Research On Integrated Analysis Of Human Body Multi-physiological Parameters And Auxiliary Diagnosis Technology

Various physiological parameters of the human body are important indicators of the health of life.By monitoring and analyzing physiological parameters,doctors can extract important physiological information of patients and take effective medical measures for the disease.In my country,the development of cardiovascular disease is not optimistic,and it has become the first major disease that endangers people’s health.The monitoring of various physiological parameters of the human body is of great significance to the prevention and treatment of cardiovascular diseases,especially in the context of the current raging new crown pneumonia epidemic,the efficient monitoring of the physiological parameters of patients in isolation scenarios has become a top priority for doctors to carry out treatment.Current monitoring methods for some physiological parameters such as heart rate,respiration,etc.often have shortcomings such as inconvenience and invasiveness,and are not suitable for long-term measurement of patients.In this paper,starting from the photoplethysmography,several physiological parameters are correlated and integrated,and an efficient monitoring method of respiratory rate and heart rate is proposed based on the photoplethysmography signal.In this paper,an arrhythmia classification algorithm based on ECG signal is proposed as a new method to assist doctors in diagnosis.The respiratory signal has a modulating effect on the pulse wave signal.Using the time domain characteristics of the photoplethysmography pulse wave,the respiratory signal is extracted from several waveform characteristics and the respiratory frequency is calculated.In addition,the filtering method is used to obtain the respiratory frequency from the pulse wave signal.The respiratory rate obtained by the fusion of the two methods has the smallest error with the true value,and the average absolute error is 2.13BPM.In terms of heart rate monitoring,the photoplethysmography signal containing motion artifacts is processed,and the pulse wave signal is transformed into frequency domain analysis.Various influencing factors are considered in the selection of spectral peak points.The heart rate detection algorithm proposed in this paper is used to obtain The heart rate value is highly similar to the real heart rate,the Pearson similarity coefficient is 0.9907,and the mean absolute error with the real value is 1.32BPM.In the classification of arrhythmias,this paper uses the ECG signal data of the Massachusetts Institute of Technology and the Boston Beth Israel Hospital Arrhythmia Database(MIT-BIH)to carry out experiments,according to the American Association for the Advancement of Medical Devices.into 5 categories.For the beat data in the MIT-BIH database,the wavelet transform is used to filter out the noise,and the generative adversarial network is used to improve the imbalance of the number of heart beats in the data set.The convolutional neural network CNN and CNNN+LSTM models with residual designed and implemented in this paper are trained on the training set and tested on the test set.The experimental results show that the model proposed in this paper has an excellent classification effect.The overall accuracy of the CNN model classification is 99.27%,the Macro-F1 score is 98.02%,the overall accuracy of the CNN+LSTM model is 99.34%,and the Macro-F1 score is 98.25%.