Oxygen-adding atomization inhaler for treating interstitial lung disease

Abstract

The invention belongs to the technical field of medical instruments, and discloses an oxygen-adding atomization inhaler for treating interstitial lung disease, which comprises a respirator assembly, an oxygen bottle, a liquid medicine box, a waste liquid collecting box, an atomization pump, a negative pressure suction pump and a monitoring mechanism, the respirator assembly is connected with the oxygen bottle; the liquid medicine box is communicated with the oxygen bottle; a medicine mist supply pipe is arranged on the respirator assembly; the respirator assembly is slidably provided with a buccal device; an atomizing pump is arranged on the liquid medicine box; the atomizing pump is connected with the medicine mist supply pipe and the mouth holding device; a negative pressure suction pump is arranged on the waste liquid collecting box; the respirator assembly is provided with a sputum suction hose; and a monitoring mechanism is fixedly arranged on the liquid medicine box. According to the device, the mouth holding device or the medicine mist supply pipe can be selected for oxygen adding atomization inhalation treatment, abnormal conditions of a patient can be found in time, data of a plurality of monitoring indexes are integrated for fusion and analysis, potential risks are found in time, and when it is monitored that the patient has dyspnea, medicine feeding is automatically stopped, oxygen conveying is started, and sputum can be cleared in time.

Description

Technical Field

[0001] The present application relates to the technical field of medical devices, and more specifically, to an oxygenating nebulizer inhaler for treating interstitial lung disease. Background Art

[0002] Interstitial lung disease is a general term for clinical pathological entities composed of different types of disease groups with diffuse lung parenchyma, alveolar inflammation and interstitial fibrosis as the basic pathological lesions, and clinical manifestations such as active dyspnea, diffuse infiltration shadows on chest X-rays, restrictive ventilation disorders, reduced diffusion function and hypoxemia. The most important thing in the treatment of interstitial lung disease is the oxygenated nebulizer.

[0003] The document with the prior art publication number CN210331306U provides an oxygenated nebulizer for treating interstitial lung disease, comprising a shell, a handle fixedly mounted on one side of the shell, a drug inlet movably mounted on the top of the shell through a thread, the shell comprising an atomizing chamber and a mixing chamber, an oxygen tank fixedly mounted on the other side of the shell, the output end of the oxygen tank connected to the mixing chamber through a pipeline, the interior of the mixing chamber connected to a breathing mask through a pipeline, a connector fixedly mounted on one end of the breathing mask, the connector and the pipeline are connected through a thread, an ultrasonic nebulizer is fixedly mounted on one end of the atomizing chamber, and a nozzle is fixedly mounted on the bottom of the atomizing chamber. By arranging a handle, a rubber anti-slip sleeve, a thread, a connector and a groove, the problem that the existing oxygenated nebulizer for treating interstitial lung disease is large in size, inconvenient for patients to carry, and the breathing mask is inconvenient to disassemble, which is not conducive to long-term use of the oxygenated nebulizer for treating interstitial lung disease is solved.

[0004] Although the above-mentioned prior art solutions can achieve relevant beneficial effects through the structure of the prior art, they still have the following defects: 1. It is impossible to monitor the process of the patient's inhalation treatment and to detect abnormal conditions and potential risks in time. 2. During the treatment process, when the patient has difficulty breathing, it is impossible to stop the delivery of medicine in time and to deliver oxygen to the patient in time to assist the patient's breathing.

[0005] In view of this, we propose an oxygenating nebulizer for treating interstitial lung disease. Summary of the invention

[0006] 1. Technical issues to be solved

[0007] The purpose of the present application is to provide an oxygenated nebulizer inhaler and method for treating interstitial lung disease, which solves the technical problems raised in the above-mentioned background technology, and realizes the technical effect that a mouthpiece or a medicine mist supply tube can be selected for oxygenated nebulizer inhalation treatment according to the specific situation of the patient, abnormal conditions of the patient can be discovered in time, the data of multiple monitoring indicators can be integrated and analyzed to obtain comprehensive evaluation results, and potential risks can be discovered in time. When it is monitored that the patient has breathing difficulties, the drug delivery is automatically stopped and oxygen delivery is started, sputum can be cleared in time, and respiratory disorders caused by sputum blockage are avoided.

[0008] 2. Technical solution

[0009] The technical solution of the present application provides an oxygenating nebulizer for treating interstitial lung disease, comprising: a breathing mask assembly, an oxygen cylinder, a liquid medicine tank, a waste liquid collection tank, an nebulizer pump, a negative pressure suction pump and a monitoring mechanism;

[0010] The breathing mask assembly is connected to the oxygen cylinder through a hose;

[0011] The medicine tank is connected to the oxygen cylinder through a hose; the oxygen cylinder has a built-in pressure sensor to monitor the remaining oxygen in real time.

[0012] A medicine mist supply pipe is fixedly arranged on the breathing mask assembly; a mouthpiece is sealably and slidably arranged on the breathing mask assembly;

[0013] An atomizing pump is fixedly arranged on the medicine liquid tank; the atomizing pump is connected to the medicine mist supply pipe and the mouthpiece through a hose;

[0014] A negative pressure suction pump is fixedly arranged on the waste liquid collection box; a sputum suction hose is slidably arranged on the breathing mask assembly; the sputum suction hose is connected to the input end of the negative pressure suction pump; and the output end of the negative pressure suction pump is connected to the waste liquid collection box.

[0015] A monitoring mechanism is fixedly installed on the medicine tank to monitor the oxygen nebulization inhalation treatment process of patients with interstitial lung disease and detect abnormal conditions (such as dyspnea, etc.) in time.

[0016] Through the above technical scheme, according to the specific conditions of patients with interstitial lung disease, a mouthpiece or a medicine mist supply tube is selected to provide oxygenated atomized inhalation treatment for patients. The oxygen cylinder continuously delivers oxygen to the medicine liquid tank, and the atomizer pump atomizes the medicine liquid in the medicine liquid tank and delivers it to the mouthpiece or the medicine mist supply tube for the patient to inhale the treatment. The monitoring agency monitors the oxygenated atomized inhalation treatment process of patients with interstitial lung disease and detects abnormal conditions in time. The sputum in the patient's throat can be sucked into the waste liquid collection box through a sputum suction hose and a negative pressure suction pump.

[0017] As an optional solution of the present invention, the breathing mask assembly includes a mask body, an oxygen supply nasal suction tube, a sputum suction hose, a mouthpiece, an exhaust pipe and a medicine mist supply pipe;

[0018] An oxygen supply nasal suction tube is fixedly arranged on the mask body; two branch tubes are fixedly arranged on the oxygen supply nasal suction tube, which can provide oxygen to two nostrils respectively;

[0019] A sputum suction hose is slidably arranged on the mask body; sputum in the throat can be sucked through the sputum suction hose.

[0020] A mouthpiece is slidably provided on the mask body; the atomized medicine can be delivered to the mouth through the mouthpiece;

[0021] The cover body is fixedly provided with a plurality of exhaust pipes; each exhaust pipe is provided with a solenoid valve; one or more exhaust pipes can be selected to operate for exhaust as required, so as to adjust the gas pressure and humidity in the cover body.

[0022] The mask body is fixedly provided with a medicine mist supply pipe. Atomized medicine can be provided to the mask body through the medicine mist supply pipe. For patients who cannot actively cooperate, the medicine mist supply pipe is selected for treatment. For patients who cannot actively hold the mouthpiece, the medicine mist supply pipe is more suitable.

[0023] Through the above technical solution, the mask body fits tightly to the patient's face, and according to the specific situation of the patient, a mouthpiece or a medicine mist supply tube can be used to perform aerosol inhalation treatment on the patient; the sputum in the throat can be sucked out through a sputum suction hose to reduce the patient's discomfort. When the patient has difficulty breathing, oxygen can be delivered to the patient through an oxygen nasal cannula, which helps to improve the patient's breathing condition.

[0024] As an optional solution of the present invention, an oxygen flow valve B is provided on the oxygen delivery nasal suction tube. An oxygen flow valve A is provided on the hose between the oxygen cylinder and the liquid medicine tank. A dosing port is provided at the upper end of the liquid medicine tank, and a sealing plug is detachably fixedly provided on the dosing port.

[0025] As an optional solution of the present invention, a medicine mist flow valve B is provided on the hose connected to the medicine mist supply pipe and the atomizing pump;

[0026] A medicine mist flow valve A is arranged on the hose connected to the mouthpiece and the atomizing pump.

[0027] As an optional solution of the present invention, two gears are rotatably provided on the cover body, and the two gears are meshed and transmission-connected;

[0028] Rollers are coaxially fixed on the gears; the rollers are in frictional contact with the sputum suction hose;

[0029] A micro motor is fixedly arranged on the cover body, and an output end of the micro motor is coaxially fixedly connected with one of the gears.

[0030] A miniature camera and LED light are fixedly installed at the front end of the suction hose;

[0031] Through the above technical solution, a micro motor drives one of the gears to rotate, so that one of the two gears rotates synchronously in the opposite direction, and then the two rollers move synchronously in the opposite direction, driving the suction hose to move into the patient's throat or out of the throat.

[0032] As an optional solution of the present invention, the monitoring mechanism includes:

[0033] Data collection module: collects patient data, including medical history, case data, and image data during treatment, collects patient images, including normal images and abnormal images, and annotates the images as reference samples;

[0034] Sound collection module: collects sound data in the ward in real time;

[0035] Respiratory monitoring module: including a respiratory rate sensor, which monitors the patient's respiratory condition through the respiratory rate sensor;

[0036] Heart rate monitoring module: including a heart rate monitoring belt to monitor the patient's heart rate;

[0037] Blood oxygen saturation monitor: monitors the patient's blood oxygen saturation; uses a blood oxygen saturation probe, clipped on the patient's finger, to measure blood oxygen saturation through optical principles.

[0038] Image acquisition module: including a high-definition camera to collect high-definition images of patients;

[0039] Image preprocessing module: preprocess the collected images, including denoising, grayscale, image enhancement and normalization, etc., to eliminate interference factors in the image.

[0040] Feature extraction module: extracts features from the preprocessed image, including color, texture and shape; uses image processing algorithms to extract feature information such as color, texture and shape in the image.

[0041] Abnormal analysis module: uses machine learning or deep learning algorithms to analyze and identify images after feature extraction, and promptly identify abnormal conditions, such as dyspnea, aggravated cough, chest tightness and pain, and allergic reactions (rash, itching), etc.

[0042] Comprehensive evaluation module: Combines the monitoring results of blood oxygen saturation, heart rate, respiratory rate and abnormal analysis modules, integrates and analyzes the data of multiple monitoring indicators through algorithms, obtains comprehensive evaluation results, conducts comprehensive evaluation of patients, and promptly discovers abnormal conditions and potential risks;

[0043] Breathing optimization management module: controls the operation of the atomizer pump, drug mist flow valve A, drug mist flow valve B, micro motor and oxygen flow valve B; when the patient is detected to have breathing difficulties, the breathing optimization management module stops the control of the atomizer pump, drug mist flow valve A and drug mist flow valve B, and stops delivering medicine; opens the oxygen flow valve B, and delivers oxygen to the nostrils through the oxygen delivery nasal suction tube for the patient to breathe, which helps to improve the patient's breathing condition. At the same time, the alarm is triggered. At the same time, the micro motor works, driving the two gears to rotate synchronously in the opposite direction, so that the two rollers rotate synchronously in the opposite direction, pushing the sputum suction hose forward to the throat; the micro camera collects the image of the throat and analyzes whether the throat is blocked by sputum; if the throat is blocked by sputum, the negative pressure suction pump is started to suck the sputum into the waste liquid collection box, and the sputum is cleaned up as soon as possible to restore the patient to normal breathing.

[0044] Alarm module: including an alarm, which will promptly sound an alarm when abnormal conditions and potential risks are detected; using an audible and visual alarm or a vibration alarm to ensure that medical staff can receive alarm information in a timely manner.

[0045] PLC control module: connected to the data collection module, respiratory monitoring module, heart rate monitoring module, blood oxygen saturation monitor, image acquisition module, image preprocessing module, feature extraction module, abnormal analysis module, comprehensive evaluation module and alarm module network.

[0046] As an optional solution of the present invention, the abnormality analysis module analyzes and identifies the image after feature extraction, and promptly identifies abnormal situations, including the following steps:

[0047] 1. Data preparation: Obtain high-definition images of patients from the image acquisition module, including images in normal and abnormal states. Manually annotate the images and clearly mark various abnormal conditions, such as dyspnea, increased coughing, chest tightness, chest pain, and allergic reactions, to form an annotated data set.

[0048] 2. Model training: Select appropriate machine learning or deep learning algorithms based on the characteristics of image data and the needs of anomaly recognition. Common algorithms include convolutional neural networks (CNN), support vector machines (SVM), random forests, etc. Use labeled data sets to train the selected algorithm so that the model can learn the association between features in the image and anomalies. By adjusting model parameters, increasing the amount of training data, using data enhancement technology, etc., the performance of the model is continuously optimized to improve the accuracy of anomaly recognition.

[0049] 3. Abnormal identification: Use the feature extraction module to extract key features such as color, texture, shape, etc. from the image. Input the extracted features into the trained model to identify abnormalities (such as dyspnea, increased coughing, chest tightness, chest pain, and allergic reactions).

[0050] Dyspnea: Analyze the patient's chest rise and fall. Use deep learning algorithms to analyze breathing movement patterns and identify abnormal breathing.

[0051] Coughing aggravation: Monitors the coughing sound of the person in the image and analyzes the frequency, intensity and duration of the cough. Observes the body posture and facial expression of the person when coughing to help judge the severity of the cough.

[0052] Chest tightness and pain: Analyze the facial expressions and postures of people in the image, especially the signals related to pain and discomfort. With the help of medical knowledge and artificial intelligence algorithms, combined with image features, determine whether there is a possibility of chest tightness and pain.

[0053] Allergic reaction: Observe the changes in the color, texture, and morphology of the person's skin in the image, especially whether there are allergic symptoms such as rash, redness, swelling, itching, etc. Use deep learning algorithms to classify and identify skin images to accurately determine whether there is an allergic reaction.

[0054] 4. Output results: The model outputs recognition results based on the input features, determines whether there are any abnormalities in the image, and gives the specific type of abnormality.

[0055] 5. Alarm feedback: Set the threshold for abnormal recognition according to actual needs. When the recognition result exceeds the threshold, the alarm is triggered. When an abnormal situation is recognized, the alarm module will issue an alarm to remind medical staff to deal with it in time. The recognition results and alarm information will be recorded for subsequent analysis and processing by medical staff. At the same time, the performance and accuracy of the model will be continuously optimized based on the feedback from medical staff.

[0056] 6. Continuous monitoring and updating: During oxygen nebulizer treatment, the abnormal analysis module is continuously used to monitor and analyze the images. As the treatment process progresses and the patient's condition changes, the model is continuously updated and optimized to adapt to new abnormal situations and data features.

[0057] Through the above technical solution, the abnormality analysis module can accurately analyze and identify the images after feature extraction, timely discover and handle abnormal situations, and provide strong support for the decision-making of medical staff.

[0058] As an optional solution of the present invention, the comprehensive evaluation module combines the monitoring results of the blood oxygen saturation, heart rate, respiratory rate and abnormal analysis module, and integrates and analyzes the data of multiple monitoring indicators through an algorithm to timely discover abnormal situations and potential risks; including the following steps:

[0059] 1. Data collection: real-time collection of blood oxygen saturation data, heart rate, respiratory rate and abnormal analysis module data;

[0060] 2. Data preprocessing: including data cleaning, data standardization and data alignment;

[0061] Data cleaning: Remove noise, missing values ​​or outliers to ensure the accuracy and completeness of data.

[0062] Data standardization: Convert data from different sources and dimensions into a unified standard form to facilitate subsequent analysis.

[0063] Data alignment: Ensure that the data of various physiological parameters are consistent in time to facilitate synchronous analysis.

[0064] 3. Data fusion: Extract key features from the preprocessed data, such as the average value of blood oxygen saturation, the rate of change of heart rate, the stability of respiratory rate, etc.

[0065] Feature fusion: Use algorithms (such as weighted summation, principal component analysis, neural network, etc.) to fuse the features of multiple physiological parameters to form a comprehensive feature vector. Perform feature fusion according to the following formula:

[0066] Where F represents the comprehensive feature value, which is the comprehensive result obtained after multiple physiological parameter features are weighted, nonlinearly transformed and normalized. n represents the total number of physiological parameter features. i Represents the weight of the i-th physiological parameter feature, reflecting the importance of this feature in the comprehensive evaluation. β is an index for adjusting the weight, which is used to change the way the weight affects the comprehensive feature value. By adjusting the value of β, the weight of certain features can be made more prominent or weakened in the comprehensive evaluation. max(f) and min(f) represent the maximum and minimum values ​​of all physiological parameter feature values, respectively. These two values ​​are used to normalize the feature values ​​so that all feature values ​​can be compared and fused on the same scale. f i Represents the original value of the i-th physiological parameter feature. i Represents the nonlinear transformation index of the i-th physiological parameter feature. By adjusting α i The value of can change the nonlinear contribution of the eigenvalue to the comprehensive eigenvalue. Different features can have different α i values ​​to reflect their different importance or nonlinear relationship.

[0067] 4. Model training: Based on historical data, train a comprehensive assessment model (such as support vector machine, random forest, deep learning model, etc.), which can predict the patient's health status or risk level based on the comprehensive feature vector.

[0068] Where y is the dependent variable, which represents the target value we want to predict or explain. 0is the intercept term, which represents the expected value of y when all eigenvalues ​​are 0. i is the linear coefficient of the i-th feature, indicating the linear influence of this feature on y. i ) is the i-th feature F i The nonlinear transformation function is performed. This function can be of any form to capture the nonlinear effects of the features. ij It is feature F i and Fj. This coefficient represents the degree of influence of two features on y together, that is, the interaction effect between them. δ is the coefficient of the square term summed after all features are nonlinearly transformed. This coefficient is used to capture the overall nonlinear interaction effect between features, that is, the degree of influence of all features on y together after all features are nonlinearly transformed. ε is the error term, which represents the random variation or noise that is not captured in the model.

[0069] The data collected in real time is input into the comprehensive assessment model to obtain the comprehensive assessment results of the current patient.

[0070] 5. Result output and anomaly detection

[0071] Result output: The comprehensive evaluation results are presented to users or medical staff in an intuitive manner (such as dashboards, reports, etc.).

[0072] Anomaly detection: Set thresholds or rules. When the comprehensive evaluation results exceed a certain range, an abnormal alarm is triggered to prompt medical staff to take timely intervention measures.

[0073] 6. Feedback and Optimization

[0074] User feedback: Collect feedback from healthcare professionals and patients to understand the accuracy and usefulness of the comprehensive assessment module.

[0075] Model optimization: Based on feedback and new data, the comprehensive evaluation model is continuously optimized to improve the accuracy and reliability of the evaluation.

[0076] Through the above steps, the comprehensive evaluation module can realize comprehensive monitoring and analysis of patients' physiological parameters, timely detect abnormal conditions and potential risks, and provide strong support for medical staff's decision-making.

[0077] As an optional solution of the present invention, the breathing optimization management module collects throat images through a micro camera to analyze whether the throat is blocked by sputum, including the following steps:

[0078] 1. Image preprocessing: Get the original image of the throat from the micro camera. Remove noise and interference in the image, such as uneven lighting and blurred images. Use filtering algorithms (such as mean filtering and Gaussian filtering) to smooth the image. Improve the contrast and clarity of the image to better identify sputum features. Use grayscale transformation, image sharpening and other technologies to enhance the image.

[0079] 2. Feature extraction: Identify features related to sputum in the image, such as color, texture, shape, etc. Use shape feature extraction algorithms (such as contour detection algorithms) to extract the contours and boundaries of sputum.

[0080] 3. Throat structure identification: Identify the anatomical structure of the throat, such as the vocal cords, epiglottis, etc. Ensure that the normal structure of the throat is not misidentified as sputum during the analysis process.

[0081] 4. Sputum detection: Based on the extracted sputum features, determine whether there is sputum in the image. Use image recognition algorithms (such as support vector machine SVM, convolutional neural network CNN) to classify and identify sputum.

[0082] 5. Sputum volume assessment: Further evaluate the quantity and distribution of sputum. Quantify the amount of sputum by measuring indicators such as sputum area, volume or density.

[0083] 6. Risk assessment: Based on the amount and distribution of sputum, determine whether the sputum poses a threat to breathing.

[0084] If the amount of sputum is large or the distribution location is critical (such as blocking the airway), it is judged as high risk.

[0085] 7. Result output: Output the analysis results in the form of images, text or numbers for medical staff to review. Clearly mark the location, quantity and risk level of sputum in the analysis results.

[0086] Through the above technical solution, the image analysis system can accurately analyze the collected throat images, determine whether there is sputum blockage, and trigger the alarm in time, providing strong support for the subsequent treatment of medical staff.

[0087] 3. Beneficial effects

[0088] One or more technical solutions provided in the technical solution of this application have at least the following technical effects or advantages:

[0089] 1. The present invention can select a mouthpiece or a medicine mist supply tube for oxygenated atomization inhalation therapy according to the specific conditions of the patient to meet personalized treatment needs.

[0090] 2. Real-time monitoring of the patient's breathing, heart rate, blood oxygen saturation and other physiological indicators to provide doctors with comprehensive patient status information.

[0091] 3. Through image acquisition, preprocessing, feature extraction and abnormality analysis, abnormal conditions of patients can be discovered in time, improving the accuracy and timeliness of diagnosis.

[0092] 4. Integrate and analyze data from multiple monitoring indicators to obtain comprehensive evaluation results, identify potential risks in a timely manner, and provide decision-making support for doctors.

[0093] 5. When it is detected that the patient has breathing difficulties, the drug delivery will be automatically stopped and oxygen delivery will be started, which will help improve the patient's breathing condition and reflect the intelligent management of the equipment.

[0094] 6. Sputum can be cleared in time. The suction hose is pushed to the throat by a micro motor and roller, and the sputum is cleared in time using a negative pressure suction pump, avoiding respiratory disorders caused by sputum blockage and ensuring the safety of patients. BRIEF DESCRIPTION OF THE DRAWINGS

[0095] Figure 1 This is an overall schematic diagram of an oxygenating nebulizer for treating interstitial lung disease disclosed in a preferred embodiment of the present application;

[0096] Figure 2 This is a schematic diagram of a breathing mask assembly of an oxygenating nebulizer inhaler for treating interstitial lung disease disclosed in a preferred embodiment of the present application.

[0097] Reference numerals:

[0098] 1. Breathing mask assembly; 2. Oxygen cylinder; 3. Medicine liquid tank; 4. Waste liquid collection tank; 5. Atomizing pump; 6. Negative pressure suction pump; 7. Push mechanism; 11. Mask body; 12. Oxygen nasal suction tube; 13. Suction hose; 14. Mouthpiece; 15. Exhaust pipe; 16. Medicine mist supply pipe; 17. Sealing partition; 18. Locking knob; 21. Oxygen flow valve A; 22. Oxygen flow valve B; 31. Dosing port; 51. Medicine mist flow valve A; 52. Medicine mist flow valve B; 71. Gear; 72. Roller. DETAILED DESCRIPTION

[0099] The present application is further described in detail below in conjunction with the accompanying drawings.

[0100] Reference Figure 1 , the embodiment of the present application provides an oxygenating nebulizer for treating interstitial lung disease, comprising: a breathing mask assembly 1, an oxygen cylinder 2, a liquid medicine tank 3, a waste liquid collection tank 4, an nebulizer pump 5, a negative pressure suction pump 6 and a monitoring mechanism;

[0101] The breathing mask assembly 1 is connected to the oxygen cylinder 2 through a hose; the breathing mask assembly 1 is made of medical-grade silicone or soft TPU material to ensure comfort for long-term wearing.

[0102] The liquid medicine tank 3 is connected to the oxygen cylinder 2 through a hose; the oxygen cylinder 2 has a built-in pressure sensor to monitor the remaining oxygen in real time and remind replacement through an LED indicator or wireless signal. Equipped with a precision flow control valve, the oxygen supply rate can be adjusted according to the patient's needs. The liquid medicine tank 3 is made of transparent medical grade plastic, which is convenient for observing the remaining liquid medicine. The liquid medicine tank 3 has a built-in constant temperature heating device to keep the liquid medicine at a suitable temperature, improve the atomization efficiency and comfort. A liquid level sensor is set in the liquid medicine tank 3 to monitor the amount of liquid medicine. When the liquid level is low, it will automatically alarm and stop atomization.

[0103] A medicine mist supply pipe 16 is fixedly provided on the breathing mask assembly 1; a mouthpiece 14 is sealably and slidably provided on the breathing mask assembly 1;

[0104] The liquid medicine tank 3 is fixed with an atomizing pump 5, which is connected to the medicine mist supply pipe 16 and the mouthpiece 14 through a hose. The atomizing pump 5 uses ultrasonic or compression atomizing technology to ensure uniform atomization of the liquid medicine. The built-in noise reduction design reduces noise interference during treatment.

[0105] A negative pressure suction pump 6 is fixedly arranged on the waste liquid collection box 4; a sputum suction hose 13 is slidably arranged on the breathing mask assembly 1; the sputum suction hose 13 is connected to the input end of the negative pressure suction pump 6; and the output end of the negative pressure suction pump 6 is connected to the waste liquid collection box 4. The negative pressure suction pump 6 can adjust the negative pressure intensity according to the viscosity of the sputum. The pressure change in the sputum suction hose 13 is detected by a sensor and automatically started and stopped.

[0106] A monitoring mechanism is fixedly arranged on the medicine tank 3, and the monitoring mechanism monitors the oxygen nebulization inhalation treatment process of the interstitial lung disease patient and promptly detects abnormal conditions (such as dyspnea, etc.).

[0107] In this technical solution, according to the specific conditions of patients with interstitial lung disease, a mouthpiece 14 or a medicine mist supply pipe 16 is selected to provide oxygenated nebulized inhalation treatment for the patient. The oxygen cylinder 2 continuously supplies oxygen to the medicine liquid tank 3, and the nebulizer pump 5 atomizes the medicine liquid in the medicine liquid tank 3 and delivers it to the mouthpiece 14 or the medicine mist supply pipe 16 for the patient to inhale the treatment. The monitoring agency monitors the oxygenated nebulized inhalation treatment process of patients with interstitial lung disease and promptly detects abnormal conditions (such as dyspnea). The sputum in the patient's throat can be sucked into the waste liquid collection box 4 through the sputum suction hose 13 and the negative pressure suction pump 6. Not only does it improve the treatment efficiency, but it also enhances the patient's comfort and safety, while realizing intelligent management and remote monitoring, providing a more comprehensive, convenient and personalized solution for the treatment of patients with interstitial lung disease.

[0108] Reference Figure 2 The breathing mask assembly 1 includes a mask body 11, an oxygen supply nasal suction tube 12, a sputum suction hose 13, a mouthpiece 14, an exhaust pipe 15 and a medicine mist supply pipe 16;

[0109] An oxygen supply nasal suction tube 12 is fixedly provided on the mask body 11; two branch tubes are fixedly provided on the oxygen supply nasal suction tube 12, which can provide oxygen to two nostrils respectively;

[0110] A sputum suction hose 13 is slidably provided on the cover body 11 ; the sputum in the throat can be sucked through the sputum suction hose 13 .

[0111] A mouthpiece 14 is slidably provided on the mask body 11; the atomized medicine can be delivered to the mouth through the mouthpiece 14; the mask body 11 is made of a soft and breathable material to improve the patient's comfort. Two sets of buckles are provided on the mask body 11 to fix the mask body 11 to the head. In the case where the patient can actively cooperate, the mouthpiece is selected to be used. The mouthpiece requires the patient to actively hold it in the mouth, and is suitable for patients who can understand and cooperate with treatment instructions. The mouthpiece usually allows more precise drug delivery, especially when the treatment requires a specific dose of the drug.

[0112] The cover body 11 is fixedly provided with a plurality of exhaust pipes 15 ; each exhaust pipe 15 is provided with a solenoid valve; one or more exhaust pipes 15 can be selected to operate for exhaust as required, so as to adjust the gas pressure and humidity in the cover body.

[0113] The mask body 11 is fixedly provided with a medicine mist supply pipe 16. Atomized medicine can be provided to the mask body 11 through the medicine mist supply pipe 16. In the case where the patient cannot actively cooperate, the medicine mist supply pipe 16 is selected for treatment. For patients who cannot actively hold the mouthpiece (such as children, comatose patients or patients with severe breathing difficulties), the medicine mist supply pipe is more suitable.

[0114] The cover body 11 is provided with a locking knob 18, through which the mouthpiece 14 can be adjusted in position and locked.

[0115] A sealing partition 17 is slidably provided on the mask body 11. When the mouthpiece 14 is not used for oxygen inhalation therapy, the mouthpiece 14 is withdrawn from the mask body 11 and the sealing partition 17 is pressed. The sealing partition 17 isolates the internal space of the mask body 11 and the mouthpiece 14 to prevent the mouthpiece 14 from being contaminated during use.

[0116] In this technical solution, the mask body 11 fits tightly to the patient's face, and according to the patient's specific conditions, the mouthpiece 14 or the medicine mist supply tube 16 can be used to perform aerosol inhalation treatment on the patient; the sputum in the throat can be sucked out through the sputum suction hose 13 to reduce the patient's discomfort. When the patient has difficulty breathing, oxygen can be delivered to the patient through the oxygen nasal suction tube 12, which helps to improve the patient's breathing condition.

[0117] Furthermore, an oxygen flow valve B22 is provided on the oxygen supply nasal suction tube 12.

[0118] An oxygen flow valve A21 is provided on the hose between the oxygen cylinder 2 and the medicine liquid tank 3 .

[0119] A drug adding port 31 is disposed at the upper end of the drug liquid tank 3 , and a sealing plug is detachably fixedly disposed on the drug adding port 31 .

[0120] Furthermore, a medicine mist flow valve B52 is provided on the hose connected to the medicine mist supply pipe 16 and the atomizing pump 5;

[0121] A medicine mist flow valve A51 is provided on the hose connected to the mouthpiece 14 and the atomizing pump 5.

[0122] Furthermore, the cover body 11 is rotatably provided with two gears 71, and the two gears 71 are meshed and transmission-connected;

[0123] A roller 72 is coaxially fixedly disposed on the gear 71; the roller 72 is frictionally contacted with the sputum suction hose 13;

[0124] A micro motor is fixedly mounted on the cover body 11 , and an output end of the micro motor is coaxially fixedly connected to one of the gears 71 .

[0125] A miniature camera and an LED light are fixedly arranged at the front end of the sputum suction hose 13;

[0126] In this technical solution, a micro motor drives one of the gears 71 to rotate, so that one of the two gears 71 rotates synchronously in opposite directions, and then the two rollers 72 move synchronously in opposite directions, driving the suction hose 13 to move into or out of the patient's throat.

[0127] Furthermore, monitoring agencies include:

[0128] Data collection module: collects patient data, including medical history, case data, and image data during treatment, collects patient images, including normal images and abnormal images, and annotates the images as reference samples;

[0129] Sound collection module: collects sound data in the ward in real time;

[0130] Respiratory monitoring module: including a respiratory rate sensor, which monitors the patient's respiratory condition through the respiratory rate sensor;

[0131] Heart rate monitoring module: including a heart rate monitoring belt to monitor the patient's heart rate;

[0132] Blood oxygen saturation monitor: monitors the patient's blood oxygen saturation; uses a blood oxygen saturation probe, clipped on the patient's finger, to measure blood oxygen saturation through optical principles.

[0133] Image acquisition module: including a high-definition camera to collect high-definition images of patients;

[0134] Image preprocessing module: preprocess the collected images, including denoising, grayscale, image enhancement and normalization, etc., to eliminate interference factors in the image.

[0135] Feature extraction module: extracts features from the preprocessed image, including color, texture and shape; uses image processing algorithms to extract feature information such as color, texture and shape in the image.

[0136] Abnormal analysis module: uses machine learning or deep learning algorithms to analyze and identify images after feature extraction, and promptly identify abnormal conditions, such as dyspnea, aggravated cough, chest tightness and pain, and allergic reactions (rash, itching), etc.

[0137] Comprehensive evaluation module: Combines the monitoring results of blood oxygen saturation, heart rate, respiratory rate and abnormal analysis modules, integrates and analyzes the data of multiple monitoring indicators through algorithms, obtains comprehensive evaluation results, conducts comprehensive evaluation of patients, and promptly discovers abnormal conditions and potential risks;

[0138] Breathing optimization management module: control the operation of the atomizing pump 5, the medicine mist flow valve A51, the medicine mist flow valve B52, the micro motor and the oxygen flow valve B22; when the patient is detected to have breathing difficulties, the breathing optimization management module controls the atomizing pump 5, the medicine mist flow valve A51 and the medicine mist flow valve B52 to stop working and stop delivering medicine; open the oxygen flow valve B22, and deliver oxygen to the nostrils through the oxygen delivery nasal suction tube 12 for the patient to breathe, which helps to improve the patient's breathing condition. At the same time, the alarm is triggered. At the same time, the micro motor works, driving the two gears 71 to rotate synchronously in the opposite direction, so that the two rollers 72 rotate synchronously in the opposite direction, and push the sputum suction hose 13 forward to the throat; collect the image of the throat through the micro camera, and analyze whether the throat is blocked by sputum; if the throat is blocked by sputum, start the negative pressure suction pump 6 to suck the sputum into the waste liquid collection box 4, and clean the sputum as soon as possible, so that the patient can resume normal breathing.

[0139] Alarm module: including an alarm, which will promptly sound an alarm when abnormal conditions and potential risks are detected; using an audible and visual alarm or a vibration alarm to ensure that medical staff can receive alarm information in a timely manner.

[0140] PLC control module: connected to the data collection module, respiratory monitoring module, heart rate monitoring module, blood oxygen saturation monitor, image acquisition module, image preprocessing module, feature extraction module, abnormal analysis module, comprehensive evaluation module and alarm module. Realize network connection and coordinated control between modules. Adopt programmable logic controller (PLC) to realize data communication and coordinated control between modules through programming, so as to ensure stable operation and efficient work of monitoring mechanism.

[0141] Furthermore, the abnormality analysis module analyzes and identifies the image after feature extraction, and promptly identifies abnormal situations, including the following steps:

[0142] 1. Data preparation: Obtain high-definition images of patients from the image acquisition module, including images in normal and abnormal states. Manually annotate the images and clearly mark various abnormal conditions, such as dyspnea, increased coughing, chest tightness, chest pain, and allergic reactions, to form an annotated data set.

[0143] 2. Model training: Select appropriate machine learning or deep learning algorithms based on the characteristics of image data and the needs of anomaly recognition. Common algorithms include convolutional neural networks (CNN), support vector machines (SVM), random forests, etc.

[0144] Train the model: Use the labeled dataset to train the selected algorithm so that the model can learn the association between features in the image and anomalies.

[0145] Optimize the model: Continuously optimize the performance of the model and improve the accuracy of anomaly identification by adjusting model parameters, increasing the amount of training data, and using data enhancement technology.

[0146] 3. Abnormal identification: Use the feature extraction module to extract key features such as color, texture, shape, etc. from the image. Input the extracted features into the trained model to identify abnormalities (such as dyspnea, increased coughing, chest tightness, chest pain, and allergic reactions).

[0147] Dyspnea: Analyze the patient's chest rise and fall. Use deep learning algorithms to analyze breathing movement patterns and identify abnormal breathing.

[0148] Coughing aggravation: Monitors the coughing sound of the person in the image and analyzes the frequency, intensity and duration of the cough. Observes the body posture and facial expression of the person when coughing to help judge the severity of the cough.

[0149] Chest tightness and pain: Analyze the facial expressions and postures of people in the image, especially the signals related to pain and discomfort. With the help of medical knowledge and artificial intelligence algorithms, combined with image features, determine whether there is a possibility of chest tightness and pain.

[0150] Allergic reaction (rash, itching): Observe the changes in the color, texture, and morphology of the person's skin in the image, especially whether there are allergic symptoms such as rash, redness, swelling, itching, etc. Use deep learning algorithms to classify and identify skin images to accurately determine whether there is an allergic reaction.

[0151] 4. Output results: The model outputs recognition results based on the input features, determines whether there are any abnormalities in the image, and gives the specific type of abnormality.

[0152] 5. Alarm feedback: Set the threshold for abnormal recognition according to actual needs. When the recognition result exceeds the threshold, an alarm is triggered.

[0153] When an abnormal situation is identified, an alarm is issued through the alarm module to remind medical staff to deal with it in time.

[0154] The recognition results and alarm information are recorded for subsequent analysis and processing by medical staff. At the same time, the performance and accuracy of the model are continuously optimized based on the feedback from medical staff.

[0155] 6. Continuous monitoring and updating: During oxygen nebulizer treatment, the abnormal analysis module is continuously used to monitor and analyze the images.

[0156] As the treatment process progresses and the patient's condition changes, the model is continuously updated and optimized to adapt to new abnormal situations and data characteristics.

[0157] In this technical solution, the abnormality analysis module can accurately analyze and identify the images after feature extraction, detect and handle abnormal situations in a timely manner, and provide strong support for the decision-making of medical staff.

[0158] Further, the cough exacerbation analysis includes the following steps:

[0159] 1. Cough frequency analysis: Count the number of cough sound segments to calculate the cough frequency. The cough frequency can be expressed as the number of coughs per minute or the total number of coughs during the entire observation period.

[0160] 2. Cough intensity analysis: Analyze the amplitude or volume of the cough sound to assess the intensity of the cough. Use audio analysis software to measure the peak amplitude or average amplitude of the cough sound. Cough intensity can be divided into mild, moderate and severe levels, and the specific classification criteria can be formulated according to actual conditions.

[0161] 3. Cough duration analysis: Measure the duration of each cough sound segment. Calculate the total duration of all cough sound segments to evaluate the duration of the cough.

[0162] 4. Body posture analysis: Analyze the body posture of the person when coughing, including the swing of the head, bending or straightening of the body, etc. Pay attention to whether there are accompanying symptoms such as difficulty breathing, wheezing or chest pain. Changes in body posture can provide additional information about the severity of the cough.

[0163] 5. Facial expression analysis: Analyze the facial expression of the character when coughing, including whether it is painful, tired or anxious, etc. Facial expressions can reflect the degree of impact of coughing on the character and the discomfort they may feel.

[0164] 6. Comprehensive judgment: Combine the results of sound analysis and observation of body posture and facial expression to comprehensively judge the severity of the cough. Pay attention to factors such as the frequency, intensity, duration and accompanying symptoms of the cough.

[0165] In this technical solution, the aggravation of cough can be quantified and analyzed more accurately, providing stronger support for taking corresponding treatment measures.

[0166] Furthermore, the dyspnea recognition includes the following steps:

[0167] 1. Image preprocessing: De-noise the image to remove noise and interference. Perform image enhancement to improve the contrast and clarity of the image so as to more accurately analyze the chest rise and fall. Perform image registration to align images at different time points for subsequent analysis.

[0168] 2. Chest rise and fall analysis: Use image processing technology (such as edge detection, contour tracking, etc.) to locate the patient's chest area. Ensure accurate positioning so that subsequent analysis can accurately reflect the rise and fall of the chest.

[0169] 3. Measure the chest rise and fall, extract the chest contour or edge in the located chest area, calculate the chest contour diameter, area or circumference and other parameters at different time points, and quantify the chest rise and fall.

[0170] 4. Calculation of chest rise and fall rate: Based on the extracted chest parameters, calculate the chest rise and fall rate, that is, the rate of change of chest parameters over time. The chest rise and fall rate can reflect the rate and depth of breathing, and is an important indicator for judging whether breathing is abnormal. Calculate the chest rise and fall rate according to the following formula:

[0171] R chest (t) = [A(t) - A(t-1)] / Δt; where R chest (t) is the chest rise and fall rate at time t. It is used to describe the rate of change of chest area per unit time. It reflects the rate and depth of breathing and is an important basis for judging whether breathing is abnormal. ΔA(t) is the change in chest area from time t-1 to time t. A(t) is the chest area at time t. This is a specific value, which represents the cross-sectional area of ​​the patient's chest at a specific time point (ie time t). This value can be measured by image processing technology. A(t-1) is the chest area at time t-1. Δt is the time interval. This is a value indicating the length of time.

[0172] 5. Abnormal breathing recognition: Extract features from chest rise and fall data, such as breathing frequency, breathing depth, average and standard deviation of chest rise and fall rate. These features can be used as the basis for subsequent abnormal breathing recognition. Establish an abnormal breathing recognition model and select appropriate machine learning or deep learning algorithms, such as support vector machine (SVM), random forest (RandomForest) or convolutional neural network (CNN).

[0173] The model is trained using a labeled training dataset that includes samples of normal and abnormal breathing.

[0174] Adjust the model parameters to optimize the model's performance and ensure that the model can accurately identify abnormal breathing.

[0175] 6. Model validation and application: Use the validation data set to evaluate the performance of the model to ensure that the model can accurately identify abnormal breathing in actual applications. Apply the trained model to the chest rise and fall data analysis of actual patients to identify whether there is abnormal breathing.

[0176] 7. Result output: Based on the output of the model, determine whether the patient's breathing is abnormal. If abnormal breathing is identified, further examination of the patient's physical condition is required to determine the cause of the breathing difficulty. Based on the identification results and the patient's physical condition, corresponding treatment suggestions are given.

[0177] Further, the chest tightness and chest pain analysis includes the following steps:

[0178] 1. Feature extraction: Use image processing techniques (such as edge detection, texture analysis, color analysis, etc.) to extract features related to facial expressions and postures. These features include the degree of frowning of eyebrows, the degree of squinting of eyes, the degree of opening and closing of mouth, the degree of facial distortion, body posture and movement, etc.

[0179] 2. Facial expression analysis:

[0180] Pain expression recognition: Pain usually causes facial expressions such as frowning eyebrows, squinting eyes, and drooping mouth corners. Use artificial intelligence algorithms (such as convolutional neural networks (CNNs)) to classify and recognize the extracted facial expression features to determine whether there is a pain expression.

[0181] Analysis of uncomfortable expressions: Uncomfortable expressions may not be as obvious as painful expressions, but they are usually accompanied by tension and distortion of facial muscles. By analyzing subtle changes in facial muscles, it is possible to determine whether there is an uncomfortable expression.

[0182] 3. Posture analysis:

[0183] Posture recognition: Uses AI algorithms to identify postures of people in images, especially those associated with pain and discomfort. For example, holding one’s hands to one’s abdomen and leaning forward may indicate abdominal or chest pain.

[0184] Movement analysis: Analyze whether the character's movements are smooth and natural, or whether they are stiff, slow, or abnormal. Abnormal movements may indicate physical discomfort or pain.

[0185] 4. Inference of pain location: Based on facial expressions and body posture characteristics, combined with medical knowledge, the possible location of pain can be inferred. For example, chest pain may be accompanied by body posture characteristics such as holding the abdomen with both hands and shortness of breath.

[0186] 5. Assessment of the possibility of chest tightness and chest pain: Comprehensively assess the facial expressions, body characteristics and possible pain sites to assess the possibility of chest tightness and chest pain. If multiple features point to chest tightness and chest pain, the possibility is high; otherwise, the possibility is low. Follow the following formula to assess the possibility of chest tightness and chest pain:

[0187] P chest_pain =w 1 *E pain +w 2 *E discomfort +w 3 *P pain_related +w 4 *A abnormal +w 5 *L pain_correlation +b;

[0188] P chest_pain Represents the estimated value of the possibility of chest tightness or chest pain. It is a numerical value, usually between 0 and 1, used to indicate the possibility of chest tightness or chest pain calculated based on the input features. 1 、w 2 、w 3 、w 4 and w 5 It is the weight of each feature, which determines the contribution of each feature to the final evaluation result. The weight is determined by the training process of the machine learning algorithm and is usually a non-negative value. The size of the weight reflects the importance of the feature in predicting the possibility of chest tightness and chest pain. pain is the recognition result of the pain expression. It can be a probability value (indicating the possibility of the presence of a pain expression) or a classification label (such as 0 for no pain and 1 for pain). This value is obtained by analyzing the facial expression features in the image. discomfort is the analysis result of uncomfortable expression. It can be a probability value or classification label, which is used to indicate the possibility of uncomfortable expression. pain_relatedis the result of pain-related posture recognition. It can be a numerical value or a classification label to indicate the degree to which the posture of the person in the image is related to pain. abnormal It is the result of abnormal motion analysis. It indicates whether the motion of the person in the image is smooth and natural, or whether there are abnormal states such as stiffness and slowness. This value can also be a numerical value or a classification label. pain_correlation It is the inference result of the correlation between the pain location and chest tightness and chest pain. It infers the possible location of pain based on facial expressions, body features and medical knowledge, and evaluates the correlation between this location and chest tightness and chest pain. This value can also be a numerical value or a classification label. b is the bias term, which represents the base value of the possibility of chest tightness and chest pain when there is no feature input (that is, when all feature values ​​are 0). The bias term is a constant used to adjust the reference point of the final evaluation result.

[0189] 6. Auxiliary diagnosis and suggestions: The analysis results are provided to doctors as auxiliary diagnosis information to help doctors judge the patient's condition more accurately.

[0190] Furthermore, the allergic reaction analysis includes the following steps:

[0191] 1. Feature extraction:

[0192] Color analysis: Analyzes the color distribution of skin images, especially whether there is abnormal redness or discoloration.

[0193] Texture analysis: Use image processing technology to extract the texture characteristics of the skin, such as roughness, uniformity, etc., to detect whether there is a rash or abnormal changes in skin texture.

[0194] Morphological change analysis: Observe the morphological changes on the skin surface, such as whether there are raised rashes, blisters, or skin swelling.

[0195] 2. Model training: Build an allergic reaction recognition model; build a dataset containing known allergic reactions (rash, itching, etc.) and normal skin images. Use deep learning algorithms (such as convolutional neural networks (CNN)) to train the dataset so that the model can learn and recognize the characteristics of allergic reactions. The allergic reaction recognition model is:

[0196] θ ★ =argmin θ {∑ N i＝1 -[y i log(σ(W T x i +d))+(1-y i )log(1-σ(W T x i +d))]};

[0197] In the formula, θ ★ Represents the optimal solution for the model parameters. θ It means to find the parameter set that minimizes the objective function (i.e., the sum of the loss functions in the brackets) in the parameter space θ. N i＝1 It means to sum all N samples in the data set. i Represents the true label or target value of the i-th sample in the dataset. σ represents the Sigmoid activation function, which maps the input to the (0,1) interval and is usually used in the output layer of a binary classification problem. W T represents the transpose of the weight matrix W. In a neural network, the weight matrix is ​​used to connect the input layer and the output layer (or hidden layer). i Represents the feature vector of the i-th sample in the data set. It is a column vector that contains all the features of the sample. d represents the bias term (or bias parameter). In neural networks, the bias term is used to adjust the output to better fit the data. log represents the natural logarithm function. In mathematics and machine learning, the logarithm function is often used to define the loss function because it can measure the difference between two probability distributions.

[0198] 3. Classification and recognition: The preprocessed skin images are input into the trained model for classification and recognition to determine whether an allergic reaction occurs.

[0199] 4. Result judgment: Based on the output of the model, determine whether there is an allergic reaction in the image. The model outputs a probability value or classification label to indicate the possibility of an allergic reaction. The judgment result is provided to doctors or professionals as auxiliary diagnostic information to help them more accurately judge the patient's condition.

[0200] 5. Validation and optimization: Use an independent validation dataset to evaluate the performance of the model to ensure its accuracy and reliability. Based on the validation results, adjust and optimize the model to improve its accuracy and generalization ability in identifying allergic reactions.

[0201] Furthermore, the comprehensive evaluation module combines the monitoring results of the blood oxygen saturation, heart rate, respiratory rate and abnormal analysis module, integrates and analyzes the data of multiple monitoring indicators through algorithms, obtains comprehensive evaluation results, conducts comprehensive evaluation of patients, and promptly discovers abnormal conditions and potential risks; including the following steps:

[0202] 1. Data collection: Collect blood oxygen saturation data, heart rate, respiratory rate and abnormal analysis module data in real time; use blood oxygen saturation sensor to measure the patient's blood oxygen level in real time. Monitor the patient's heart rate through heart rate sensor. Monitor the patient's respiratory rate using respiratory belt, chest impedance change or breath sound recognition. Obtain abnormal alarm or warning information about the patient's physiological parameters from the abnormal analysis module.

[0203] 2. Data preprocessing: including data cleaning, data standardization and data alignment;

[0204] Data cleaning: Remove noise, missing values ​​or outliers to ensure the accuracy and completeness of data.

[0205] Data standardization: Convert data from different sources and dimensions into a unified standard form to facilitate subsequent analysis.

[0206] Data alignment: Ensure that the data of various physiological parameters are consistent in time to facilitate synchronous analysis.

[0207] 3. Data fusion: Extract key features from the preprocessed data, such as the average value of blood oxygen saturation, the rate of change of heart rate, the stability of respiratory rate, etc.

[0208] Feature fusion: Use algorithms (such as weighted summation, principal component analysis, neural network, etc.) to fuse the features of multiple physiological parameters to form a comprehensive feature vector. Perform feature fusion according to the following formula:

[0209] Where F represents the comprehensive feature value, which is the comprehensive result obtained after multiple physiological parameter features are weighted, nonlinearly transformed and normalized. n represents the total number of physiological parameter features. i Represents the weight of the i-th physiological parameter feature, reflecting the importance of this feature in the comprehensive evaluation. β is an index for adjusting the weight, which is used to change the way the weight affects the comprehensive feature value. By adjusting the value of β, the weight of certain features can be made more prominent or weakened in the comprehensive evaluation. max(f) and min(f) represent the maximum and minimum values ​​of all physiological parameter feature values, respectively. These two values ​​are used to normalize the feature values ​​so that all feature values ​​can be compared and fused on the same scale. f i Represents the original value of the i-th physiological parameter feature. i Represents the nonlinear transformation index of the i-th physiological parameter feature. By adjusting α i The value of can change the nonlinear contribution of the eigenvalue to the comprehensive eigenvalue. Different features can have different α i values ​​to reflect their different importance or nonlinear relationship.

[0210] 4. Model training: Based on historical data, train a comprehensive assessment model (such as support vector machine, random forest, deep learning model, etc.), which can predict the patient's health status or risk level based on the comprehensive feature vector.

[0211] Where y is the dependent variable, which represents the target value we want to predict or explain. 0 is the intercept term, which represents the expected value of y when all eigenvalues ​​are 0. i is the linear coefficient of the i-th feature, indicating the linear influence of this feature on y. i ) is the i-th feature F i The nonlinear transformation function is performed. This function can be of any form to capture the nonlinear effects of the features. ij It is feature F i and Fj. This coefficient represents the degree of influence of two features on y together, that is, the interaction effect between them. δ is the coefficient of the square term summed after all features are nonlinearly transformed. This coefficient is used to capture the overall nonlinear interaction effect between features, that is, the degree of influence of all features on y together after all features are nonlinearly transformed. ε is the error term, which represents the random variation or noise that is not captured in the model.

[0212] The data collected in real time is input into the comprehensive assessment model to obtain the comprehensive assessment results of the current patient.

[0213] 5. Result output and anomaly detection

[0214] Result output: The comprehensive evaluation results are presented to users or medical staff in an intuitive manner (such as dashboards, reports, etc.).

[0215] Anomaly detection: Set thresholds or rules. When the comprehensive evaluation results exceed a certain range, an abnormal alarm is triggered to prompt medical staff to take timely intervention measures.

[0216] 6. Feedback and Optimization

[0217] User feedback: Collect feedback from healthcare professionals and patients to understand the accuracy and usefulness of the comprehensive assessment module.

[0218] Model optimization: Based on feedback and new data, the comprehensive evaluation model is continuously optimized to improve the accuracy and reliability of the evaluation.

[0219] Through the above steps, the comprehensive evaluation module can realize comprehensive monitoring and analysis of patients' physiological parameters, timely detect abnormal conditions and potential risks, and provide strong support for medical staff's decision-making.

[0220] Furthermore, the breathing optimization management module collects throat images through a micro camera and analyzes whether the throat is blocked by sputum, including the following steps:

[0221] 1. Image preprocessing: Get the original image of the throat from the micro camera. Remove noise and interference in the image, such as uneven lighting and blurred images. Use filtering algorithms (such as mean filtering and Gaussian filtering) to smooth the image. Improve the contrast and clarity of the image to better identify sputum features. Use grayscale transformation, image sharpening and other technologies to enhance the image.

[0222] 2. Feature extraction: Identify features related to sputum in the image, such as color, texture, shape, etc. Use shape feature extraction algorithms (such as contour detection algorithms) to extract the contours and boundaries of sputum.

[0223] 3. Throat structure identification: Identify the anatomical structure of the throat, such as the vocal cords, epiglottis, etc. Ensure that the normal structure of the throat is not misidentified as sputum during the analysis process.

[0224] 4. Sputum detection: Based on the extracted sputum features, determine whether there is sputum in the image. Use image recognition algorithms (such as support vector machine SVM, convolutional neural network CNN) to classify and identify sputum.

[0225] 5. Sputum volume assessment: Further evaluate the quantity and distribution of sputum. Quantify the amount of sputum by measuring indicators such as sputum area, volume or density.

[0226] 6. Risk assessment: Based on the amount and distribution of sputum, determine whether the sputum poses a threat to breathing.

[0227] If the amount of sputum is large or the distribution location is critical (such as blocking the airway), it is judged as high risk.

[0228] 7. Result output: Output the analysis results in the form of images, text or numbers for medical staff to review. Clearly mark the location, quantity and risk level of sputum in the analysis results.

[0229] In this technical solution, the image analysis system can accurately analyze the collected throat images, determine whether there is sputum blockage, and trigger an alarm in time, providing strong support for the subsequent treatment of medical staff.

[0230] The working principle of the oxygenated atomizer inhaler for treating interstitial lung disease of the present invention is as follows: according to the specific conditions of the interstitial lung disease patient, a mouthpiece 14 or a medicine mist supply pipe 16 is selected to perform oxygenated atomized inhalation treatment on the patient. The breathing mask assembly 1 is fixedly worn on the patient's face; the oxygen cylinder 2 continuously delivers oxygen to the medicine liquid tank 3, and the atomizer pump 5 atomizes the medicine liquid in the medicine liquid tank 3 and delivers it to the mouthpiece 14 or the medicine mist supply pipe 16 for the patient to inhale the treatment. If the mouthpiece 14 is adopted, multiple exhaust pipes 15 are made to work, and if the medicine mist supply pipe 16 is adopted, one exhaust pipe 15 is made to work; at the same time, the sound collection module collects the sound data in the ward in real time; the breathing monitoring module monitors the patient's breathing condition; the heart rate monitoring module monitors the patient's heart rate; the blood oxygen saturation monitor monitors the patient's blood oxygen saturation; the image collection module collects high-definition images of the patient; and then the image preprocessing module preprocesses the collected images, including denoising, grayscale, image enhancement and normalization, etc., to eliminate interference factors in the image. The feature extraction module extracts features from the preprocessed image, and the extracted features include color, texture and shape; the abnormal analysis module uses machine learning or deep learning algorithms to analyze and identify the images after feature extraction, and promptly identify abnormal situations; the comprehensive evaluation module combines the monitoring results of blood oxygen saturation, heart rate, respiratory rate and abnormal analysis module, and integrates and analyzes the data of multiple monitoring indicators through algorithms to obtain comprehensive evaluation results, conduct a comprehensive evaluation of patients, and promptly discover abnormal situations and potential risks; when the patient is monitored to have breathing difficulties, the respiratory optimization management module controls the atomization pump 5, the drug mist flow valve A51 and the drug mist flow valve B52 to stop working and stop delivering medicine; the oxygen flow valve B22 is opened, and oxygen is delivered to the nostrils through the oxygen delivery nasal suction tube 12 for the patient to breathe, which helps to improve the patient's breathing condition. At the same time, an alarm is triggered. At the same time, the micro motor works, driving the two gears 71 to rotate synchronously in the opposite direction, causing the two rollers 72 to rotate synchronously in the opposite direction, pushing the sputum suction hose 13 forward to the throat; the micro camera collects the throat image to analyze whether the throat is blocked by sputum; if the throat is blocked by sputum, the negative pressure suction pump 6 is started to suck the sputum into the waste liquid collection box 4, and the sputum is cleaned up as soon as possible to restore normal breathing of the patient. When abnormal conditions and potential risks are detected, the alarm module issues an alarm in time.

[0231] The present invention can select a mouthpiece or a medicine mist supply tube for oxygenated atomization inhalation therapy according to the specific situation of the patient to meet the needs of personalized treatment. Real-time monitoring of the patient's breathing, heart rate, blood oxygen saturation and other physiological indicators, providing doctors with comprehensive patient status information. Through image acquisition, preprocessing, feature extraction and abnormal analysis, abnormal conditions of patients can be discovered in time, improving the accuracy and timeliness of diagnosis. The data of multiple monitoring indicators are integrated and analyzed to obtain a comprehensive evaluation result, and potential risks are discovered in time to provide decision support for doctors. When it is monitored that the patient has difficulty breathing, the drug delivery is automatically stopped and the oxygen delivery is turned on, which helps to improve the patient's breathing condition and reflects the intelligent management of the equipment. Sputum can be cleaned in time, and the suction hose is pushed to the throat by a micro motor and a roller, and the sputum is cleaned in time by a negative pressure suction pump, avoiding respiratory disorders caused by sputum blockage and ensuring the safety of patients.

[0232] As described above, the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some of the technical features thereof may be replaced by equivalents. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. An oxygenating nebulizer for treating interstitial lung disease, comprising a breathing mask assembly, an oxygen cylinder, a liquid medicine tank, a waste liquid collection tank, an atomizing pump, a negative pressure suction pump and a monitoring mechanism, characterized in that: The breathing mask assembly is connected to the oxygen cylinder through a hose; the liquid medicine tank is connected to the oxygen cylinder through a hose; the oxygen cylinder has a built-in pressure sensor; A medicine mist supply pipe is fixedly arranged on the breathing mask assembly; a mouthpiece is sealably and slidably arranged on the breathing mask assembly; An atomizing pump is fixedly arranged on the medicine liquid tank; the atomizing pump is connected to the medicine mist supply pipe and the mouthpiece through a hose; A negative pressure suction pump is fixedly arranged on the waste liquid collection box; a sputum suction hose is slidably arranged on the breathing mask assembly; the sputum suction hose is connected to the input end of the negative pressure suction pump; and the output end of the negative pressure suction pump is connected to the waste liquid collection box; A monitoring mechanism is fixedly installed on the medicine tank to monitor the oxygen nebulization inhalation treatment process of patients with interstitial lung disease and promptly detect abnormal conditions and potential risks.

2. The oxygenating nebulizer for treating interstitial lung disease according to claim 1, characterized in that: The breathing mask assembly comprises a mask body, an oxygen supply nasal suction tube, a sputum suction hose, a mouthpiece, an exhaust pipe and a medicine mist supply pipe; An oxygen supply nasal suction tube is fixedly arranged on the mask body; two branch tubes are fixedly arranged on the oxygen supply nasal suction tube; A sputum suction hose is slidably arranged on the mask body; A mouthpiece is slidably provided on the mask body; the atomized medicine can be delivered to the mouth through the mouthpiece; The cover body is fixedly provided with a plurality of exhaust pipes; each exhaust pipe is provided with an electromagnetic valve; and the cover body is fixedly provided with a medicine mist supply pipe.

3. The oxygenating nebulizer for treating interstitial lung disease according to claim 2, characterized in that: The oxygen supply nasal suction tube is provided with an oxygen flow valve B; the hose between the oxygen cylinder and the medicine tank is provided with an oxygen flow valve A; A medicine mist flow valve B is provided on the hose connected to the medicine mist supply pipe and the atomizing pump; a medicine mist flow valve A is provided on the hose connected to the mouthpiece and the atomizing pump; Two gears are rotatably arranged on the cover body, and the two gears are meshed and transmission connected; rollers are coaxially fixedly arranged on the gears; the rollers are arranged in friction contact with the suction hose; a micro motor is fixedly arranged on the cover body, and the output end of the micro motor is coaxially fixedly connected to one of the gears.

4. The oxygenating nebulizer for treating interstitial lung disease according to claim 1, characterized in that: The monitoring agencies include: Data collection module: collects patient data, including medical history, case history, and image data during treatment, and annotates the images as reference samples; Sound collection module: collects sound data in the ward in real time; Respiratory monitoring module: including respiratory rate sensor to monitor the patient's respiratory condition; Heart rate monitoring module: including a heart rate monitoring belt to monitor the patient's heart rate; Blood oxygen saturation monitor: monitors the patient's blood oxygen saturation; Image acquisition module: including a high-definition camera to collect high-definition images of patients; Image preprocessing module: preprocess the collected images, including denoising, grayscale, image enhancement and normalization; Feature extraction module: extracts features from the preprocessed image. The extracted features include color, texture and shape. Abnormal analysis module: analyzes and identifies the image after feature extraction to identify abnormal situations; Comprehensive assessment module: Combines the monitoring results of blood oxygen saturation, heart rate, respiratory rate and abnormal analysis modules to conduct a comprehensive assessment of the patient and promptly identify abnormal conditions and potential risks; Breathing optimization management module: controls the operation of the atomizer pump, drug mist flow valve A, drug mist flow valve B, micro motor and oxygen flow valve B; when the patient is detected to have breathing difficulties, stops delivering medicine and starts delivering oxygen; analyzes whether there is sputum blockage in the throat. If there is sputum blockage, the sputum is sucked into the waste liquid collection box through a negative pressure suction pump to restore the patient's normal breathing; Alarm module: including alarm, which will send out alarm in time when abnormal situation and potential risk are detected; PLC control module: connected to the data collection module, respiratory monitoring module, heart rate monitoring module, blood oxygen saturation monitor, image acquisition module, image preprocessing module, feature extraction module, abnormal analysis module, comprehensive evaluation module and alarm module network.

5. The oxygenating nebulizer for treating interstitial lung disease according to claim 4, characterized in that: The abnormal analysis module analyzes and identifies the image after feature extraction, including the following: Data preparation: Obtain high-definition images of patients from the image acquisition module, annotate the images, clearly mark various abnormalities, and form annotated data sets; Model training: Select the convolutional neural network (CNN) model and use the labeled data set for training so that the model can learn the association between features in the image and abnormal conditions; continuously optimize the performance of the model by adjusting the model parameters, increasing the amount of training data, and using data enhancement techniques; Abnormal identification: Use the feature extraction module to extract key features of color, texture, and shape from the image; input the extracted features into the trained model to perform abnormal identification, including dyspnea analysis, cough aggravation analysis, chest tightness and pain analysis, and allergic reaction analysis; Output result: The model outputs the recognition result based on the input features, determines whether there is an abnormality in the image, and gives the specific type of abnormality; Alarm feedback: Set the threshold for abnormal recognition according to actual needs; when the recognition result exceeds the threshold, an alarm is triggered; Continuous monitoring and updating: During oxygen nebulizer treatment, the anomaly analysis module is used to continuously monitor and analyze images; as the treatment progresses and the patient's condition changes, the model is continuously updated and optimized to adapt to new anomalies and data features.

6. The oxygenating nebulizer for treating interstitial lung disease according to claim 5, characterized in that: The cough exacerbation analysis includes the following: Cough frequency analysis: Count the number of cough sound clips to calculate the cough frequency; Cough intensity analysis: Analyze the amplitude or volume of the cough sound to assess the intensity of the cough; Cough duration analysis: The total duration of all cough sound segments was calculated to assess the duration of the cough; Body posture analysis: Analyze the body posture of the person when coughing, including the shaking of the head, bending or straightening of the body; Facial expression analysis: Analyze the facial expression of a person when coughing, including whether it is painful, tired or anxious; Comprehensive judgment: Combine the results of sound analysis and observation of body posture and facial expressions to comprehensively judge the severity of the cough.

7. The oxygenating nebulizer for treating interstitial lung disease according to claim 5, characterized in that: The chest tightness and chest pain analysis includes the following contents: Feature extraction: Use image processing techniques to extract features related to facial expressions and posture; This includes the frowning of eyebrows, the squinting of eyes, the opening and closing of mouth, the distortion of face, body posture and movement; Facial expression analysis: Use convolutional neural network (CNN) to classify and recognize the extracted facial expression features to determine whether there is pain expression; By analyzing subtle changes in facial muscles, determine whether there is an uncomfortable expression; Posture analysis: Using AI algorithms to identify the postures of people in images, especially those associated with pain and discomfort; analyzing whether the person's movements are smooth and natural, or whether they are stiff or slow; Inference of pain location: Infer the possible location of pain based on facial expressions and body posture characteristics combined with medical knowledge; Possibility assessment of chest tightness and chest pain: Comprehensive facial expressions, body characteristics and possible pain locations are used to assess the possibility of chest tightness and chest pain; the chest tightness and chest pain assessment model is: P chest_pain =w1*E pain +w2*E discomfort +w3*P pain_related +w4*A abnormal +w5*L pain_correlation +b; P chest_pain represents the evaluation value of the possibility of chest tightness or chest pain; w1, w2, w3, w4, and w5 are the weights of each feature; E pain is the recognition result of the painful expression; E discomfort is the analysis result of uncomfortable expression; P pain_related is the result of pain-related posture recognition; A abnormal It is the analysis result of abnormal action; L pain_correlation is the inference result of the correlation between pain location and chest tightness and chest pain; b is the bias term; Auxiliary diagnosis and suggestions: The analysis results are provided to doctors as auxiliary diagnosis information to help doctors judge the patient's condition more accurately.

8. The oxygenating nebulizer for treating interstitial lung disease according to claim 5, characterized in that: The allergic reaction analysis includes the following: Feature extraction: Analyze the color distribution of skin images and pay attention to whether there is abnormal redness, swelling or discoloration; use image processing technology to extract skin texture features and detect whether there is rash or abnormal changes in skin texture; observe morphological changes on the skin surface; Model training: Build an allergic reaction recognition model; build a dataset containing known allergic reactions and normal skin images; train the dataset so that the model can learn and recognize the characteristics of allergic reactions; the allergic reaction recognition model is: i ★ =argmin θ {∑ N i＝1 -[y i log(σ(W T x i +d))+(1-y i )log(1-σ(W T x i +d))]};expression In, θ ★ Represents the optimal solution of model parameters; argmin θ It means to find the parameter set that minimizes the objective function in the parameter space θ; i represents the true label or target value of the i-th sample in the dataset; σ represents the Sigmoid activation function; W T represents the transpose of the weight matrix W; x i represents the feature vector of the i-th sample in the data set; d represents the bias term; log represents the natural logarithm function; Classification and recognition: The pre-processed skin images are input into the trained model for classification and recognition to determine whether there is an allergic reaction; Result judgment: Based on the output of the model, determine whether there is an allergic reaction in the image; Verification and optimization: Use an independent validation data set to evaluate the performance of the model to ensure the accuracy and reliability of the model; adjust and optimize the model based on the validation results.

9. The oxygenating nebulizer for treating interstitial lung disease according to claim 5, characterized in that: The comprehensive assessment module combines the monitoring results of blood oxygen saturation, heart rate, respiratory rate and abnormal analysis module to conduct a comprehensive assessment of the patient and promptly identify abnormal conditions and potential risks; it includes the following: Data collection: real-time collection of blood oxygen saturation data, heart rate, respiratory rate and abnormal analysis module data; Data preprocessing: including data cleaning, data standardization and data alignment; Data fusion: Extract key features from the preprocessed data, including the average value of blood oxygen saturation, the rate of change of heart rate, and the stability of respiratory rate; Feature fusion: The features of multiple physiological parameters are fused using a weighted summation algorithm to form a comprehensive feature vector; feature fusion is performed according to the following formula: Where, F represents the comprehensive characteristic value; n represents the total number of physiological parameter characteristics; u i represents the weight of the ith physiological parameter feature; β is an exponent for adjusting the weight; max(f) and min(f) represent the maximum and minimum values ​​of all physiological parameter feature values, respectively; f i Represents the original value of the i-th physiological parameter feature; α i Represents the nonlinear transformation index of the i-th physiological parameter characteristics; Model training: Based on historical data, a support vector machine comprehensive evaluation model is trained, which can predict the patient's health status or risk level based on the comprehensive feature vector; the model is: In the formula, y is the dependent variable; β0 is the intercept term; β i is the linear coefficient of the i-th feature; φ(F i ) is the i-th feature F i The nonlinear transformation function performed; γ ij It is feature F i The coefficient of the interaction term between and Fj; δ is the coefficient of the square term summed after all features are nonlinearly transformed; ε is the error term; the data collected in real time are input into the comprehensive evaluation model to obtain the comprehensive evaluation results of the current patient; Result output and anomaly detection: Display the comprehensive evaluation results to users or medical staff in an intuitive way; set thresholds or rules to trigger abnormal alarms when the comprehensive evaluation results exceed a certain range; Feedback and optimization: Collect feedback from medical staff and patients to understand the accuracy and practicality of the comprehensive assessment module; based on feedback and new data, continuously optimize the comprehensive assessment model to improve the accuracy and reliability of the assessment.

10. The oxygenating nebulizer for treating interstitial lung disease according to claim 1, characterized in that: The breathing optimization management module analyzes whether there is phlegm blockage in the throat, including the following: Image preprocessing: obtain the original image of the throat from the micro camera; remove noise and interference in the image; Feature extraction: Identify features related to sputum in the image, including color, texture, and shape; extract the contours and boundaries of sputum; Throat structure identification: Identify the anatomy of the throat, including the vocal cords and epiglottis; ensure that normal structures of the throat are not mistaken for sputum during analysis; Sputum detection: Based on the extracted sputum features, determine whether there is sputum in the image; use the support vector machine (SVM) image recognition algorithm to classify and identify sputum; Sputum volume assessment: Assess the amount and distribution of sputum; quantify the amount of sputum by measuring the area, volume or density of sputum; Risk assessment: Based on the amount and distribution of sputum, determine whether the sputum poses a threat to breathing; Result output: The analysis results are output in the form of images, text or numbers for medical staff to review, and the location, quantity and risk level of sputum are clearly marked in the analysis results.

Images