Based On Non-Contact Heart Rate Variability Parameters Extraction

With the rapid development of society,people are facing increasing pressure from work and life,and being in a state of stress for a long time is harmful to physical and mental health.Therefore,it is of great significance to identify stress for busy people.Use the Heart Rate Variability(HRV)parameter to identify the state of human stress.The traditional measurement of heart rate variability parameters generally uses a contact method,and the contact acquisition method generally requires the tester to remain stationary during the measurement process,which hinders people’s normal work.Therefore,this article focuses on non-contact measurement methods,the purpose of which is to measure physiological parameters without hindering people’s normal working conditions,so as to perform emotion recognition and achieve the goal of daily pressure monitoring.Based on Imaging Photoplethysmography(IPPG),this paper proposes a measurement method that uses a common camera to capture face video.Because the physiological signal of the human body is weak,the signal-to-noise ratio of the physiological signal extracted from the face video is relatively low.Therefore,this paper uses a method based on Euler amplification to enhance the change of human skin color to enhance the human face information related to physiological signals in the image.This article collected 25 videos of 5 testers for experiments.The experimental results show that after using Euler amplification the average absolute error of the heart rate extracted from the face video and the heart rate extracted by the contact method has been reduced from 19 times / minute before using Euler to 1 / minute,and the maximum absolute error has been reduced from 40 times / minute to 2 times / minute.It is proved that Euler amplification can effectively enhance the heart rate related information in face videos.Due to the non-contact measurement method is susceptible to noise interference caused by ambient light,acquisition equipment,etc.,this paper uses wavelet transform to denoise.The best wavelet basis function is selected through analysis and experiments,and a good denoising effect is obtained.In order to solve the problem of extracting characteristics of heart rate variability when covering individual areas of the face,this paper divides the human face into four measurement areas: forehead,cheeks,below the mouth to the mouth,and the entire face,and extracts heart rate and HRV parameters for comparative analysis.The HRV time domain parameters as an example.The average absolute error of the RR interval mean of the HRV time domain parameters extracted from the above four regions is reduced from 46,40,57,56 ms to 15,16,10,and 13 ms,respectively.Experiments show that after Euler amplification is used,more accurate physiological parameters can be extracted from different measurement areas of the face.In order to verify the effect of heart rate variability features extracted by noncontact measurement method for emotion recognition,this article takes stress recognition as an example,uses Stroop Color-World Test as a stress-inducing scheme,and uses the CART decision tree algorithm to perform stress and non-stress emotion Identify.The experimental results show that using the non-contact method proposed in this paper to obtain heart rate variability(HRV parameters)for stress emotion recognition,compared with the HRV parameters extracted by contact method for pressure recognition,the average accuracy rate is 2.65% lower,which proves that based on non-contact It is feasible to extract the HRV parameters for stress emotion recognition using the new method.