Respiratory tract drug delivery analysis method based on deep analysis of lung function curve of chronic obstructive pulmonary disease

Abstract

A respiratory tract drug delivery analysis method based on deep analysis of a lung function curve of a chronic obstructive pulmonary disease comprises the following steps: deeply and accurately analyzing the lung function curve of a patient, preprocessing image information of the lung function curve, then carrying out digital processing, carrying out feature extraction on the preprocessed image by using a convolutional neural network, and obtaining a feature extraction result; analyzing the curve data to obtain multi-type digital characteristic parameters; weight coefficients are distributed for the multiple types of digital characteristic parameters according to the influence degree, the multiple types of digital characteristic parameters and database data are compared and analyzed through a neural network, then the respiratory function of the patient is accurately evaluated, and the corresponding drug curative effect is predicted; according to the respiratory function evaluation effect and the drug prediction effect of the patient, the optimal drug inhalation condition is provided for the patient, and the treatment compliance and the treatment effect of the patient are improved by setting a personalized guidance mode.

Description

Technical field:

[0001] The present invention relates to a respiratory tract drug administration analysis method based on in-depth analysis of COPD lung function curves. Background technology:

[0002] Chronic obstructive pulmonary disease (COPD) is a chronic, progressive respiratory disease usually caused by long-term smoking, air pollution, or occupational exposure to harmful dust. Inhalation therapy is currently the main treatment for COPD.

[0003] Since each patient's lung indicators are different, the treatment effect will also be affected by the individual differences of the patients. Traditional respiratory medication guidance relies on the doctor's clinical experience and cannot accurately meet the treatment needs of each patient. The drug utilization rate is low, resulting in uneven treatment effects and a lack of dynamic adjustment mechanism, which has a certain negative impact on the treatment effect. Summary of the invention:

[0004] The embodiment of the present invention provides a respiratory drug administration analysis method based on deep analysis of the lung function curve of COPD. The method is reasonably designed. By performing accurate and deep analysis of the lung function curve, the image information is converted into digital information. Different digital feature parameters are weighted according to their influence on the patient's respiratory function. The use of digital processing and weight distribution can accurately capture subtle changes in the patient's respiratory function, reduce the subjectivity of the evaluation, and make the evaluation results more accurate and objective; at the same time, it can provide the best usage guidance for the patient's drug inhalation, so that the patient can inhale the drug according to the correct steps, and ensure the effective use of the drug to the greatest extent. The personalized guidance method set for different patients improves the patient's treatment compliance and treatment effect, and solves the problems existing in the prior art.

[0005] The technical solution adopted by the present invention to solve the above technical problems is:

[0006] A respiratory drug administration analysis method based on in-depth analysis of COPD lung function curves, the analysis method comprising the following steps:

[0007] S1, deeply and accurately analyzes the patient's lung function curve, pre-processes the image information of the lung function curve and then digitizes it, uses a convolutional neural network to extract features from the pre-processed image, analyzes the curve data, and obtains multiple types of digital feature parameters;

[0008] S2, assign weight coefficients to multiple types of digital feature parameters according to their influence, use neural networks to compare and analyze multiple types of digital feature parameters with database data to achieve accurate evaluation of the patient's respiratory function and predict the efficacy of corresponding drugs;

[0009] S3, based on the patient's respiratory function assessment results and drug prediction effects, provides patients with the best drug inhalation conditions and sets up personalized guidance methods to improve patients' treatment compliance and treatment effects.

[0010] The pulmonary function curve is a flow-volume correlation curve, and the size, color and contrast of the curve need to be adjusted during the preprocessing process; the multi-type digital feature parameters include PEF, MMEF, FEF25%, FEF50%, FEF75%, PIF, MIF50, curve area and slope of the related position curve;

[0011] The PEF represents the maximum flow rate during forced exhalation, the MMEF represents the average flow rate when the forced exhalation vital capacity is 25%-75%; the FEF25%, FEF50%, and FEF75% are the expiratory flow rates when 25%, 50%, and 75% of the vital capacity are exhaled, respectively; the PIF represents the maximum inspiratory peak flow rate, and the MIF 50 represents the instantaneous inspiratory flow rate when 50% of the forced vital capacity is exhaled.

[0012] The curve areas are the area between the inhalation curve and the abscissa axis, and the area between the exhalation curve and the abscissa axis.

[0013] The database stores lung function curves of normal people and patients of all ages and simulation results of patients' lung function curves during drug inhalation in the respiratory tract.

[0014] The following steps are involved in deeply and accurately analyzing the patient's lung function curve, preprocessing the image information of the lung function curve and then digitizing it, extracting features from the preprocessed image using a convolutional neural network, analyzing the curve data, and obtaining multiple types of digital feature parameters:

[0015] S1.1, the convolution kernel of the convolutional neural network slides on the input image, performs multiplication and summation operations on the local area of ​​the image, and obtains a feature map, which contains different digital feature parameters of the image;

[0016] S1.2, use the pooling layer to sample the feature map to reduce the dimension and calculation amount of the feature map;

[0017] S1.3, transfer the digitized image features to the data analysis module for data screening, filter out data with large differences, incomplete and unnecessary data, use the random forest algorithm to evaluate the importance of features, and use the importance as a reference for weight allocation. Multiply the feature weight by the corresponding feature value to obtain the weighted feature set. Use the Euclidean distance and cosine similarity calculations to match the weighted data with the data in the database, and use MLP to compare and analyze the data in the database.

[0018] The MLP includes a model framework, an activation function, a loss function and an optimizer; the model framework includes an input layer, a hidden layer and an output layer, the input data is forward propagated through each layer of the MLP, the output of each layer is calculated and the prediction result is obtained in the output layer; the activation function is a Sigma activation function, which can perform nonlinear changes in neuron output;

[0019] The loss function uses mean square error to measure the difference between the model prediction result and the actual result; the optimizer is an Adam optimizer, which is used to minimize the loss function.

[0020] The optimal drug inhalation conditions include an optimal inhalation flow rate, a breath-holding time, and an optimal exhalation flow rate adapted to the patient when inhaling the drug.

[0021] The analysis system of the analysis method comprises a curve analysis module, a data analysis module, a drug administration guidance module and a feedback module;

[0022] The curve analysis module includes a curve recognition unit, a preprocessing unit and a feature extraction unit. The curve recognition unit is used to recognize the flow-volume curve of the patient's lung function curve. The preprocessing unit is used to adjust the size, color and contrast of the image and perform data enhancement to improve the generalization ability of the model. The feature extraction unit is used to convert image information into digital information and process the image into data that can be compared with the same.

[0023] The data analysis module includes a feature screening unit, a weight allocation unit, a database and a therapeutic effect prediction unit;

[0024] The medication guidance module includes an instruction generation unit, a text output unit and a real-time monitoring unit, which outputs the predicted optimal drug inhalation conditions for the patient in text form according to the situation predicted in the data analysis module; the feedback module is used to store the patient's actual inhalation drug process data and feed it back to the data analysis module, and the data analysis module adjusts the efficacy prediction and inhalation conditions according to the feedback data.

[0025] The feature screening unit screens the transmitted data and selects the required data to transmit to the weight distribution unit. According to the input data, characteristic values ​​are assigned to different factors to determine the importance of each factor in the drug deposition process. The feature similarity of the data with assigned weights is calculated, and a neural network is used to compare the data with a large amount of data already stored in the database to find the data that best matches the patient, judge the patient's lung function, check the drug particle deposition distribution score and deposition efficiency corresponding to the data, predict the patient's drug efficacy according to the data in the database, and provide the conditions required for the best drug efficacy.

[0026] The present invention adopts the above structure, and through the accurate analysis of the lung function curve, the image information is digitized and weighted according to the influence of different parameters on the patient's respiratory function. The use of digital processing and weight distribution can accurately capture the subtle changes in the patient's respiratory function, reduce the subjectivity of the evaluation, and make the evaluation results more accurate and objective; by setting up a feedback module to collect and accurately record the patient's drug inhalation data, and seamlessly integrate these data into the system, the system performs comprehensive analysis and evaluation, and realizes dynamic adjustment and optimization of the current treatment plan. This mechanism ensures the high personalization and accuracy of the treatment process, and effectively improves the treatment effect and patient experience; through the neural network, the automatic analysis and evaluation of the lung function curve is realized, which reduces the time and workload of doctors manually interpreting the report, greatly improves medical efficiency, and has the advantages of practicality, efficiency, safety and reliability. Description of the drawings:

[0027] Figure 1 It is a schematic diagram of the structure of the analysis system of the present invention.

[0028] Figure 2 It is a structural schematic diagram of the lung function curve of the present invention.

[0029] Figure 3 It is a flow chart of the analysis of the lung function curve of the present invention.

[0030] Figure 4 The present invention is a flowchart of the drug administration guidance evaluation. Specific implementation method:

[0031] In order to clearly illustrate the technical features of the present solution, the present invention is described in detail below through specific implementation methods and in conjunction with the accompanying drawings.

[0032] like Figure 1-4 As shown in, a respiratory drug administration analysis method based on deep analysis of COPD lung function curves, the analysis method comprises the following steps:

[0033] S1, deeply and accurately analyzes the patient's lung function curve, pre-processes the image information of the lung function curve and then digitizes it, uses a convolutional neural network to extract features from the pre-processed image, analyzes the curve data, and obtains multiple types of digital feature parameters;

[0034] S2, assign weight coefficients to multiple types of digital feature parameters according to their influence, use neural networks to compare and analyze multiple types of digital feature parameters with database data to achieve accurate evaluation of the patient's respiratory function and predict the efficacy of corresponding drugs;

[0035] S3, based on the patient's respiratory function assessment results and drug prediction effects, provides patients with the best drug inhalation conditions and sets up personalized guidance methods to improve patients' treatment compliance and treatment effects.

[0036] The pulmonary function curve is a flow-volume correlation curve, and the size, color and contrast of the curve need to be adjusted during the preprocessing process; the multi-type digital feature parameters include PEF, MMEF, FEF25%, FEF50%, FEF75%, PIF, MIF50, curve area and slope of the related position curve;

[0037] The PEF represents the maximum flow rate during forced exhalation, the MMEF represents the average flow rate when the forced exhalation vital capacity is 25%-75%; the FEF25%, FEF50%, and FEF75% are the expiratory flow rates when 25%, 50%, and 75% of the vital capacity are exhaled, respectively; the PIF represents the maximum inspiratory peak flow rate, and the MIF 50 represents the instantaneous inspiratory flow rate when 50% of the forced vital capacity is exhaled.

[0038] The curve areas are the area between the inhalation curve and the abscissa axis, and the area between the exhalation curve and the abscissa axis.

[0039] The database contains lung function curves of normal people and patients of all ages and simulation results of patients' lung function curves during the process of drug inhalation in the respiratory tract.

[0040] The following steps are involved in deeply and accurately analyzing the patient's lung function curve, preprocessing the image information of the lung function curve and then digitizing it, extracting features from the preprocessed image using a convolutional neural network, analyzing the curve data, and obtaining multiple types of digital feature parameters:

[0041] S1.1, the convolution kernel of the convolutional neural network slides on the input image, performs multiplication and summation operations on the local area of ​​the image, and obtains a feature map, which contains different digital feature parameters of the image;

[0042] S1.2, use the pooling layer to sample the feature map to reduce the dimension and calculation amount of the feature map;

[0043] S1.3, transfer the digitized image features to the data analysis module for data screening, filter out data with large differences, incomplete and unnecessary data, use the random forest algorithm to evaluate the importance of features, and use the importance as a reference for weight allocation. Multiply the feature weight by the corresponding feature value to obtain the weighted feature set. Use the Euclidean distance and cosine similarity calculations to match the weighted data with the data in the database, and use MLP to compare and analyze the data in the database.

[0044] The MLP includes a model framework, an activation function, a loss function and an optimizer; the model framework includes an input layer, a hidden layer and an output layer, the input data is forward propagated through each layer of the MLP, the output of each layer is calculated and the prediction result is obtained in the output layer; the activation function is a Sigma activation function, which can perform nonlinear changes in neuron output;

[0045] The loss function uses mean square error to measure the difference between the model prediction result and the actual result; the optimizer is an Adam optimizer, which is used to minimize the loss function.

[0046] The optimal drug inhalation conditions include an optimal inhalation flow rate, a breath-holding time, and an optimal exhalation flow rate adapted to the patient when inhaling the drug.

[0047] The analysis system of the analysis method comprises a curve analysis module, a data analysis module, a drug administration guidance module and a feedback module;

[0048] The curve analysis module includes a curve recognition unit, a preprocessing unit and a feature extraction unit. The curve recognition unit is used to recognize the flow-volume curve of the patient's lung function curve. The preprocessing unit is used to adjust the size, color and contrast of the image and perform data enhancement to improve the generalization ability of the model. The feature extraction unit is used to convert image information into digital information and process the image into data that can be compared with the same.

[0049] The data analysis module includes a feature screening unit, a weight allocation unit, a database and a therapeutic effect prediction unit;

[0050] The medication guidance module includes an instruction generation unit, a text output unit and a real-time monitoring unit, which outputs the predicted optimal drug inhalation conditions for the patient in text form according to the situation predicted in the data analysis module; the feedback module is used to store the patient's actual inhalation drug process data and feed it back to the data analysis module, and the data analysis module adjusts the efficacy prediction and inhalation conditions according to the feedback data.

[0051] The feature screening unit screens the transmitted data and selects the required data to transmit to the weight distribution unit. According to the input data, characteristic values ​​are assigned to different factors to determine the importance of each factor in the drug deposition process. The feature similarity of the data with assigned weights is calculated, and a neural network is used to compare the data with a large amount of data already stored in the database to find the data that best matches the patient, judge the patient's lung function, check the drug particle deposition distribution score and deposition efficiency corresponding to the data, predict the patient's drug efficacy according to the data in the database, and provide the conditions required for the best drug efficacy.

[0052] The working principle of the respiratory drug administration analysis method based on deep analysis of COPD lung function curve in the embodiment of the present invention is: through accurate and deep analysis of the lung function curve, the image information is converted into digital information, and weights are assigned to different digital feature parameters according to the degree of influence on the patient's respiratory function. The use of digital processing and weight assignment can accurately capture subtle changes in the patient's respiratory function, reduce the subjectivity of the evaluation, and make the evaluation results more accurate and objective; at the same time, it can provide the best usage guidance for the patient's drug inhalation, so that the patient can inhale the drug according to the correct steps, and ensure the effective use of the drug to the greatest extent. The personalized guidance method set for different patients improves the patient's treatment compliance and treatment effect.

[0053] In the overall solution, the analysis method mainly includes the following steps: deeply and accurately analyze the patient's lung function curve, digitize the image information of the lung function curve after preprocessing, use convolutional neural network to extract features of the preprocessed image, analyze the curve data, and obtain multiple types of digital feature parameters; assign weight coefficients to multiple types of digital feature parameters according to the degree of influence, use neural network to compare and analyze the multiple types of digital feature parameters with database data respectively to achieve accurate evaluation of the patient's respiratory function and predict the efficacy of corresponding drugs; provide patients with the best drug inhalation conditions based on the patient's respiratory function evaluation effect and drug prediction effect, and set up personalized guidance methods to improve patients' treatment compliance and treatment effect.

[0054] Furthermore, the pulmonary function curve is a flow-volume correlation curve graph, and the size, color and contrast of the curve graph need to be adjusted during the preprocessing process; the multi-type digital feature parameters include PEF, MMEF, FEF25%, FEF50%, FEF75%, PI F, MI F50, curve area and slope of the related position curve;

[0055] The PEF represents the maximum flow rate during forced exhalation, the MMEF represents the average flow rate when the forced exhalation vital capacity is 25%-75%; the FEF25%, FEF50%, and FEF75% are the expiratory flow rates when 25%, 50%, and 75% of the vital capacity are exhaled, respectively; the PIF represents the maximum inspiratory peak flow rate, and the MIF50 represents the instantaneous inspiratory flow rate when 50% of the forced vital capacity is used.

[0056] The drug efficacy of the present application mainly includes the distribution fraction of drug particles in various parts of the respiratory tract and the deposition efficiency of drug particles in the respiratory tract. The efficacy of drug particles is quantitatively analyzed by detecting the above parameters.

[0057] In order to improve the accuracy of respiratory function assessment, the lung function curves of normal people and patients of all ages and the simulation results of patients' lung function curves during respiratory drug inhalation should be saved in the database as much as possible to avoid data omissions. When the assessment results appear, they can be compared quickly and accurately to improve the overall assessment efficiency.

[0058] Preferably, the patient's lung function curve is deeply and accurately analyzed, the image information of the lung function curve is preprocessed and then digitized, the preprocessed image is extracted using a convolutional neural network, the curve data is analyzed, and multiple types of digital feature parameters are obtained, including the following steps: the convolution kernel of the convolutional neural network slides on the input image, and multiplication and summation operations are performed in the local area of ​​the image to obtain a feature map, wherein the feature map contains different digital feature parameters of the image; the feature map is sampled using a pooling layer to reduce the dimension and calculation amount of the feature map; the digitized image features are transmitted to the data analysis module for data screening to filter out data with large differences, incomplete and unnecessary data, and the importance of the features is evaluated using a random forest algorithm. The importance is used as a reference for weight allocation, and the weight of the feature is multiplied by the corresponding feature value to obtain a weighted feature set, and the weighted data is matched with the data in the database through Euclidean distance and cosine similarity calculations, and the data is compared and analyzed in the database using MLP.

[0059] For MLP, it is a neural network model, including a model framework, an activation function, a loss function and an optimizer; the model framework includes an input layer, a hidden layer and an output layer, the input data is forward propagated through each layer of the MLP, the output of each layer is calculated and the prediction result is obtained at the output layer; the activation function is a Sigmoid activation function, which can perform nonlinear changes in neuron output; the loss function uses mean square error to measure the difference between the model prediction result and the actual result; the optimizer is an Adam optimizer, which is used to minimize the loss function.

[0060] Furthermore, in the present application, the drug guidance given to the patient includes: the inhalation flow rate in the inhalation phase, the breath-holding time in the breath-holding phase, and the expiratory flow rate in the exhalation phase.

[0061] The analysis system corresponding to the analysis method includes a curve analysis module, a data analysis module, a drug administration guidance module and a feedback module which are arranged in coordination with each other.

[0062] Specifically, the curve analysis module includes a curve recognition unit, a preprocessing unit and a feature extraction unit. The curve recognition unit is used to identify the flow-volume curve of the patient's lung function curve. The preprocessing unit is used to adjust the size, color and contrast of the image and perform data enhancement to improve the generalization ability of the model. The feature extraction unit is used to convert image information into digital information and process the image into data that can be compared.

[0063] The data analysis module includes a feature screening unit, a weight allocation unit, a database and an efficacy prediction unit. In actual application, the feature screening unit will screen the transmitted data, select the required data and transmit it to the weight allocation unit. The weight allocation unit will allocate characteristic values ​​to different factors according to the input data, determine the importance of each factor in the drug deposition process, calculate the feature similarity of the weighted data, use a neural network to compare it with a large amount of data already stored in the database, find the data that best matches the patient, judge the patient's lung function, check the drug particle deposition distribution score and deposition efficiency corresponding to the data, predict the patient's drug efficacy according to the data in the database, and provide the conditions required for the best drug efficacy.

[0064] The medication guidance module includes an instruction generation unit, a text output unit and a real-time monitoring unit, which outputs the predicted optimal drug inhalation conditions for the patient in text form according to the situation predicted in the data analysis module; the feedback module is used to store the patient's actual inhalation drug process data and feed it back to the data analysis module, and the data analysis module adjusts the efficacy prediction and inhalation conditions according to the feedback data.

[0065] The feedback module can store the patient's actual inhaled drug process data and feed it back to the data analysis module. The data analysis module adjusts the efficacy prediction and inhalation conditions based on the feedback data.

[0066] It should be noted in particular that this application involves the field of medical health information technology, and is mainly intended to improve the efficiency of drug use in patients taking respiratory medications and to achieve customized respiratory medication guidance based on individual conditions.

[0067] In summary, the respiratory drug administration analysis method based on deep analysis of COPD lung function curve in the embodiment of the present invention converts image information into digital information by accurately and deeply analyzing the lung function curve, and assigns weights to different digital feature parameters according to the degree of influence on the patient's respiratory function. The use of digital processing and weight assignment can accurately capture subtle changes in the patient's respiratory function, reduce the subjectivity of the evaluation, and make the evaluation results more accurate and objective; at the same time, it can provide the best usage guidance for the patient's drug inhalation, so that the patient can inhale the drug according to the correct steps, and ensure the effective use of the drug to the greatest extent. The personalized guidance method set for different patients improves the patient's treatment compliance and treatment effect.

[0068] The above specific implementation manner cannot be used as a limitation on the protection scope of the present invention. For those skilled in the art, any substitution, improvement or change made to the implementation manner of the present invention falls within the protection scope of the present invention.

[0069] The matters not described in detail in the present invention are all known technologies to those skilled in the art.

Claims

1. A respiratory drug administration analysis method based on in-depth analysis of COPD lung function curves, characterized in that: The analytical method comprises the following steps: S1, deeply and accurately analyzes the patient's lung function curve, pre-processes the image information of the lung function curve and then digitizes it, uses a convolutional neural network to extract features from the pre-processed image, analyzes the curve data, and obtains multiple types of digital feature parameters; S2, assign weight coefficients to multiple types of digital feature parameters according to their influence, use neural networks to compare and analyze multiple types of digital feature parameters with database data to achieve accurate evaluation of the patient's respiratory function and predict the efficacy of corresponding drugs; S3, based on the patient's respiratory function assessment results and drug prediction effects, provides patients with the best drug inhalation conditions and sets up personalized guidance methods to improve patients' treatment compliance and treatment effects.

2. The respiratory drug administration analysis method based on deep analysis of COPD lung function curve according to claim 1, characterized in that: The pulmonary function curve is a flow-volume correlation curve, and the size, color and contrast of the curve need to be adjusted during the preprocessing process; the multi-type digital feature parameters include PEF, MMEF, FEF25%, FEF50%, FEF75%, PIF, MIF50, curve area and slope of the related position curve; The PEF represents the maximum flow rate during forced exhalation, the MMEF represents the average flow rate when the forced exhalation vital capacity is 25%-75%; the FEF25%, FEF50%, and FEF75% are the expiratory flow rates when 25%, 50%, and 75% of the vital capacity are exhaled, respectively; the PIF represents the maximum inspiratory peak flow rate, and the MIF 50 represents the instantaneous inspiratory flow rate when 50% of the forced vital capacity is exhaled.

3. The respiratory drug administration analysis method based on deep analysis of COPD lung function curve according to claim 2, characterized in that: The curve areas are the area between the inhalation curve and the abscissa axis, and the area between the exhalation curve and the abscissa axis.

4. The respiratory drug administration analysis method based on deep analysis of COPD lung function curve according to claim 1, characterized in that: The database contains lung function curves of normal people and patients of all ages and simulation results of patients' lung function curves during the process of drug inhalation in the respiratory tract.

5. The respiratory drug administration analysis method based on deep analysis of COPD lung function curve according to claim 1, characterized in that: The following steps are involved in deeply and accurately analyzing the patient's lung function curve, preprocessing the image information of the lung function curve and then digitizing it, extracting features from the preprocessed image using a convolutional neural network, analyzing the curve data, and obtaining multiple types of digital feature parameters: S1.1, the convolution kernel of the convolutional neural network slides on the input image, performs multiplication and summation operations on the local area of ​​the image, and obtains a feature map, which contains different digital feature parameters of the image; S1.2, use the pooling layer to sample the feature map to reduce the dimension and calculation amount of the feature map; S1.3, transfer the digitized image features to the data analysis module for data screening, filter out data with large differences, incomplete and unnecessary data, use the random forest algorithm to evaluate the importance of features, and use the importance as a reference for weight allocation. Multiply the feature weight by the corresponding feature value to obtain the weighted feature set. Use the Euclidean distance and cosine similarity calculations to match the weighted data with the data in the database, and use MLP to compare and analyze the data in the database.

6. The respiratory drug administration analysis method based on deep analysis of COPD lung function curve according to claim 5, characterized in that: The MLP includes a model framework, an activation function, a loss function and an optimizer; the model framework includes an input layer, a hidden layer and an output layer, the input data is forward propagated through each layer of the MLP, the output of each layer is calculated and the prediction result is obtained in the output layer; the activation function is a Sigmoid activation function, which can perform nonlinear changes in neuron output; The loss function uses mean square error to measure the difference between the model prediction result and the actual result; the optimizer is an Adam optimizer, which is used to minimize the loss function.

7. The respiratory drug administration analysis method based on deep analysis of COPD lung function curve according to claim 1, characterized in that: The optimal drug inhalation conditions include an optimal inhalation flow rate, a breath-holding time, and an optimal exhalation flow rate adapted to the patient when inhaling the drug.

8. The respiratory drug administration analysis method based on deep analysis of COPD lung function curve according to claim 1, characterized in that: The analysis system of the analysis method comprises a curve analysis module, a data analysis module, a drug administration guidance module and a feedback module; The curve analysis module includes a curve recognition unit, a preprocessing unit and a feature extraction unit. The curve recognition unit is used to recognize the flow-volume curve of the patient's lung function curve. The preprocessing unit is used to adjust the size, color and contrast of the image and perform data enhancement to improve the generalization ability of the model. The feature extraction unit is used to convert image information into digital information and process the image into data that can be compared with the same. The data analysis module includes a feature screening unit, a weight allocation unit, a database and a therapeutic effect prediction unit; The medication guidance module includes an instruction generation unit, a text output unit and a real-time monitoring unit, which outputs the predicted optimal drug inhalation conditions for the patient in text form according to the situation predicted in the data analysis module; the feedback module is used to store the patient's actual inhalation drug process data and feed it back to the data analysis module, and the data analysis module adjusts the efficacy prediction and inhalation conditions according to the feedback data.

9. The respiratory drug administration analysis method based on deep analysis of COPD lung function curve according to claim 8, characterized in that: The feature screening unit screens the transmitted data and selects the required data to transmit to the weight distribution unit. According to the input data, characteristic values ​​are assigned to different factors to determine the importance of each factor in the drug deposition process. The feature similarity of the data with assigned weights is calculated, and a neural network is used to compare the data with a large amount of data already stored in the database to find the data that best matches the patient, judge the patient's lung function, check the drug particle deposition distribution score and deposition efficiency corresponding to the data, predict the patient's drug efficacy according to the data in the database, and provide the conditions required for the best drug efficacy.

Images