CN119028596A - A diabetes risk early warning system and method - Google Patents

Abstract

The present invention discloses a diabetes risk early warning system and method, comprising: obtaining known patient information, physical sign information, facial images and case information, generating a patient database, dividing the patient database into a first database and a second database, wherein the patients in the first database are diabetic patients, and the patients in the second database are non-diabetic patients; obtaining patient information, case information and physical sign information of diabetic patients from the first database, outputting disease control decision information, obtaining patient information, case information and physical sign information of non-diabetic patients from the second database, outputting prevention and control decision information; real-time monitoring of the patient's physical sign information and facial images, and outputting disease control feedback information; receiving the disease control feedback information, and updating the disease control decision information and the prevention and control decision information in real time; using an early warning model to obtain the patient's disease control feedback information in real time, thereby improving the timeliness and accuracy of disease monitoring.

Description

Diabetes risk early warning system and method

Technical Field

The invention relates to the technical field of diabetes data processing, in particular to a diabetes risk early warning system and a diabetes risk early warning method.

Background

In recent years, diabetes has become a serious public health problem seriously harming national health, brings heavy economic burden to society, is urgent in prevention and treatment work, and has great significance in preventing diabetes by carrying out individualized data monitoring on crowds.

At present, the Chinese patent application publication No. CN116189896A discloses a cloud-based diabetes health data early warning method, and the disease symptom characteristics and reference sign data of each stage of different types of diabetes are called from a database and stored in a cloud server; detecting physical sign parameters of a target person in real time through portable wearable equipment of the target person, and uploading the real-time physical sign parameters to the cloud server; judging whether the physical sign parameters are abnormal or not, and setting a physical sign monitoring schedule for a target person according to a judging result; according to the sign monitoring plan, a target person is lifted to carry out periodic sign measurement, and a sign measurement result is obtained; comparing the physical sign measurement result with reference physical sign data of different types of diabetes mellitus, determining the type and the disease stage of the diabetes mellitus patient of a target person according to the comparison result, and carrying out early warning, so that the physical sign index detection can be accurately judged according to the measurement result on the premise of not affecting the self time arrangement of the patient; however, the related technology does not perform early warning of diabetes mellitus aiming at non-ill people, is unfavorable for early defense, obtains the case characteristics of ill people, performs periodic monitoring and analysis, and lacks timeliness and accuracy of monitoring and comprehensiveness of analysis.

Disclosure of Invention

The invention solves the technical problems that: in the related technology, early warning of diabetes mellitus is not carried out on non-ill people, early defense is not facilitated, case characteristics of ill people are obtained, periodic monitoring and analysis are carried out, timeliness and accuracy of monitoring are lacking, and comprehensiveness of analysis are lacking.

In order to overcome the defects in the prior art, in a first aspect, the invention provides a diabetes risk early warning method, which comprises the following steps:

Step S1, acquiring known patient information, sign information, facial images and case information, generating a patient database, and dividing the patient database into a first database and a second database, wherein the patients in the first database are diabetics, and the patients in the second database are non-diabetics;

Step S2, acquiring patient information, case information and sign information of a diabetic patient according to the first database, outputting disease control decision information according to the patient information, the case information and the sign information of the diabetic patient, acquiring patient information, case information and sign information of a non-diabetic patient according to the second database, and outputting prevention control decision information according to the patient information, the case information and the sign information of the non-diabetic patient;

Step S3, monitoring the physical sign information and the facial image of the patient in real time, and outputting disease control feedback information;

s4, receiving the disease control feedback information, and updating the disease control decision information and the prevention control decision information in real time;

As a preferable scheme of the diabetes risk early warning method, the invention comprises the following steps:

The patient information includes name, age, sex, height, blood pressure, blood sugar, weight, eating habits, acquired diseases and genetic diseases;

the case information includes symptoms, cause of onset, type of onset, and stage of onset;

the physical sign information comprises blood glucose information, heart rate information, activity information and sleep information;

As a preferable scheme of the diabetes risk early warning method, the invention comprises the following steps:

The step S1 specifically includes:

acquiring known patient information, sign information, facial images and case information;

judging whether the patient is a diabetic patient according to key indexes in the sign information;

If the patient is a diabetic patient, storing patient information, sign information, facial images and case information of the patient into a first database;

If the patient is a non-diabetic patient, storing patient information, sign information, facial images and case information of the patient to a second database;

As a preferable scheme of the diabetes risk early warning method, the invention comprises the following steps:

The step S2 specifically includes:

Inputting patient information of a diabetic patient into a first database, matching case information and sign information of the diabetic patient in the first database, taking the case information and the sign information as input of a first machine learning model, taking disease control decision information as output of the first machine learning model, training the first machine learning model until the disease control decision information is subjected to expert diagnosis, and completing training of the first machine learning model with accuracy being a percentage;

Inputting patient information of a non-diabetic patient into a second database, matching case information and sign information of the non-diabetic patient in the second database, taking the case information and the sign information as input of a second machine learning model, taking prevention control decision information as output of the second machine learning model, training the second machine learning model until the prevention control decision information is subjected to expert diagnosis, and completing training of the second machine learning model with accuracy being hundred percent;

As a preferable scheme of the diabetes risk early warning method, the invention comprises the following steps:

The step S3 specifically includes:

Invoking a first database to acquire known patient information, physical sign information, facial images and case information corresponding to a diabetic patient, acquiring diabetes types corresponding to the corresponding case information, identifying and extracting keywords of the case information corresponding to the diabetic patient, constructing a keyword joint model based on deep learning, updating joint parameters in real time, inputting the keyword joint model as keywords, and outputting the keywords as a first comprehensive phrase;

invoking a second database to acquire case information, patient information and case information corresponding to a non-diabetic patient, constructing a keyword joint model based on deep learning, updating joint parameters in real time, inputting the keyword joint model as keywords, and outputting keywords related to the keywords;

acquiring case information, patient information and sign information of a diabetic patient of a first database, and case information, patient information and sign information of a non-diabetic patient of a second database;

Invoking the first database and the second database, respectively taking patient information, physical sign information and case information of the diabetes patients and the non-diabetes patients as data sets, training the data sets by using a machine learning model, outputting sentences in the patient information and the physical sign information as case information, taking the case information as primary sentences, encoding and labeling the primary sentences, generating secondary sentences, and converting the secondary sentences into tertiary sentences which are the same as the primary sentences in length and fixed settable dimensions;

Extracting keywords in the three-level sentences, calculating the similarity between the keywords corresponding to the second database and the keywords corresponding to the data set, and comparing the similarity with a similarity threshold value:

If the similarity is larger than the similarity threshold, combining the keywords to generate a keyword group;

If the similarity is smaller than the similarity threshold, returning the keywords to the three-level sentences to generate a third database;

As a preferable scheme of the diabetes risk early warning method, the invention comprises the following steps:

The third database comprises patient information and corresponding keyword groups, keywords and facial images;

corresponding the patient information of the third database to the facial images one by one;

As a preferable scheme of the diabetes risk early warning method, the invention comprises the following steps:

the step S3 further includes:

Monitoring the physical sign information and the facial image of the patient in real time;

Taking blood glucose information, heart rate information, activity information and sleep information which are acquired in real time as keywords in a first database and a second database, inputting a keyword joint model, and outputting keywords related to the keywords;

Inputting the keywords and keywords related to the keywords into a machine learning model, and establishing a first diabetes risk early warning model;

the facial image in the third database is called, the facial image of the patient is matched by utilizing the face recognition technology, and patient information corresponding to the facial image is obtained;

Extracting and analyzing the characteristics of the facial image of the patient by utilizing an image recognition technology to obtain keywords and keyword groups corresponding to the facial image characteristics of the patient;

inputting keywords and keyword groups corresponding to facial image features of a patient into a machine learning model, and establishing a second diabetes risk early warning model;

Fusing the first diabetes risk early-warning model and the second diabetes risk early-warning model according to patient information by using a model fusion technology, and obtaining a diabetes risk early-warning model;

inputting the sign information and the facial image of the patient, and acquiring the disease control feedback information of the patient;

The condition control feedback information includes symptoms, type of illness, and stage of illness;

As a preferable scheme of the diabetes risk early warning method, the invention comprises the following steps:

The real-time monitoring of the patient's sign information and facial images includes:

Monitoring blood sugar of a patient by using a blood glucose meter to acquire blood sugar information of the patient;

the heart rate monitoring bracelet is used for monitoring the heart rate of the patient, and heart rate information of the patient is obtained;

Physical activity monitoring is carried out on a patient by using an accelerometer and a pedometer in the intelligent bracelet, and activity information of the patient is obtained;

the method comprises the steps of performing sleep monitoring on a patient by using an intelligent bracelet to obtain sleep information of the patient;

As a preferable scheme of the diabetes risk early warning method, the invention comprises the following steps:

the step S4 specifically includes:

Acquiring disease control feedback information of a patient, and judging whether the patient is a diabetic patient according to the critical disease index in the disease control feedback information;

If the diabetes mellitus patient is, inputting the disease control feedback information into a machine learning model, acquiring disease control decision information, and outputting and updating the disease control decision information after the disease control decision information is confirmed by an expert;

If the patient is a non-diabetic patient, the disease control feedback information is input into a machine learning model, the preventive control decision information is obtained, and after the expert confirms, the preventive control decision information is output and updated.

In a second aspect, a diabetes risk early warning system includes a monitoring module, an early warning module and a decision module;

the monitoring module is used for monitoring and acquiring the physical sign information and the facial image of the patient;

The early warning module is used for establishing an early warning model according to known patient information, sign information, facial images and case information, and acquiring disease condition control feedback information of a patient according to the sign information and the facial images of the patient by using the early warning model;

the decision module is used for comprehensively analyzing the disease control feedback of the patient and acquiring the disease control decision information of the diabetic patient and the prevention control decision information of the non-diabetic patient.

The invention has the beneficial effects that: according to the invention, the early warning model is built according to the known patient information, the sign information, the case information and the facial image, the sign information and the facial image of the patient are monitored in real time, and the early warning model is utilized to obtain the disease condition control feedback information of the patient, so that the timeliness and the accuracy of disease condition monitoring are improved.

The first diabetes risk early warning model and the second diabetes risk early warning model are established, the two models are fused to obtain the early warning model, and the disease control feedback information of the patient is obtained according to the blood sugar information, the heart rate information, the activity information, the sleep information and the facial image of the patient, so that the accuracy and the comprehensiveness of the diabetes risk early warning are improved through comprehensive analysis of various body data of the patient.

Drawings

Fig. 1 is a basic flow chart of a diabetes risk early warning method according to an embodiment of the present invention;

Fig. 2 is a basic flow chart of a diabetes risk early warning system according to an embodiment of the present invention.

Detailed Description

So that the manner in which the above recited objects, features and advantages of the present invention can be understood in detail, a more particular description of the invention, briefly summarized above, may be had by reference to the embodiments, some of which are illustrated in the appended drawings.

Embodiment 1, referring to fig. 1, provides a diabetes risk early warning method according to an embodiment of the present invention, including the following steps:

Step S1, acquiring known patient information, sign information, facial images and case information, generating a patient database, and dividing the patient database into a first database and a second database, wherein the patients in the first database are diabetics, and the patients in the second database are non-diabetics;

Step S2, acquiring patient information, case information and sign information of a diabetic patient according to the first database, outputting disease control decision information according to the patient information, the case information and the sign information of the diabetic patient, acquiring patient information, case information and sign information of a non-diabetic patient according to the second database, and outputting prevention control decision information according to the patient information, the case information and the sign information of the non-diabetic patient;

Step S3, monitoring the physical sign information and the facial image of the patient in real time, and outputting disease control feedback information;

s4, receiving the disease control feedback information, and updating the disease control decision information and the prevention control decision information in real time;

the acquiring patient information, case information and physical sign information of the non-diabetic patient according to the second database comprises:

Collecting physical sign information of a patient, judging whether the patient is a diabetic patient according to key indexes of the physical sign information of the patient, if the patient is a non-diabetic patient, matching corresponding patient information, case information and physical sign information of the non-diabetic patient in a second database according to physical sign identification of the characteristic information, and if the patient is not matched with the corresponding non-diabetic patient in the second database, collecting the patient information, the case information and the physical sign information of the patient to be stored in the second database together to serve as new patient data of the non-diabetic patient, wherein the physical sign identification is a unique identification for distinguishing each diabetic patient internally.

In the embodiment, an early warning model is established according to the known patient information, the sign information, the case information and the facial image, the sign information and the facial image of the patient are monitored in real time, and the early warning model is utilized to obtain the disease control feedback information of the patient, so that timeliness and accuracy of disease monitoring are improved;

The method comprises the steps of establishing a first diabetes risk early warning model and a second diabetes risk early warning model, fusing the two models to obtain an early warning model, obtaining disease control feedback information of a patient according to blood glucose information, heart rate information, activity information, sleep information and facial images of the patient, and comprehensively analyzing various physical conditions of the patient to improve accuracy and comprehensiveness of diabetes risk early warning;

The patient information includes name, age, sex, height, blood pressure, blood sugar, weight, eating habits, acquired diseases and genetic diseases;

the case information includes symptoms, cause of onset, type of onset, and stage of onset;

the physical sign information comprises blood glucose information, heart rate information, activity information and sleep information;

In the embodiment, accurate and comprehensive data support is provided for establishing a machine learning model by acquiring patient information, case information and sign information of a patient;

The step S1 specifically includes:

acquiring known patient information, sign information, facial images and case information;

judging whether the patient is a diabetic patient according to key indexes in the sign information;

If the patient is a diabetic patient, storing patient information, sign information, facial images and case information of the patient into a first database;

If the patient is a non-diabetic patient, storing patient information, sign information, facial images and case information of the patient to a second database;

in the embodiment, patient data is divided into diabetic patient data and non-diabetic patient data according to key indexes in physical sign information, and the diabetic patient data and the non-diabetic patient data are respectively stored in a first database and a second database, so that accurate data support is provided for establishing a disease control decision machine learning model and a prevention control decision machine learning model;

The step S2 specifically includes:

Inputting patient information of a diabetic patient into a first database, matching case information and sign information of the diabetic patient in the first database, taking the case information and the sign information as input of a first machine learning model, taking disease control decision information as output of the first machine learning model, training the first machine learning model until the disease control decision information is subjected to expert diagnosis, and completing training of the first machine learning model with accuracy being a percentage;

Inputting patient information of a non-diabetic patient into a second database, matching case information and sign information of the non-diabetic patient in the second database, taking the case information and the sign information as input of a second machine learning model, taking prevention control decision information as output of the second machine learning model, training the second machine learning model until the prevention control decision information is subjected to expert diagnosis, and completing training of the second machine learning model with accuracy being hundred percent;

The machine learning model training comprises:

Data preparation: preparing a diabetes patient data set and a non-diabetes patient data set which need training, cleaning and preprocessing the data, including missing value processing, abnormal value processing and repeated value processing of the data, and dividing the cleaned and processed data set into a training data set and a test data set;

Model selection: selecting a proper machine learning algorithm and model, wherein common algorithms comprise linear regression, logistic regression, decision trees, random forests, support vector machines, neural networks and the like;

Model training: training the selected machine learning model by using a training data set, and iteratively updating parameters of the machine learning model by adopting an optimization algorithm such as gradient descent and the like so as to minimize a loss function;

Model evaluation: in the training process of the machine learning model, evaluating the machine learning model to determine the performance of the machine learning model, wherein evaluation indexes generally comprise precision, recall rate, F1 value and the like;

Model parameter adjustment: according to the evaluation result of the machine learning model, parameter adjustment is carried out on the machine learning model so as to further improve the performance of the machine learning model;

Model preservation and deployment: after the machine learning model is trained, the trained machine learning model is saved, and the decision of the illness state control decision information and the prevention control decision information is carried out;

In the embodiment, the machine learning training is respectively carried out according to the diabetes patient data and the non-diabetes patient data, the disease control decision information of the diabetes patient and the prevention control decision information of the non-diabetes patient are respectively obtained through the machine learning model, and diagnosis confirmation is carried out through an expert, so that the accuracy of the diabetes disease control and the prevention control is improved, and medical resources are saved;

The step S3 specifically includes:

Invoking a first database to acquire known patient information, physical sign information, facial images and case information corresponding to a diabetic patient, acquiring diabetes types corresponding to the corresponding case information, identifying and extracting keywords of the case information corresponding to the diabetic patient, constructing a keyword joint model based on deep learning, updating joint parameters in real time, inputting the keyword joint model as keywords, and outputting the keywords as a first comprehensive phrase;

invoking a second database to acquire case information, patient information and case information corresponding to a non-diabetic patient, constructing a keyword joint model based on deep learning, updating joint parameters in real time, inputting the keyword joint model as keywords, and outputting keywords related to the keywords;

acquiring case information, patient information and sign information of a diabetic patient of a first database, and case information, patient information and sign information of a non-diabetic patient of a second database;

Invoking the first database and the second database, respectively taking patient information, physical sign information and case information of the diabetes patients and the non-diabetes patients as data sets, training the data sets by using a machine learning model, outputting sentences in the patient information and the physical sign information as case information, taking the case information as primary sentences, encoding and labeling the primary sentences, generating secondary sentences, and converting the secondary sentences into tertiary sentences which are the same as the primary sentences in length and fixed settable dimensions;

Extracting keywords in the three-level sentences, calculating the similarity between the keywords corresponding to the second database and the keywords corresponding to the data set, and comparing the similarity with a similarity threshold value:

If the similarity is larger than the similarity threshold, combining the keywords to generate a keyword group;

If the similarity is smaller than the similarity threshold, returning the keywords to the three-level sentences to generate a third database;

In the embodiment, comprehensive and accurate model input is provided for constructing a first diabetes risk early warning model by constructing a keyword combined model, and accurate and comprehensive model input is provided for constructing a second diabetes risk early warning model by constructing a third database;

The third database comprises patient information and corresponding keyword groups, keywords and facial images;

corresponding the patient information of the third database to the facial images one by one;

in the embodiment, corresponding fusion parameters are provided for fusing the first diabetes risk early warning model and the second diabetes early warning model to obtain the early warning model by corresponding patient information to the facial images one by one;

the step S3 further includes:

Monitoring the physical sign information and the facial image of the patient in real time;

Taking blood glucose information, heart rate information, activity information and sleep information which are acquired in real time as keywords in a first database and a second database, inputting a keyword joint model, and outputting keywords related to the keywords;

Inputting the keywords and keywords related to the keywords into a machine learning model, and establishing a first diabetes risk early warning model;

the facial image in the third database is called, the facial image of the patient is matched by utilizing the face recognition technology, and patient information corresponding to the facial image is obtained;

Extracting and analyzing the characteristics of the facial image of the patient by utilizing an image recognition technology to obtain keywords and keyword groups corresponding to the facial image characteristics of the patient;

inputting keywords and keyword groups corresponding to facial image features of a patient into a machine learning model, and establishing a second diabetes risk early warning model;

Fusing the first diabetes risk early-warning model and the second diabetes risk early-warning model according to patient information by using a model fusion technology, and obtaining a diabetes risk early-warning model;

inputting the sign information and the facial image of the patient, and acquiring the disease control feedback information of the patient;

The condition control feedback information includes symptoms, type of illness, and stage of illness;

the face recognition technology comprises the following steps:

Image quality evaluation: after the image is acquired, evaluating the image quality, screening out low-quality images, and reducing noise and errors in subsequent processing;

Face detection: and using a face detection algorithm to automatically locate a face region in the image, and ensuring that subsequent processing is concentrated in the face region.

Face alignment: the acquisition of the face image can be changed due to different shooting angles and postures, and in the preprocessing process, the common method is to align the face so that the positions of characteristic points such as eyes, noses and mouths in the image are kept consistent, and the difficulty of subsequent recognition is reduced;

image enhancement: enhancement processing is carried out on the image, such as image denoising, image contrast enhancement, histogram equalization and the like, so as to improve the image quality and enhance the face characteristics;

Feature extraction: after preprocessing, converting face features in the image into mathematical vector representation by using a feature extraction algorithm, so that subsequent recognition and comparison are facilitated;

the method for carrying out feature extraction and feature analysis on the facial image of the patient comprises an HOG facial feature extraction method and a DeepFace facial feature extraction method;

The HOG facial feature extraction method includes:

image preprocessing: the input image is subjected to operations such as gray level conversion, size adjustment, background elimination and the like, so that the subsequent feature extraction is facilitated;

Calculating the gradient: carrying out gradient calculation on the preprocessed image to obtain a gradient map of the image;

partitioning: dividing the gradient map into a plurality of unit areas, each unit area containing a certain number of gradients;

calculating a direction histogram: carrying out direction statistics on gradients in each unit area to obtain a direction histogram;

normalization: the direction histogram is normalized, so that subsequent feature matching and comparison are facilitated;

the DeepFace facial feature extraction method comprises the following steps:

data preprocessing: the input image is subjected to operations such as gray level conversion, size adjustment, background elimination and the like, so that the subsequent feature extraction is facilitated;

convolution layer: learning low-level features of the image using the convolutional layer;

Pooling layer: the pooling layer is used for reducing the spatial resolution of the image, and reducing the number of parameters and the computational complexity;

Full tie layer: learning advanced features using fully connected layers;

output layer: outputting the image features using an output layer;

In the embodiment, the early warning model is obtained by carrying out model fusion on the first diabetes risk early warning model and the second diabetes risk early warning model, so that the disease control feedback information obtained by using the early warning model is more comprehensive and accurate;

The real-time monitoring of the patient's sign information and facial images includes:

Monitoring blood sugar of a patient by using a blood glucose meter to acquire blood sugar information of the patient;

the heart rate monitoring bracelet is used for monitoring the heart rate of the patient, and heart rate information of the patient is obtained;

Physical activity monitoring is carried out on a patient by using an accelerometer and a pedometer in the intelligent bracelet, and activity information of the patient is obtained;

the method comprises the steps of performing sleep monitoring on a patient by using an intelligent bracelet to obtain sleep information of the patient;

the step S4 specifically includes:

Acquiring disease control feedback information of a patient, and judging whether the patient is a diabetic patient according to the critical disease index in the disease control feedback information;

If the diabetes mellitus patient is, inputting the disease control feedback information into a machine learning model, acquiring disease control decision information, and outputting and updating the disease control decision information after the disease control decision information is confirmed by an expert;

If the patient is a non-diabetic patient, inputting disease control feedback information into a machine learning model, acquiring preventive control decision information, and outputting and updating the preventive control decision information after expert confirmation;

In the embodiment, the machine learning model is utilized to acquire the disease control decision information and the prevention control decision information according to the disease control feedback information, so that the timeliness and the comprehensiveness of the diabetes risk early warning are improved.

Embodiment 2, referring to fig. 2, is another embodiment of the present invention, which is different from the first embodiment, and provides a remote control method for urinary surgery, including the foregoing diabetes risk early warning method, including a monitoring module, an early warning module, and a decision module;

the monitoring module is used for monitoring and acquiring the physical sign information and the facial image of the patient;

The early warning module is used for establishing an early warning model according to known patient information, sign information, facial images and case information, and acquiring disease condition control feedback information of a patient according to the sign information and the facial images of the patient by using the early warning model;

The decision module is used for comprehensively analyzing the disease control feedback of the patient to obtain the disease control decision information of the diabetic patient and the prevention control decision information of the non-diabetic patient;

In the embodiment, the monitoring module can monitor and acquire the physical sign information and the facial image of the patient in real time, the early warning module can acquire the disease condition control feedback information of the patient according to the physical sign information and the facial image of the patient, and the decision module can update the disease condition control decision information and the prevention control decision information according to the disease condition control feedback information of the patient, so that the accuracy, timeliness and comprehensiveness of the diabetes risk early warning are improved.

It should be appreciated that embodiments of the invention may be implemented or realized by a combination of computer hardware and software, or by computer instructions stored in a non-transitory computer-readable memory. The methods may be implemented in a computer program using standard programming techniques, including a non-transitory computer readable storage medium configured with a computer program, where the storage medium so configured causes a computer to operate in a specific and predefined manner, in accordance with the methods and drawings described in the specific embodiments. Each program may implement communication with the computer system in a high-level procedural or object-oriented programming language. However, the program(s) can be implemented in assembly or machine language, if desired. In any case, the language may be a compiled or interpreted language. Furthermore, the program can be run on a programmed application specific integrated circuit for this purpose.

It should be noted that the above embodiments are only for illustrating the technical solution of the present invention and not for limiting the same, and although the present invention has been described in detail with reference to the preferred embodiments, it should be understood by those skilled in the art that the technical solution of the present invention may be modified or substituted without departing from the spirit and scope of the technical solution of the present invention, which is intended to be covered in the scope of the claims of the present invention.

Claims (10)

1. The diabetes risk early warning method is characterized by comprising the following steps of:

Step S1, acquiring known patient information, sign information, facial images and case information, generating a patient database, and dividing the patient database into a first database and a second database, wherein the patients in the first database are diabetics, and the patients in the second database are non-diabetics;

Step S2, acquiring patient information, case information and sign information of a diabetic patient according to the first database, outputting disease control decision information according to the patient information, the case information and the sign information of the diabetic patient, acquiring patient information, case information and sign information of a non-diabetic patient according to the second database, and outputting prevention control decision information according to the patient information, the case information and the sign information of the non-diabetic patient;

Step S3, monitoring the physical sign information and the facial image of the patient in real time, and outputting disease control feedback information;

And S4, receiving the disease control feedback information, and updating the disease control decision information and the prevention control decision information in real time.

2. The diabetes risk early warning method according to claim 1, wherein:

The patient information includes name, age, sex, height, blood pressure, blood sugar, weight, eating habits, acquired diseases and genetic diseases;

the case information includes symptoms, cause of onset, type of onset, and stage of onset;

the sign information includes blood glucose information, heart rate information, activity information, and sleep information.

3. The diabetes risk early warning method according to claim 1, wherein: the step S1 specifically includes:

acquiring known patient information, sign information, facial images and case information;

judging whether the patient is a diabetic patient according to key indexes in the sign information;

If the patient is a diabetic patient, storing patient information, sign information, facial images and case information of the patient into a first database;

if the patient is a non-diabetic patient, the patient information, the sign information, the facial image, and the case information of the patient are stored to a second database.

4. The diabetes risk early warning method according to claim 1, wherein: the step S2 specifically includes:

Inputting patient information of a diabetic patient into a first database, matching case information and sign information of the diabetic patient in the first database, taking the case information and the sign information as input of a first machine learning model, taking disease control decision information as output of the first machine learning model, training the first machine learning model until the disease control decision information is subjected to expert diagnosis, and completing training of the first machine learning model with accuracy being a percentage;

Inputting the patient information of the non-diabetic patient into a second database, matching the case information and the sign information of the non-diabetic patient in the second database, taking the case information and the sign information as the input of a second machine learning model, taking the prevention control decision information as the output of the second machine learning model, and training the second machine learning model until the prevention control decision information is subjected to expert diagnosis, wherein the accuracy is a percentage, and completing the training of the second machine learning model.

5. The diabetes risk early warning method according to claim 1, wherein: the step S3 specifically includes:

Invoking a first database to acquire known patient information, physical sign information, facial images and case information corresponding to a diabetic patient, acquiring a diabetes type corresponding to the corresponding case information, identifying and extracting keywords of the case information corresponding to the diabetic patient, constructing a keyword joint model based on deep learning, updating joint parameters in real time, inputting the keyword joint model as keywords, and outputting the keywords as a first comprehensive phrase;

invoking a second database to acquire case information, patient information and case information corresponding to a non-diabetic patient, constructing a keyword joint model based on deep learning, updating joint parameters in real time, inputting the keyword joint model as keywords, and outputting keywords related to the keywords;

acquiring case information, patient information and sign information of a diabetic patient of a first database, and case information, patient information and sign information of a non-diabetic patient of a second database;

Invoking the first database and the second database, respectively taking patient information, physical sign information and case information of the diabetes patients and the non-diabetes patients as data sets, training the data sets by using a machine learning model, outputting sentences in the patient information and the physical sign information as case information, taking the case information as primary sentences, encoding and labeling the primary sentences, generating secondary sentences, and converting the secondary sentences into tertiary sentences which are the same as the primary sentences in length and fixed settable dimensions;

Extracting keywords in the three-level sentences, calculating the similarity between the keywords corresponding to the second database and the keywords corresponding to the data set, and comparing the similarity with a similarity threshold value:

If the similarity is larger than the similarity threshold, combining the keywords to generate a keyword group;

If the similarity is smaller than the similarity threshold, returning the keywords to the three-level sentences to generate a third database.

6. The diabetes risk early warning method according to claim 5, wherein:

The third database comprises patient information and corresponding keyword groups, keywords and facial images;

And corresponding the patient information of the third database to the facial images one by one.

7. The diabetes risk early warning method according to claim 1, wherein: the step S3 further includes:

Monitoring the physical sign information and the facial image of the patient in real time;

Taking blood glucose information, heart rate information, activity information and sleep information which are acquired in real time as keywords in a first database and a second database, inputting a keyword joint model, and outputting keywords related to the keywords;

Inputting the keywords and keywords related to the keywords into a machine learning model, and establishing a first diabetes risk early warning model;

the facial image in the third database is called, the facial image of the patient is matched by utilizing the face recognition technology, and patient information corresponding to the facial image is obtained;

Extracting and analyzing the characteristics of the facial image of the patient by utilizing an image recognition technology to obtain keywords and keyword groups corresponding to the facial image characteristics of the patient;

inputting keywords and keyword groups corresponding to facial image features of a patient into a machine learning model, and establishing a second diabetes risk early warning model;

Fusing the first diabetes risk early-warning model and the second diabetes risk early-warning model according to patient information by using a model fusion technology, and obtaining a diabetes risk early-warning model;

inputting the sign information and the facial image of the patient, and acquiring the disease control feedback information of the patient;

the condition control feedback information includes symptoms, type of illness, and stage of illness.

8. The diabetes risk early warning method according to claim 1, wherein: the real-time monitoring of the patient's sign information and facial images includes:

Monitoring blood sugar of a patient by using a blood glucose meter to acquire blood sugar information of the patient;

the heart rate monitoring bracelet is used for monitoring the heart rate of the patient, and heart rate information of the patient is obtained;

Physical activity monitoring is carried out on a patient by using an accelerometer and a pedometer in the intelligent bracelet, and activity information of the patient is obtained;

and carrying out sleep monitoring on the patient by using the intelligent bracelet to acquire sleep information of the patient.

9. The diabetes risk early warning method according to claim 1, wherein: the step S4 specifically includes:

Acquiring disease control feedback information of a patient, and judging whether the patient is a diabetic patient according to the critical disease index in the disease control feedback information;

If the diabetes mellitus patient is, inputting the disease control feedback information into a machine learning model, acquiring disease control decision information, and outputting and updating the disease control decision information after the disease control decision information is confirmed by an expert;

If the patient is a non-diabetic patient, the disease control feedback information is input into a machine learning model, the preventive control decision information is obtained, and after the expert confirms, the preventive control decision information is output and updated.

10. A diabetes risk early warning system comprising a diabetes risk early warning method according to any one of claims 1-9, characterized by comprising a monitoring module, an early warning module and a decision module;

the monitoring module is used for monitoring and acquiring the physical sign information and the facial image of the patient;

The early warning module is used for establishing an early warning model according to known patient information, sign information, facial images and case information, and acquiring disease condition control feedback information of a patient according to the sign information and the facial images of the patient by using the early warning model;

the decision module is used for comprehensively analyzing the disease control feedback of the patient and acquiring the disease control decision information of the diabetic patient and the prevention control decision information of the non-diabetic patient.