Supervision auxiliary system for children to take medicine

Abstract

The invention relates to the technical field of medicine taking supervision and assistance, and provides a supervision and assistance system for medicine taking of children, and the system comprises a plurality of modules which work cooperatively. The medicine distribution module distributes medicine according to a preset administration scheme, and the medicine taking monitoring module monitors the medicine taking behavior of children in real time through a camera and analyzes captured images through an image recognition algorithm. And the medicine taking analysis module receives and processes the monitoring data, and judges the medicine taking success rate of the child through the evaluation model. And the feedback module generates feedback information according to the analysis result to guide subsequent operation. The medicine taking optimization module continuously optimizes a medicine administration scheme according to historical data, and a higher medicine taking success rate is ensured. In addition, the system further comprises a data processing unit, a model building unit and a prediction unit, and by analyzing historical medicine taking data of children, a gradient boosting decision tree model is used for predicting and identifying reasons causing the medicine taking success rate to be lower than a set threshold value.

Description

Technical Field

[0001] The present invention relates to the technical field of medication-assisting supervision, and more particularly to a supervision-assisting system for children taking medication. Background Art

[0002] With the advancement of modern medical technology, the treatment effect of children's diseases has been significantly improved. However, compared with adults, children face many unique difficulties in the process of taking medicine, which not only affects the treatment effect, but also may bring potential health risks. First, children's cognitive ability and self-control have not yet fully developed, and it is difficult for them to understand the importance of taking medicine on time and in the right amount; second, many children are very sensitive to the taste and texture of medicine, and often resist taking medicine; in addition, children's physiological characteristics also bring difficulties to taking medicine, such as weak swallowing ability, easy to choke or spit medicine. Due to their young age, poor compliance, and inability to adapt to the taste of medicine or the process of taking medicine, child patients often refuse to take medicine, do not take medicine on time, or take medicine irregularly. These problems not only affect the treatment effect, but may also lead to waste of medicine and even cause potential health risks.

[0003] Existing children's medication management mainly relies on the experience and manual supervision of parents or caregivers. However, this traditional supervision method has multiple defects. First, manual supervision is subjective and easily affected by the inattention of parents or caregivers, resulting in the inability to timely discover and correct children's irregular medication behavior. Busy parents may find it difficult to ensure that their children take medicine on time every time, or cannot accurately judge whether the children have actually swallowed the medicine. Some parents may compromise because of their children's crying, resulting in missed doses or insufficient doses. More seriously, some parents may try to cover up the taste of the medicine with food or drinks, which may change the effect of the medicine or cause inaccurate doses. Secondly, traditional supervision methods cannot provide accurate data records and feedback, and it is difficult to provide reliable reference information for subsequent medical decisions. In addition, parents' verbal reminders or the use of paper record sheets have limited effects and lack objectivity. With the development of technology, some electronic reminder devices have been introduced into children's medication management, but these devices can usually only provide simple time reminders and cannot truly monitor and analyze children's medication behavior. In addition, when multiple children take medicine at the same time, the efficiency and accuracy of supervision are even more difficult to guarantee.

[0004] In order to solve the above problems, a technical solution is now provided. Summary of the invention

[0005] In order to overcome the above-mentioned defects of the prior art, the present invention provides a supervision assistance system for children's medication, which includes a drug dispensing module, a medication monitoring module, a medication analysis module, a medication optimization module and a feedback module; the system monitors children's medication behavior in real time and provides accurate data records and feedback through camera monitoring, image recognition algorithm, data analysis and feedback optimization, thereby improving the standardization and success rate of children's medication, and solving the problems of strong subjectivity, inaccurate data and low efficiency of manual supervision.

[0006] To achieve the above object, the present invention provides the following technical solutions:

[0007] A supervision auxiliary system for children's medication, comprising a medication distribution module, a medication monitoring module, a medication analysis module, a medication optimization module and a feedback module, wherein the medication distribution module is connected to the medication monitoring module, the medication monitoring module is connected to the medication analysis module, the medication analysis module is connected to the medication optimization module, the medication optimization module is connected to the feedback module, and the feedback module is connected to the medication analysis module, wherein the medication distribution module is used to distribute medication according to a preset medication regimen; the medication monitoring module is used to monitor children's medication behavior in real time, including monitoring children's medication actions through a camera and analyzing captured images through an image recognition algorithm; the medication analysis module is used to receive and process monitoring data transmitted by the medication monitoring module, input a constructed medication success rate evaluation model, and judge children's medication success rate; the feedback module is used to generate feedback information according to the judgment result of the data processing module; the medication optimization module is used to optimize the medication regimen according to historical medication data; wherein the medication analysis module calculates a medication success rate evaluation value according to the medication success rate evaluation model, wherein the formula of the medication success rate evaluation model is:

[0008]

[0009] Where S is the evaluation value of the medication success rate, D is the drug distribution dose, A is the evaluation value of oral cooperation, T is the duration of medication, P is the pixel area occupied by the drug in the image, and t is the total time of the medication process.

[0010] As a further solution of the present invention, the medication monitoring module includes a camera, an image processing unit and an image recognition unit. The camera is used to capture the child's medication actions; the image processing unit is used to perform real-time denoising, enhancement and stabilization processing on the image captured by the camera.

[0011] As a further solution of the present invention, the image recognition unit is based on a deep learning framework, and analyzes the preprocessed image through a multi-layer convolutional neural network to identify the child's medication behavior and output medication characteristic parameters. The medication characteristic parameters include the child's mouth opening amplitude, medication duration, pixel area occupied by the drug in the image, and the total time of the medication process.

[0012] As a further solution of the present invention, the monitoring data transmitted by the medication analysis module and the medication monitoring module are input into the collected children's mouth opening amplitude, swallowing times and emotional index into the oral cooperation evaluation model to evaluate the oral cooperation degree of the children when taking the medicine, wherein the formula of the oral cooperation evaluation model is:

[0013]

[0014] In the formula, A is the oral cooperation assessment value, O is the child's mouth opening width, C is the number of swallowing times, and EI is the emotional index.

[0015] As a further scheme of the present invention, the medication optimization module includes a data processing unit, a model building unit and a prediction unit, wherein the data processing unit is used to collect and process each child's historical medication data group, wherein the medication data group includes the medication success rate value, the child's mouth opening amplitude, the duration of medication, the pixel area occupied by the drug in the image and the total time of the medication process, the medication time, the emotional index, the child's age, weight, past medication compliance records, drug type, drug dosage, drug taste, number of swallowings, and medication location, and performs feature extraction, cleaning and standardization on the data, and divides the processed data into a training set and a test set; the prediction unit is used to receive the child's real-time medication data group, input the trained gradient boosting decision tree model, and output the predicted reasons for the medication success rate being lower than the set threshold.

[0016] As a further solution of the present invention, the model building unit uses the training set data to train the gradient boosting decision tree model, and annotates the reasons for the medication data groups in the training set whose medication success rate is lower than the set threshold. The medication success rate is used as the target variable and the medication data group is used as the feature variable. The trained gradient boosting decision tree model is tested using the test set data.

[0017] As a further solution of the present invention, the feedback module includes an information generating unit, a display unit and an alarm unit, wherein the information generating unit is used to generate feedback information according to the judgment result of the medication analysis module, and the feedback information includes the result of whether the medication is successful, the accuracy of the medication time and the correctness of the medication dosage; the display unit is used to display the feedback information and the reasons output by the medication optimization module that cause the medication success rate to be lower than the set threshold, and the alarm unit is used to trigger an alarm signal when the judgment result is that the medication success rate is lower than the set threshold.

[0018] As a further solution of the present invention, the result of whether the medication is successful is obtained by comparing the medication success rate with a preset threshold; the accuracy of the medication time is judged by comparing the actual medication duration with the preset medication time; the correctness of the medication dosage is judged by comparing the actual dispensed dosage with the prescribed dosage.

[0019] Compared with the prior art, the beneficial effects of the monitoring auxiliary system for children taking medicine of the present invention are:

[0020] The present invention realizes real-time monitoring of children's medication behavior by integrating computer vision technology and deep learning algorithms, thereby improving the accuracy and objectivity of supervision; secondly, the medication success rate evaluation model and the oral cooperation evaluation model comprehensively consider factors such as drug distribution dosage, medication duration, and oral cooperation degree, and can comprehensively and scientifically evaluate children's medication situation; thirdly, the medication optimization module adopts an advanced gradient boosting decision tree model, which can intelligently predict the reasons for the low medication success rate based on historical medication data, providing a scientific basis for formulating personalized improvement strategies. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] Figure 1 The present invention is a schematic structural diagram of a monitoring auxiliary system for children taking medicine. DETAILED DESCRIPTION

[0022] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0023] Example 1

[0024] A supervision auxiliary system for children's medication includes a medication distribution module, a medication monitoring module, a medication analysis module, a medication optimization module and a feedback module.

[0025] The drug dispensing module in the embodiment of the present invention is connected to the medication monitoring module, the medication monitoring module is connected to the medication analysis module, the medication analysis module is connected to the medication optimization module, the medication optimization module is connected to the feedback module, and the feedback module is connected to the medication analysis module.

[0026] The drug dispensing module is used to dispense drugs according to a preset dosage regimen, and the drug dispensing module includes a drug storage bin, a dispenser and a control unit, wherein the drug storage bin is used to store drugs; the dispenser is used to control the output of drugs; and the control unit controls the dispenser based on the preset dosage regimen.

[0027] The drug storage bin in the embodiment of the present invention is made of transparent material, has good sealing and visibility, has a capacity of 500ml, and is provided with a detachable cover on the top for easy addition of drugs; the dispenser adopts a spiral conveying mechanism driven by a micro-stepping motor, which can accurately control the amount of drugs output, and the minimum distribution accuracy reaches 0.1ml; the control unit adopts an ARM Cortex-M4 microcontroller, with a built-in real-time clock and memory, which can store multiple groups of drug administration plans; the drug administration plan includes information such as drug type, dosage, frequency and time. When working, the control unit triggers the drug administration program according to the built-in clock, sends a control signal to the stepper motor, drives the spiral conveying mechanism to rotate a predetermined number of circles, and delivers the precise dose of drugs to the drug outlet. At the same time, the control unit communicates with the main system through the built-in Bluetooth to receive updated drug administration plans or send distribution records. The module is also equipped with a weight sensor to monitor the remaining amount of drugs in the drug storage bin, and automatically issues a replenishment reminder when it is lower than the preset threshold.

[0028] The drug dispensing module in the embodiment of the present invention realizes the automation and precision of drug dispensing, significantly reduces human errors, and ensures the accuracy of each drug dosage. At the same time, through the function of remotely updating the drug administration plan, the flexibility of medication management is improved, allowing doctors to adjust the medication strategy at any time according to the actual situation of the child; in addition, the built-in drug dosage monitoring and reminder functions effectively avoid the situation of drug exhaustion and ensure the continuity of treatment. Overall, the drug dispensing module not only improves the accuracy and convenience of drug administration, but also provides a reliable drug supply basis for the entire children's medication auxiliary supervision system through its intelligent and humanized design, effectively improving the safety and compliance of children's medication.

[0029] The medication monitoring module is used to monitor children's medication behavior in real time, including monitoring children's medication actions through a camera and analyzing the captured images through an image recognition algorithm.

[0030] The medication monitoring module in the embodiment of the present invention includes a camera, an image processing unit and an image recognition unit. The camera is used to capture the child's medication action; the image processing unit is used to perform real-time denoising, enhancement and stabilization processing on the image captured by the camera.

[0031] The image recognition unit in the embodiment of the present invention is based on a deep learning framework, and analyzes the preprocessed image through a multi-layer convolutional neural network to identify the child's medication behavior and output medication characteristic parameters, wherein the medication characteristic parameters include the child's mouth opening amplitude, medication duration, pixel area occupied by the drug in the image, and the total medication process time.

[0032] The medication monitoring module in the embodiment of the present invention uses a high-definition camera to capture the child's medication action in real time. The camera is installed in an appropriate position to ensure that the child's face and medication process can be clearly captured. The image processing unit receives the video stream transmitted by the camera and first performs preprocessing, including noise reduction, contrast enhancement and image stabilization; these processes are implemented through common image processing algorithms, such as Gaussian filtering denoising, histogram equalization contrast enhancement, and optical flow method for image stabilization. The processed image is passed to the image recognition unit, which uses a pre-trained convolutional neural network model. The model has been fine-tuned through transfer learning using a large number of annotated children's medication images, and can effectively identify medication-related features. The model output includes the child's mouth opening amplitude, medication duration, the area occupied by the drug in the image, and the total time of the entire medication process; it can objectively and accurately record the child's medication behavior, provide detailed medication parameters, and provide reliable data support for evaluating medication conditions and optimizing medication regimens.

[0033] The steps of training the convolutional neural network model in the embodiment of the present invention include:

[0034] Step 1: Collect high-definition images of children taking medicine. The image resolution should be no less than 1280x720 pixels. These images should include children of different ages from 2 to 12 years old, covering various ways of taking medicine such as oral liquid, tablets, and capsules. At the same time, ensure that the collected images cover different lighting conditions such as natural light, indoor lighting, and diverse shooting angles to increase the diversity of the data. Finally, all collected images are randomly divided into training set, validation set, and test set in a ratio of 8:1:1 to prepare for subsequent model training and evaluation;

[0035] Step 2: Use the open source annotation tool LabelImg to annotate the collected images. Annotate the position of the child's face for each image with a rectangular frame; annotate the mouth opening state, divided into open or closed; annotate the position of the medicine, also marked with a rectangular frame; finally, annotate the medication action, including the three stages of start, in progress or end; the purpose of annotation is to provide a clear learning goal for the model, save all annotation information as XML format files, and ensure that each XML file corresponds to the corresponding image file one by one;

[0036] Step 3: Perform data enhancement on the original image. The enhancement methods include random rotation, with a rotation range of negative 15 degrees to 15 degrees; horizontal flipping, with a flipping probability of 50%; brightness adjustment, with an adjustment range of 80% to 120% of the original brightness; and contrast adjustment, with an adjustment range of 80% to 120% of the original contrast. Five enhanced images are generated for each original image;

[0037] Step 4: Use the TensorFlow deep learning framework and select the pre-trained ResNet-50 model as the base network. ResNet-50 is a deep convolutional neural network that performs well in large-scale image recognition tasks. Download the pre-trained ResNet-50 model weights from TensorFlow Hub. These pre-trained weights contain features learned from a large number of general images.

[0038] Step 5: Keep the convolutional layer of ResNet-50 as the feature extractor, remove the original fully connected layer, and add a new global average pooling layer; on this basis, add two new fully connected layers: the first layer contains 1024 neurons and uses the ReLU activation function; the second layer is used as the output layer, which contains 4 neurons, which are used to predict the mouth opening amplitude, medication duration, drug area ratio and total medication time respectively;

[0039] Step 6: Use the transfer learning strategy to freeze the weights of the first 100 layers of ResNet-50 and keep the parameters of these layers unchanged. Set the remaining convolutional layers and the newly added fully connected layers to the trainable state;

[0040] Step 7: Set the training parameters, including batch size of 32, initial learning rate of 0.0001, use Adam optimizer, and select mean square error as loss function; write a training loop, which includes loading batch data, forward propagation, calculating loss, back propagation, and updating weights; evaluate the model performance on the validation set after each training epoch; set the training to a total of 50 epochs, or stop training early when the validation set loss has not improved for 5 consecutive epochs to prevent overfitting;

[0041] Step 8: Evaluate the performance of the trained model on the test set. Calculate the mean absolute error, root mean square error, and R 2 The performance of the model on each prediction task is comprehensively evaluated by using indicators such as scores. At the same time, a scatter plot of the predicted value and the true value is drawn to intuitively evaluate the model performance and observe whether there are systematic deviations or outliers in the prediction results. These evaluation indicators and visualization results will help us fully understand the performance of the model and provide a basis for subsequent optimization.

[0042] The medication analysis module is used to receive and process the monitoring data transmitted by the medication monitoring module, input the constructed medication success rate evaluation model, and determine the medication success rate of children.

[0043] The monitoring data transmitted by the medication analysis module and the medication monitoring module in the embodiment of the present invention are input into the collected child's mouth opening amplitude, swallowing frequency and emotional index into the oral cooperation evaluation model to evaluate the oral cooperation degree of the child when taking the medicine, wherein the formula of the oral cooperation evaluation model is:

[0044]

[0045] In the formula, A is the oral cooperation assessment value, O is the child's mouth opening width, C is the number of swallowing times, and EI is the emotional index.

[0046] The emotion index in the embodiment of the present invention is obtained by facial expression recognition technology, which uses a deep learning model to analyze the facial features of children, including the position and shape changes of eyebrows, eyes and mouth, so as to infer the emotional state of children. The oral coordination evaluation model is an oral coordination evaluation model that comprehensively considers multiple factors. Among them, ln(1+O) reflects the effect of mouth opening amplitude on coordination. The use of a logarithmic function can provide a more sensitive response when the mouth opening amplitude is small, and the growth tends to be slow when the mouth opening amplitude is large; represents the square root of the number of swallowing times. This design is because the effect of increasing the number of swallowing times on the degree of cooperation is decreasing, and the square root function can capture this relationship well; The term considers the impact of emotional factors on compliance, where e^{EI} reflects the exponential impact of the emotional index on compliance, and the integral considers the cumulative effect of emotions during the entire medication process. Finally, taking the cube root of the entire expression is to adjust the weight of each factor so that the final compliance evaluation value is within a reasonable range.

[0047] The medication analysis module in the embodiment of the present invention calculates the medication success rate evaluation value according to the medication success rate evaluation model, wherein the formula of the medication success rate evaluation model is:

[0048]

[0049] Where S is the evaluation value of the medication success rate, D is the drug distribution dose, A is the evaluation value of oral cooperation, T is the duration of medication, P is the pixel area occupied by the drug in the image, and t is the total time of the medication process.

[0050] In the medication success rate evaluation model formula in the embodiment of the present invention, ln(1+D) reflects the influence of the drug distribution dose. The use of a logarithmic function can provide a more sensitive response when the dose is small, while the growth tends to be flat when the dose is large; (1-e -T ) item takes into account the effect of duration of medication. As time increases, the value of this item will gradually approach 1; The effect of oral coordination is introduced. The square root is used because the influence of coordination on the success rate is increasing; the integral term in the denominator It reflects the cumulative effect of the change in drug area during the entire medication process, among which Represents the absolute value of the rate of change of the drug area. This design takes into account the movement and disappearance of the drug in the mouth. Combining these factors together, the formula is able to capture the dynamic changes in the medication process while balancing the importance of each factor. For example, even if the dose is appropriate and the oral compliance is high, if the drug area changes dramatically, it may mean that the drug is spit out and the success rate will be reduced.

[0051] The medication optimization module includes a data processing unit, a model building unit and a prediction unit, wherein the data processing unit is used to collect and process each child's historical medication data group, wherein the medication data group includes the medication success rate value, the child's mouth opening amplitude, the duration of medication, the pixel area occupied by the drug in the image and the total time of the medication process, the medication time, the emotional index, the child's age, weight, past medication compliance records, drug type, drug dosage, drug taste, number of swallowings, and medication location, and performs feature extraction, cleaning and standardization on the data, and divides the processed data into a training set and a test set; the prediction unit is used to receive the child's real-time medication data group, input the trained gradient boosting decision tree model, and predict the reasons why the medication success rate is lower than the set threshold.

[0052] The model building unit in the embodiment of the present invention uses the training set data to train the gradient boosting decision tree model, and annotates the reasons for the medication data groups in the training set whose medication success rate is lower than the set threshold. The medication success rate is used as the target variable and the medication data group is used as the feature variable. The trained gradient boosting decision tree model is tested using the test set data.

[0053] The following is a Python code example that uses a gradient boosting decision tree model to predict the reasons why the medication success rate is lower than the set threshold. Please note that this example is only a starting point and may need to be adjusted according to actual conditions and device interfaces in actual applications.

[0054] import pandas as pd

[0055] from sklearn.model_selection import train_test_split

[0056] from sklearn.preprocessing import StandardScaler

[0057] from sklearn.ensemble import GradientBoostingClassifier

[0058] from sklearn.metrics import accuracy_score,classification_report

[0059] #Assume that the dataset is stored in DataFrame format

[0060] #The data set includes: medication success rate, mouth opening amplitude, medication duration, medication pixel area, total time, medication time, emotion index, age, weight, compliance record, medication type, dosage, taste, number of swallows, medication location

[0061] #Data loading and processing

[0062] data=pd.read_csv('medication_data.csv')

[0063] features = ['mouth opening width', 'drug taking duration', 'drug pixel area', 'total time', 'drug taking time', 'emotional index', 'age', 'weight', 'compliance record', 'drug type', 'dose', 'taste', 'swallowing times', 'drug taking location']

[0064] target = "Medication success rate"

[0065] #Data cleaning and standardization

[0066] X = data[features]

[0067] y=data[target]

[0068] scaler = StandardScaler()

[0069] X_scaled=scaler.fit_transform(X)

[0070] # Divide into training set and test set

[0071] X_train,X_test,y_train,y_test=train_test_split(X_scaled,y,test_size=0.2,random_state=42)

[0072] #Training gradient boosting decision tree model

[0073] model=GradientBoostingClassifier()

[0074] model.fit(X_train,y_train)

[0075] #Prediction and evaluation

[0076] y_pred=model.predict(X_test)

[0077] print("Accuracy:",accuracy_score(y_test,y_pred))

[0078] print(classification_report(y_test,y_pred))

[0079] #Model predicts the reasons for the low success rate of medication

[0080] #Assume that the new real-time medication data set is stored in the new_data variable

[0081] new_data=pd.read_csv('new_medication_data.csv')

[0082] new_data_scaled=scaler.transform(new_data[features])

[0083] predictions=model.predict(new_data_scaled)

[0084] # Output prediction results

[0085] This code is only used as an example. In actual applications, appropriate modifications and adjustments need to be made according to specific circumstances.

[0086] The feedback module is used to generate feedback information according to the judgment result of the data processing module; the medication optimization module is used to optimize the medication regimen according to the historical medication data.

[0087] The feedback module in the embodiment of the present invention includes an information generating unit, a display unit and an alarm unit, wherein the information generating unit is used to generate feedback information according to the judgment result of the medication analysis module, and the feedback information includes the result of whether the medication is successful, the accuracy of the medication time and the correctness of the medication dosage; the display unit is used to display the feedback information and the reasons output by the medication optimization module that cause the medication success rate to be lower than the set threshold, and the alarm unit is used to trigger an alarm signal when the judgment result is that the medication success rate is lower than the set threshold.

[0088] The result of whether the medication is successful in the embodiment of the present invention is obtained by comparing the medication success rate with a preset threshold; the accuracy of the medication time is determined by comparing the actual medication duration with the preset medication time; the correctness of the medication dosage is determined by comparing the actual dispensed dosage with the prescribed dosage.

[0089] The judgment of whether the medication is successful or not, the evaluation of the accuracy of the medication time and the correctness of the dosage in the embodiment of the present invention are processed by using the numerical calculation library of Python. The judgment of whether the medication is successful is performed by comparing the calculated medication success rate with the preset threshold; specifically, if the medication success rate is greater than or equal to the preset threshold, it is judged as "successful", otherwise it is judged as "failed"; the accuracy of the medication time is evaluated by comparing the actual medication duration with the preset medication time, calculating the absolute value of the time difference, if the time difference is less than or equal to one minute, the time is considered accurate; the correctness of the medication dosage is judged by comparing the actual distribution dose with the prescribed dose, calculating the dosage error percentage, if the error percentage is less than or equal to five percent, the dosage is considered correct. These judgment results are recorded in the data framework, which is convenient for subsequent analysis and report generation, and can provide objective and quantitative medication evaluation standards, reducing the errors caused by subjective judgment; it can quickly and accurately evaluate multiple key indicators of medication, which helps to find problems and intervene in time; by setting reasonable thresholds, it can flexibly adapt to the special needs of different drugs and patients; the results are easy to understand and explain, which is conducive to medical staff and patients' families to understand the medication situation.

[0090] In the embodiment of the present invention, the drug dispensing module first accurately distributes the drug according to the preset drug administration scheme to ensure the correct dosage; then, the drug taking monitoring module uses a high-definition camera to capture the child's drug taking action in real time, uses image processing technology to denoise and enhance the captured image, and then uses a multi-layer convolutional neural network in a deep learning framework to perform image analysis, identify drug taking behavior and extract key feature parameters; these parameters are transmitted to the drug taking analysis module, which uses a specially designed drug taking success rate evaluation model and oral cooperation evaluation model, comprehensively considers factors such as drug distribution dose, drug taking duration, and oral cooperation degree, and calculates the drug taking success rate evaluation value; at the same time, the drug taking optimization module collects and processes historical drug taking data, uses a gradient boosting decision tree algorithm to build a prediction model, and provides a scientific basis for optimizing the drug taking scheme by analyzing the reasons for the low drug taking success rate; finally, the feedback module generates detailed feedback information based on the drug taking analysis results, including whether the drug taking is successful, the time accuracy and the dosage correctness, and presents the results through the display screen, and triggers an alarm reminder when necessary.

[0091] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art who is familiar with the present technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

[0092] Finally: The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A supervision and auxiliary system for children's medication, including a medication distribution module, a medication monitoring module, a medication analysis module, a medication optimization module and a feedback module, It is characterized in that The drug distribution module is connected to the medication monitoring module, the medication monitoring module is connected to the medication analysis module, the medication analysis module is connected to the medication optimization module, the medication optimization module is connected to the feedback module, and the feedback module is connected to the medication analysis module, wherein the drug distribution module is used to distribute drugs according to a preset medication regimen; the medication monitoring module is used to monitor children's medication behavior in real time, including monitoring children's medication actions through a camera and analyzing captured images through an image recognition algorithm; the medication analysis module is used to receive and process monitoring data transmitted by the medication monitoring module, input the constructed medication success rate evaluation model, and judge the children's medication success rate; the feedback module is used to generate feedback information according to the judgment result of the data processing module; the medication optimization module is used to optimize the medication regimen according to historical medication data; wherein the medication analysis module calculates the medication success rate evaluation value according to the medication success rate evaluation model, wherein the formula of the medication success rate evaluation model is: Where S is the evaluation value of the medication success rate, D is the drug distribution dose, A is the evaluation value of oral cooperation, T is the duration of medication, P is the pixel area occupied by the drug in the image, and t is the total time of the medication process.

2. A monitoring auxiliary system for children taking medicine according to claim 1, It is characterized in that The medication monitoring module includes a camera, an image processing unit and an image recognition unit. The camera is used to capture the child's medication action; the image processing unit is used to perform real-time denoising, enhancement and stabilization processing on the image captured by the camera.

3. A monitoring auxiliary system for children taking medicine according to claim 1, It is characterized in that The image recognition unit is based on a deep learning framework and analyzes the preprocessed image through a multi-layer convolutional neural network to identify the child's medication behavior and output medication characteristic parameters. The medication characteristic parameters include the child's mouth opening amplitude, medication duration, pixel area occupied by the drug in the image, and the total medication process time.

4. A monitoring auxiliary system for children taking medicine according to claim 1, It is characterized in that The monitoring data transmitted by the medication analysis module and the medication monitoring module are input into the collected children's mouth opening amplitude, swallowing times and emotional index into the oral cooperation evaluation model to evaluate the oral cooperation degree of children when taking medicine. The formula of the oral cooperation evaluation model is: In the formula, A is the oral cooperation assessment value, O is the child's mouth opening amplitude, C is the number of swallowing times, and EI is the emotional index.

5. A monitoring auxiliary system for children taking medicine according to claim 1, It is characterized in that The medication optimization module includes a data processing unit, a model building unit and a prediction unit, wherein the data processing unit is used to collect and process each child's historical medication data group, wherein the medication data group includes the medication success rate value, the child's mouth opening amplitude, the duration of medication, the pixel area occupied by the drug in the image and the total time of the medication process, the medication time, the emotional index, the child's age, weight, past medication compliance records, drug type, drug dosage, drug taste, number of swallowings, and medication location, and performs feature extraction, cleaning and standardization on the data, and divides the processed data into a training set and a test set; the prediction unit is used to receive the child's real-time medication data group, input the trained gradient boosting decision tree model, and output the predicted reasons for the medication success rate being lower than the set threshold.

6. A monitoring auxiliary system for children taking medicine according to claim 1, It is characterized in that The model building unit uses the training set data to train the gradient boosting decision tree model, and annotates the reasons for the low medication success rate and the medication data group with a set threshold in the training set. The medication success rate is used as the target variable and the medication data group is used as the feature variable. The trained gradient boosting decision tree model is tested using the test set data.

7. A monitoring auxiliary system for children taking medicine according to claim 1, It is characterized in that The feedback module includes an information generating unit, a display unit and an alarm unit, wherein the information generating unit is used to generate feedback information according to the judgment result of the medication analysis module, and the feedback information includes the result of whether the medication is successful, the accuracy of the medication time and the correctness of the medication dosage; the display unit is used to display the feedback information and the reasons output by the medication optimization module that cause the medication success rate to be lower than the set threshold, and the alarm unit is used to trigger an alarm signal when the judgment result is that the medication success rate is lower than the set threshold.

8. A monitoring auxiliary system for children taking medicine according to claim 1 or 6, It is characterized in that The result of whether the medication is successful is obtained by comparing the medication success rate with a preset threshold; the accuracy of the medication time is determined by comparing the actual medication duration with the preset medication time; the correctness of the medication dosage is determined by comparing the actual dispensed dosage with the prescribed dosage.

Images