Development Of Pulmonary Function Evaluation Model And Intelligent Rehabilitation Training Assessment Algorithm For COPD Patients

Chronic obstructive pulmonary disease(COPD)is a respiratory disease,known as "COPD",which can be prevented and treated by drugs and rehabilitation.Pulmonary rehabilitation is the most important part of the management of COPD patients in stable stage.It can enhance their exercise endurance,reduce medical expenses and improve the quality of the life.An important way of lung rehabilitation is exercise rehabilitation.Traditional rehabilitation training is usually carried out in hospitals,but the high cost makes it difficult for patients to adhere to it for a long time.The implementation of community and family oriented pulmonary rehabilitation intervention system can greatly improve the rehabilitation efficiency and save the cost.However,the current problem is to formulate the best way,intensity,time and frequency of exercise training according to the patient’s own health status,which needs to consume a lot of medical resources;In addition,when patients exercise by themselves in the community or family,they can’t control their amount and quality of the exercise,and also there is a lack of evaluation of the effectiveness of movement.Baduanjin is a method of rehabilitation training for COPD patients.Based on the needs of low-cost health management and personalized intervention tracking of Baduanjin rehabilitation training,this paper mainly puts forward the following main contents:(1)Study on pulmonary function evaluation model of COPD patients based on physiological parametersBased on the study of the relationship between demographic parameters,peripheral blood parameters CRP,WBC and neu%and pulmonary function indexes FEV1 and FVC,a prediction model of pulmonary function indexes FEV1 and FVC based on machine learning algorithm is proposed and verified to predict the pulmonary function of COPD patients.The results show that the evaluation model can predict the gold classification of COPD patients with 90.24%accuracy and evaluate lung function.It can solve the problem of insufficient diagnosis of COPD caused by the lack of lung function tester in medical and health care institutions or the lack of lung function tester in primary health care,and provide a basis for generating a personalized Baduanjin rehabilitation training program.(2)Research on action recognition and evaluation method of intelligent rehabilitation trainingTaking into account the scientificity and motivation of rehabilitation training,it provides a theoretical basis for the customization and difficulty classification of patients’ training tasks.The system provides three difficulty levels to adapt to patients with chronic obstructive pulmonary disease with different degrees of disease and pulmonary function.At the same time,in the absence of professional physicists to guide and correct the patient’s actions,in order to standardize the patient’s rehabilitation training actions,the training evaluation score is given according to the similarity between the training actions and the standard actions at the end of each Baduanjin rehabilitation training.In order to realize the action recognition and evaluation based on ordinary RGB camera on ordinary household computers(such as NVIDIA GTX 1650 platform),the algorithm used must be as fast and efficient as possible.This study uses the structure of mobilenetv3 to replace the vggnet part of the original openpose and improves the openpose algorithm.At the same time,referring to the structure of lightweight openpose,the shared convolution layer and residual expansion structure are used to improve the network structure in the initial and strengthening stages of openpose.On the ordinary home computer platform,the average frame rate of Baduanjin action recognition can reach about 30fps,and the average classification accuracy of action can reach 97.3%,which can meet the real-time requirements of the algorithm.In the process of motion evaluation,there is an attention mechanism in human perception of motion quality evaluation,and subjective preference may significantly affect the judgment results.In order to simulate the influence of visual attention mechanism on judging sports activities,an LSTM action evaluation model based on spatio-temporal attention mechanism is proposed.Through spatial and temporal attention,the differentiated joints and key frames with different importance are automatically determined and allocated.The results show that the average accuracy of LSTM action evaluation method based on additional spatio-temporal attention mechanism is 94%,while the average accuracy of traditional LSTM evaluation method is 87%,The LSTM movement evaluation model based on spatiotemporal attention mechanism improves the accuracy of Baduanjin movement quality evaluation,and provides guarantee for the application of Baduanjin rehabilitation training system for patients with chronic obstructive pulmonary disease.(3)Application evaluation and effect verification of intelligent rehabilitation training system for patients with chronic obstructive pulmonary diseaseIn order to verify the effectiveness of the proposed intelligent rehabilitation training system for patients with chronic obstructive pulmonary disease,firstly,the overall structure of Baduanjin rehabilitation training system for patients with chronic obstructive pulmonary disease is designed.The system service process includes physiological parameter monitoring,lung function index prediction,Baduanjin rehabilitation monitoring and evaluation,exercise intervention guidance and other steps.Through self-help non-invasive detection,obtain a variety of physiological parameters such as the user’s height and electronic questionnaire information for COPD evaluation,and then comprehensively evaluate the health status of COPD patients and pulmonary function indicators.Based on the key information such as physiological parameter monitoring and rehabilitation training result feedback,combined with the characteristics of Baduanjin training,a personalized Baduanjin rehabilitation training scheme is generated,and the implementation feedback and optimization of exercise scheme are provided to improve the pertinence of sports rehabilitation.Based on the overall structure of the design,a set of Baduanjin rehabilitation training system is realized,and the application deployment,analysis and evaluation are carried out.Relying on Suzhou science and technology city hospital,we are responsible for clinical application research,carry out clinical application in Dongzhu and Hengtang communities in Suzhou,and build a lung rehabilitation intervention system covering communities,families and hospitals.COPD patients were recruited to form the experimental group,and the appropriate COPD patients were selected as the control group to participate in the application experiment for 12 weeks;The patients were evaluated before and 12 weeks after the intervention.The main evaluation indexes were 6-minute walking distance(6WMD)and the score of St.George’s Hospital Respiratory Questionnaire(SGRQ).The experimental results show that the Baduanjin rehabilitation training system for COPD patients has a good intervention effect.After 12 weeks of training,the use of the intelligent rehabilitation system can increase the 6wmd of patients by 21.64%and reduce the total SGRQ score of patients by 17.92%,significantly improve the aerobic capacity and endurance of COPD patients,improve the exercise function of patients,and effectively reduce the health damage of COPD patients.