This thesis examines the research current state of pulse wave correlation detection and BP measurement,as well as the related underlying theory of the motion amplification-based pulse wave extraction method.The motion magnification techniques suited for the human common carotid artery and facial artery are determined by evaluating related motion amplification techniques and their advantages and disadvantages.2)This thesis carries out experimental research,build an experimental system,capture the subject’s face and neck video with the Casio EX-ZR370 high frame rate mode,and utilize the LED array to create a stable light source to eliminate interference from ambient light variations.Through experimental analysis,the effect of video color magnification on neck region color magnification is poor.This thesis no longer pays attention to blood volume pulse and proposes to extract the pressure of pulse waves based on the motion amplification of video magnification.Combined with independent component analysis(ICA)based on blind source separation(BSS),principal component analysis(PCA),and wavelet transform,the noise is filtered to extract the fine two-way pulse wave signal.3)Using fast Fourier transform of the pulse wave to estimate HR,and calculating the dual pulse wave PTT of the common carotid artery and face artery.On this basis,this thesis proposes three multi-modulus BP measurement models combined with PTT and HR,in addition to the relevant BP measurement models proposed by predecessors.The real HR and BP of the human body are measured by a sphygmomanometer in synchrony,and the model parameters are fitted and evaluated to determine the model’s fitting effect.Finally,it is determined that the above three multi-modulus BP measurement models have high correlation and consistency with the BP measured by the cuff compression electronic sphygmomanometer using correlation analysis,error analysis,and Bland Altman consistency analysis with the measured value of the sphygmomanometer.The results of this study suggest that the method proposed in this work can effectively replace the traditional cuff compression electronic sphygmomanometer in the measurement of human BP.