Research On Health Monitoring Methods Based On Skin Characteristics

The sub-health state is a two-way intermediate dynamic between the health state and the disease state,and its effective diagnosis is of great help to the prevention of diseases.Traditional Chinese medicine believes that the viscera and qi and blood are important substances for maintaining the human body's daily life,and the face is the place where the meridians of the human body converge.The viscera has a corresponding distribution area on the face.The corresponding facial distribution area is diagnosed through the face inspection of Chinese medicine.,Can make a preliminary diagnosis of the health of the viscera.By estimating the heart rate from the facial video,the heart can be monitored more rigorously again.This article will quantitatively study the basic knowledge of facial inspection to judge visceral health and heart rate estimation to reach a preliminary judgment of human health.First,the facial video is collected and divided into frames,and the down-sampling of the image obtained by the video divided into frames is used as the input image of the next stage,and the facial area is accurately located by the MTCNN-ERT facial key point alignment algorithm.According to the basic elements of facial inspection,this paper divides the facial area into four parts that are easy to do image processing,that is,the forehead corresponds to the liver area,the root of the nose corresponds to the heart area,the middle and the lips correspond to the bladder area,the vertical line of the outer corner of the eye and the horizontal line of the middle of the ear.The area of the intercourse down to the chin corresponds to the kidney area.Among them,Mask-RCNN is used to identify areas such as the forehead,the middle of the person and around the lips,the vertical line of the outer corner of the eye and the horizontal line of the middle of the ear,down to the chin and other areas to diagnose the health of the viscera in the corresponding area,and Gabor must be used for the root area of the nose.The filter extracts the image texture feature,and then performs Otsu threshold segmentation on it,and judges whether there are horizontal stripes here by the preset value,and determines the heart health.In order to further determine the health of the heart,combined with modern medical heart rate(HR)knowledge,the use of non-contact photoplethysmography(r PPG)to extract the heart rate signal from the facial video,and then estimate the heart rate.First of all,by using the MTCNN-ERT face key point alignment algorithm to accurately locate the region of interest in the picture from the collected facial video frame,the input signal is obtained from the determined facial region of interest: the green channel signal,using the collective experience model The method of state decomposition decomposes the input signal into several eigenmode components,and then filters and saves the mode with the largest frequency spectrum in the heart rate frequency range [1,3] Hz.At the same time,the Kalman filter is used to reduce noise,and finally it will meet the requirements.All modal reconstructions are recorded as heart rate signals.After that,the fast Fourier transform or the peak reference point can be used to estimate the heart rate from the heart rate signal,and the heart health can be diagnosed again.