Ophthalmic disease intelligent pre-diagnosis system and method based on big data recognition

Abstract

The invention provides an intelligent ophthalmic disease pre-diagnosis system and method based on big data recognition, and the method comprises the steps: constructing a corresponding cataract pre-diagnosis model based on CNN training, extracting knowledge elements and association relationships related to cataract based on an LLM big language model, and constructing a corresponding cataract LLM knowledge base; after cataract clinical examination data of a patient is obtained, a first cataract pre-diagnosis strategy and a second cataract pre-diagnosis strategy corresponding to the first cataract pre-diagnosis strategy and the second cataract pre-diagnosis strategy are predicted through the cataract pre-diagnosis model and the cataract LLM knowledge base respectively; finally, strategy fusion is conducted on the two strategies through an LLM large language model, and a cataract treatment recommendation strategy is generated. According to the method, clinical examination data of the ophthalmic disease patient can be intelligently analyzed, a corresponding prediction scheme can be given, prediction results of the AI model and the LLM knowledge base are fused, rich and accurate eye diagnosis analysis results are generated for reference of doctors, the diagnosis service efficiency of the doctors is greatly improved, and the method is more intelligent and efficient compared with manual diagnosis analysis.

Description

Technical Field

[0001] The present invention relates to the field of intelligent medical technology, and in particular to an intelligent pre-diagnosis system and method for ophthalmic diseases based on big data recognition. Background Art

[0002] Cataract is a disease in which the lens of the eye becomes cloudy due to various reasons, causing visual impairment. Common types include senile cataract, complicated cataract, traumatic cataract, metabolic cataract, etc. Senile cataract is particularly common.

[0003] There are many causes of cataracts, such as aging, genetics, local nutritional disorders, immune and metabolic abnormalities, etc. In addition, exposure to strong light, smoking and alcohol abuse, malnutrition, long-term use of glucocorticoids, etc. will increase the risk of cataracts. Cataracts are the world's leading cause of blindness, especially among the elderly.

[0004] The initial symptoms of cataracts are usually not obvious, but as the disease progresses, symptoms such as blurred vision, diplopia, myopia, glare, etc. will appear, and may cause the patient to become completely blind.

[0005] The main treatment for cataracts is surgery, which involves removing the cloudy lens and implanting an artificial lens. Cataract surgery is now mature and has good results. Drug treatment is only suitable for patients with mild symptoms who have not yet reached the surgical criteria. During the recovery period, patients need to keep their eyes clean, prevent infection, and follow the doctor's instructions for follow-up examinations. The degree of improvement in vision after surgery depends on the condition of the fundus and cornea.

[0006] The existing clinical ophthalmology examination method for cataracts is mainly through the collection of eye PET or CT images, such as analyzing the presence of orbital nodules, lymph nodes and other symptoms in the eye from the eye CT images, and then the doctor analyzes the patient's cataract type and symptom status, prescribes the right medicine or gives the corresponding surgery and drug plan. However, the clinical diagnosis and analysis of cataract disease mainly relies on the attending ophthalmologist to perform cataract examination and treatment with superb clinical diagnosis experience and professional judgment ability. Doctors need to be able to analyze the image features on the eye CT images and identify the patient's pharyngeal disease type and disease progression status. This poses a challenge to most doctors, especially interns, and increases the difficulty of diagnosing the type and status of cataract disease and the application of medication / surgery.

[0007] In addition, although with the development of convolutional neural network (CNN) technology, CNN can be used to identify the characteristics of eye lesions in eye CT images, in the face of the diversified development of cataract diseases, the existing CNN model is still unable to accurately locate the symptoms and characteristics of cataracts and intelligently recommend corresponding cataract diagnosis strategies, and cannot provide doctors with more clinical pre-diagnosis auxiliary services. In addition, the existing scheme of using CNN to identify the characteristics of eye lesions in eye CT images and make predictions requires manual processing of clinical data in the early stage, and extracting corresponding data from patient examination data for AI analysis, which is not intelligent enough. Summary of the invention

[0008] In order to solve the technical problems existing in the prior art, the present invention provides the following technical solutions:

[0009] On the one hand, a method for intelligent pre-diagnosis of ophthalmic diseases based on big data recognition is provided, the method being implemented by an electronic device, and the method comprising:

[0010] S1. Collect cataract CT image sets and build corresponding cataract pre-diagnosis models based on CNN training;

[0011] S2. Collect big data on clinical diagnosis of cataracts, extract knowledge elements and associations related to cataracts based on the LLM large language model, and build a corresponding cataract LLM knowledge base;

[0012] S3, obtaining the patient's cataract clinical examination data and importing them into the cataract pre-diagnosis model and the cataract LLM knowledge base respectively, predicting the corresponding first cataract pre-diagnosis strategy by the cataract pre-diagnosis model, and retrieving the corresponding second cataract pre-diagnosis strategy by the cataract LLM knowledge base;

[0013] S4. Perform strategy fusion on the first cataract pre-diagnosis strategy and the second cataract pre-diagnosis strategy through the LLM large language model to generate and output a cataract treatment recommendation strategy.

[0014] Optionally, the step S1, collecting a cataract CT image set and building a corresponding cataract pre-diagnosis model based on CNN training, includes:

[0015] Collecting cataract CT images of multiple patients and performing preprocessing to obtain the cataract CT image set;

[0016] Performing feature engineering on the cataract CT image set, and using CNN to extract cataract CT image features in each cataract CT image, including: orbital lateral wall / nodule soft tissue morphology, SUV value, and color and texture features;

[0017] Performing feature annotation on the cataract CT image features, wherein the annotation information includes: the symptom stage corresponding to the cataract CT image features and the disease evolution trend thereof and the adopted clinical drug, surgery or a combination thereof treatment strategy;

[0018] After labeling is completed, a feature set consisting of a plurality of cataract CT image features is obtained;

[0019] Dividing the feature set into a training set and a validation set according to a preset ratio;

[0020] Inputting the training set into a preset CNN model to perform feature training and learning to generate the cataract pre-diagnosis model;

[0021] The prediction performance of the cataract pre-diagnosis model is verified using the validation set:

[0022] If the verification is successful, the cataract pre-diagnosis model is deployed and applied;

[0023] Otherwise, repeat the above steps to reconstruct the cataract pre-diagnosis model.

[0024] Optionally, the S2, collecting big data on clinical diagnosis of cataracts, extracting knowledge elements and associations related to cataracts based on the LLM large language model, and constructing a corresponding cataract LLM knowledge base, includes:

[0025] Collecting and preprocessing clinical diagnosis data of cataracts of multiple patients to obtain the clinical diagnosis big data of cataracts;

[0026] According to the symptom stage, the cataract clinical diagnosis big data is divided into a plurality of sub-datasets corresponding to different symptom stages;

[0027] Constructing entity keywords and onset association logic corresponding to different symptom stages respectively, and constructing large language model prompt words in combination with the entity keywords and the onset association logic, and inputting the large language model prompt words into a preset LLM large language model;

[0028] Traversing each of the sub-datasets through the LLM large language model, identifying and extracting cataract knowledge elements in the sub-datasets and association relationships between the cataract knowledge elements based on the large language model prompt words, wherein the association relationships include pathogenic factors and treatment strategies related to the symptom characteristics of cataract patients, and the treatment strategies include clinical drugs, surgery, or a combination thereof;

[0029] The cataract knowledge elements and their correlations in different symptom stages are sorted out through the LLM large language model to form a structured data set;

[0030] The structured data set is written into a preset knowledge base through the LLM large language model to obtain the cataract LLM knowledge base, and the cataract LLM knowledge base is deployed and applied.

[0031] Optionally, the step S3, obtaining the patient's cataract clinical examination data and importing the data into the cataract pre-diagnosis model and the cataract LLM knowledge base, respectively, predicting a corresponding first cataract pre-diagnosis strategy by the cataract pre-diagnosis model, and retrieving a corresponding second cataract pre-diagnosis strategy by the cataract LLM knowledge base, comprises:

[0032] Pre-configure parsing rules and input the LLM large language model, use the LLM large language model to read the cataract clinical examination data of the patient stored in the HIS system, and parse the cataract clinical examination data based on the parsing rules to obtain the cataract knowledge elements in the cataract CT image and the cataract clinical diagnosis data, and import the cataract knowledge elements in the cataract CT image and the cataract clinical diagnosis data into the cataract pre-diagnosis model and the cataract LLM knowledge base respectively;

[0033] Identifying cataract CT image features in the cataract CT image through the cataract pre-diagnosis model and predicting and outputting a first cataract pre-diagnosis strategy corresponding to the cataract CT image features, wherein the first cataract pre-diagnosis strategy includes the symptom stage of the cataract CT image features and the disease evolution trend thereof and the adopted clinical drug, surgery or a combination thereof treatment strategy;

[0034] The second cataract prognosis strategy associated with the cataract knowledge element is retrieved and output through the cataract LLM knowledge base, wherein the second cataract prognosis strategy includes the cataract knowledge element and the pathogenic factors and treatment strategies associated therewith, and the treatment strategies include clinical drugs, surgery or a combination thereof.

[0035] Optionally, the step S4, fusing the first cataract pre-diagnosis strategy and the second cataract pre-diagnosis strategy through an LLM large language model, generating and outputting a cataract treatment recommendation strategy, includes:

[0036] Pre-build the strategy clustering logic and input it into the LLM large language model;

[0037] The LLM large language model traverses the first cataract pre-diagnosis strategy and the second cataract pre-diagnosis strategy, clusters similar items in the first cataract pre-diagnosis strategy and the second cataract pre-diagnosis strategy based on the strategy clustering logic, and outputs corresponding clustering integration results, including: cataract symptom period and disease evolution trend, pathogenic factors and treatment strategies;

[0038] Writing the clustering integration results into the patient's electronic medical record in the HIS system, generating and saving the corresponding cataract treatment recommendation strategy for the patient;

[0039] Output and push the recommended cataract treatment strategy to the PC.

[0040] On the other hand, an intelligent pre-diagnosis system for ophthalmic diseases based on big data identification is provided, and the intelligent pre-diagnosis system for ophthalmic diseases based on big data identification is used to implement the above-mentioned intelligent pre-diagnosis method for ophthalmic diseases based on big data identification, and the system comprises:

[0041] Ophthalmological examination system, used to obtain patients' cataract clinical examination data and upload it to the backend server;

[0042] A backend server is used to import the patient's cataract clinical examination data into a preset cataract pre-diagnosis model and a preset cataract LLM knowledge base, respectively, predict the corresponding first cataract pre-diagnosis strategy by the cataract pre-diagnosis model, and retrieve the corresponding second cataract pre-diagnosis strategy by the cataract LLM knowledge base; perform strategy fusion on the first cataract pre-diagnosis strategy and the second cataract pre-diagnosis strategy through the LLM large language model, generate and output a cataract treatment recommendation strategy;

[0043] PC side, used for receiving and displaying the recommended cataract treatment strategy;

[0044] The ophthalmological examination system and the PC are respectively connected to the backend server for communication.

[0045] On the other hand, an electronic device is provided, comprising: a processor; and a memory, wherein the memory stores computer-readable instructions, and when the computer-readable instructions are executed by the processor, any one of the above-mentioned intelligent pre-diagnosis methods for ophthalmic diseases based on big data identification is implemented.

[0046] On the other hand, a computer-readable storage medium is provided, in which at least one instruction is stored, and the at least one instruction is loaded and executed by a processor to implement any one of the above-mentioned intelligent pre-diagnosis methods for ophthalmic diseases based on big data identification.

[0047] The beneficial effects brought about by the technical solution provided by the embodiment of the present invention include at least:

[0048] The present invention constructs a corresponding cataract pre-diagnosis model based on CNN training, extracts knowledge elements and associations related to cataracts based on the LLM large language model, and constructs a corresponding cataract LLM knowledge base; after obtaining the patient's cataract clinical examination data, the cataract pre-diagnosis model and the cataract LLM knowledge base respectively predict the corresponding first cataract pre-diagnosis strategy and the second cataract pre-diagnosis strategy; finally, the two strategies are fused through the LLM large language model to generate a cataract treatment recommendation strategy. It can intelligently analyze the clinical examination data of patients with ophthalmic diseases and give corresponding prediction plans, integrate the prediction results of the AI ​​model and the LLM knowledge base, and generate relatively rich and accurate eye diagnosis and analysis results for doctors' reference, greatly improving the efficiency of doctors' diagnostic services, and is more intelligent and efficient than manual diagnostic analysis.

[0049] By adopting the present invention, it is possible to further perform result analysis and prediction based on the AI ​​model in combination with the LLM large language model. On the basis of AI prediction, LLM is used to enrich and expand the prediction results, so that the disease diagnosis of cataract patients is more comprehensive, the prediction range is wider and more accurate, and more AI prediction suggestions are provided to doctors. Compared with traditional single or a few prediction results, the present invention enriches the AI ​​pre-diagnosis analysis dimensions of doctors for ophthalmic patients, and is more capable of controlling the patients' eye disease conditions. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0051] Figure 1 This is a flow chart of an intelligent pre-diagnosis method for ophthalmic diseases based on big data recognition provided by an embodiment of the present invention;

[0052] Figure 2 It is a schematic diagram of a process of strategy identification and fusion provided by an embodiment of the present invention;

[0053] Figure 3 is a training flow chart of a cataract pre-diagnosis model provided by an embodiment of the present invention;

[0054] Figure 4 This is a block diagram of an intelligent pre-diagnosis system for ophthalmic diseases based on big data recognition provided by an embodiment of the present invention;

[0055] Figure 5 It is a structural schematic diagram of an electronic device provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0056] The technical solution of the present invention is described below in conjunction with the accompanying drawings.

[0057] In the embodiments of the present invention, words such as "exemplarily" and "for example" are used to indicate examples, illustrations or explanations. Any embodiment or design described as "example" in the present invention should not be interpreted as being more preferred or more advantageous than other embodiments or designs. Specifically, the use of the word "example" is intended to present the concept in a specific way. In addition, in the embodiments of the present invention, the meaning expressed by "and / or" can be both, or it can be either of the two.

[0058] In the embodiments of the present invention, "image" and "picture" can sometimes be used interchangeably. It should be noted that when the difference between them is not emphasized, the meanings they intend to express are the same. "of", "corresponding, relevant" and "corresponding" can sometimes be used interchangeably. It should be noted that when the difference between them is not emphasized, the meanings they intend to express are the same.

[0059] In the embodiments of the present invention, sometimes a subscript such as W1 may be expressed in a non-subscript form such as W1. When the difference is not emphasized, the meanings to be expressed are consistent.

[0060] In order to make the technical problems, technical solutions and advantages to be solved by the present invention more clear, a detailed description will be given below with reference to the accompanying drawings and specific embodiments.

[0061] The present invention takes cataract as an example. For other ophthalmic diseases, the principles of the present invention can be referred to for implementation.

[0062] The embodiment of the present invention provides an intelligent pre-diagnosis method for ophthalmic diseases based on big data recognition, which can be implemented by an electronic device, which can be a terminal or a server. Figure 1 The flowchart of the intelligent pre-diagnosis method of ophthalmic diseases based on big data recognition is shown in the figure. The processing flow of the method may include the following steps:

[0063] S1. Collect cataract CT image sets and build corresponding cataract pre-diagnosis models based on CNN training;

[0064] S2. Collect big data on clinical diagnosis of cataracts, extract knowledge elements and associations related to cataracts based on the LLM large language model, and build a corresponding cataract LLM knowledge base;

[0065] S3, obtaining the patient's cataract clinical examination data and importing them into the cataract pre-diagnosis model and the cataract LLM knowledge base respectively, predicting the corresponding first cataract pre-diagnosis strategy by the cataract pre-diagnosis model, and retrieving the corresponding second cataract pre-diagnosis strategy by the cataract LLM knowledge base;

[0066] S4. Perform strategy fusion on the first cataract pre-diagnosis strategy and the second cataract pre-diagnosis strategy through the LLM large language model to generate and output a cataract treatment recommendation strategy.

[0067] like Figure 2 As shown, the present invention extracts knowledge elements and associations related to cataracts based on the LLM large language model, and constructs a corresponding cataract LLM knowledge base; after obtaining the patient's cataract clinical examination data, the cataract pre-diagnosis model and the cataract LLM knowledge base respectively predict the corresponding first cataract pre-diagnosis strategy and the second cataract pre-diagnosis strategy; finally, the two strategies are integrated through the LLM large language model to generate a cataract treatment recommendation strategy. It can intelligently analyze the clinical examination data of patients with ophthalmic diseases and give corresponding prediction plans, integrate the prediction results of the AI ​​model and the LLM knowledge base, and generate relatively rich and accurate eye diagnosis and analysis results for doctors' reference, greatly improving the efficiency of doctors' diagnostic services, and is more intelligent and efficient than manual diagnosis and analysis.

[0068] On the basis of the AI ​​model, the LLM large language model is combined to further perform result analysis and prediction. On the basis of AI prediction, LLM is used to enrich and expand the prediction results, so that the disease diagnosis of cataract patients is more comprehensive, the prediction range is wider and more accurate, and more AI prediction suggestions are provided to doctors. Compared with traditional single or a few prediction results, the present invention enriches the AI ​​pre-diagnosis analysis dimensions of doctors for ophthalmic patients and is better able to control the patients' eye diseases.

[0069] The following will describe in detail the construction principles of the model and knowledge base as well as the fusion of strategy recommendations.

[0070] like Figure 3 As shown, optionally, the S1, collecting a cataract CT image set, and building a corresponding cataract pre-diagnosis model based on CNN training, includes:

[0071] Collecting cataract CT images of multiple patients and performing preprocessing to obtain the cataract CT image set;

[0072] Performing feature engineering on the cataract CT image set, and using CNN to extract cataract CT image features in each cataract CT image, including: orbital lateral wall / nodule soft tissue morphology, SUV value, and color and texture features;

[0073] Performing feature annotation on the cataract CT image features, wherein the annotation information includes: the symptom stage corresponding to the cataract CT image features and the disease evolution trend thereof and the adopted clinical drug, surgery or a combination thereof treatment strategy;

[0074] After labeling is completed, a feature set consisting of a plurality of cataract CT image features is obtained;

[0075] Dividing the feature set into a training set and a validation set according to a preset ratio;

[0076] Inputting the training set into a preset CNN model to perform feature training and learning to generate the cataract pre-diagnosis model;

[0077] The prediction performance of the cataract pre-diagnosis model is verified using the validation set:

[0078] If the verification is successful, the cataract pre-diagnosis model is deployed and applied;

[0079] Otherwise, repeat the above steps to reconstruct the cataract pre-diagnosis model.

[0080] The present invention uses a convolutional neural network to identify and extract cataract image features on CT images of cataract patients, and performs feature annotation, and annotates the cataract image features with corresponding symptom stages, disease evolution trends, and treatment strategies, so as to train a model using CNN to generate a corresponding cataract prediagnosis model. The model can be deployed in the background and applied to intelligent image recognition of cataract patients, and can identify the current symptom stage of the cataract patient and the disease evolution trend that matches the CT image features of the patient, so that the doctor can predict the possible disease development trend of the cataract; at the same time, strategy matching and output are performed according to the identified features, and the model automatically matches and identifies the corresponding image features, and matches and outputs the corresponding treatment strategies, including medication or surgery or a combination of medication treatment strategies, so that the model can quickly perform clinical prediagnosis on cataract patients, output and recommend corresponding cataract disease analysis results, and provide corresponding treatment strategies for doctors' reference.

[0081] CNN extracts image features, and the extracted features can be nodules at different symptomatic stages, or the morphology of soft tissue lesions and the SUV value and color and texture features on the image. In different stages, such as the early and late stages of cataract, the color, color, texture and other features of the cataract soft tissue will change. Therefore, the corresponding color and texture features can be extracted to distinguish the image features of cataracts at different symptomatic stages.

[0082] The annotated information can be the symptom stage of the current cataract patient, such as the early stage of cataract. At the same time, the possible evolution and development trend of the symptom can be annotated, such as the early stage cataract may evolve into the corresponding evolution trend of the later stage, etc., and all the evolution results can be annotated on the image features. The specific treatment strategy can be annotated by the administrator, such as the attending physician, and the corresponding drug or surgical treatment strategy can be annotated according to the cataract symptom stage corresponding to the current image. After the annotation is completed, a feature set composed of multiple features can be obtained, and it can be divided into a training level and a validation set according to a ratio of 7:3 or 8:2, for example. The training set is input into the CNN model for feature training and learning to construct the AI ​​model; the validation set is then used to verify the recognition and prediction performance of the AI ​​model, for example, af1 or accuracy, precision, ROC curve, etc. can be used to verify the model performance. It can be implemented by the administrator.

[0083] After the AI ​​model is built, it can be deployed on the background server. Later, when a clinical eye CT examination is performed on the patient, the corresponding CT image can be read from the patient's electronic medical record in the HIS system (after the system collects the image, it is uploaded and saved by the imaging department) and imported into the AI ​​model. The AI ​​model recognizes the image features and matches and outputs the corresponding clinical analysis results for the doctor's reference.

[0084] The structure and general training process of CNN are well-known technologies and will not be described in detail in this embodiment. For example, the steps of training a model using a convolutional neural network (CNN) generally include the following key stages:

[0085] 1. Data Preparation

[0086] ‌Data Collection‌: Acquire large amounts of task-related image data from reliable sources.

[0087] ‌Data Labeling‌: Add labels or annotations to image data so that the model can learn during the training process.

[0088] ‌Data preprocessing‌: Scaling, cropping, and normalizing images to improve model training efficiency and performance.

[0089] ‌Data Partition‌: Divide the dataset into training set, validation set, and test set, usually in a ratio of 70%, 15%, and 15% or adjust according to specific circumstances.

[0090] 2. Model Construction

[0091] ‌Select a framework‌: Choose a suitable deep learning framework, such as TensorFlow, PyTorch, etc.

[0092] ‌Define model structure‌: Design the structure of CNN according to task requirements, including input layer, convolution layer, pooling layer, fully connected layer, output layer, etc.

[0093] ‌Initialization parameters‌: Set initial values ​​for the model’s weights and biases, typically using random initialization or weights from a pre-trained model.

[0094] 3. Model Training

[0095] ‌Define loss function‌: Choose an appropriate loss function to evaluate the difference between the model's prediction results and the actual labels, such as cross entropy loss, mean squared error, etc.

[0096] ‌Select Optimizer‌: Select an optimization algorithm to update the parameters of the model, such as SGD, Adam, RMSprop, etc.

[0097] ‌Set training parameters‌: including learning rate, batch size, number of training rounds, etc.

[0098] Start training: Input the training data into the model, calculate the loss through forward propagation, and then update the parameters through backpropagation. Repeat this process until the predetermined number of training rounds is reached or the loss converges.

[0099] 4. Model Validation and Evaluation

[0100] Validation model: Evaluate the performance of the model on the validation set, such as accuracy, recall, F1 score, etc., to monitor whether the model is overfitting or underfitting.

[0101] ‌Adjust hyperparameters‌: Adjust the model’s hyperparameters (such as learning rate, convolution kernel size, etc.) based on the validation results and retrain the model.

[0102] ‌Model Evaluation‌: Evaluate the final performance of the model on the test set to obtain the performance of the model on unknown data.

[0103] 5. Model deployment and optimization

[0104] ‌Model Export‌: Export the trained model to a deployable format, such as TensorFlow SavedModel, PyTorch model file, etc.

[0105] ‌Model deployment‌: Deploy the model to a production environment or actual application, such as a mobile device, cloud server, etc.

[0106] ‌Performance optimization‌: Further optimize the model, such as quantization and pruning, to improve the model's operating efficiency and reduce resource consumption.

[0107] 6. Continuous monitoring and updating

[0108] ‌Monitor model performance‌: Continuously monitor the performance of the model in actual applications to identify and resolve problems in a timely manner.

[0109] ‌Update the Model‌: As new data emerges and task requirements change, the model is regularly updated to maintain its performance and accuracy.

[0110] The above are the general steps for using CNN training models. When implementing them, appropriate adjustments and optimizations may be required based on task requirements and data characteristics.

[0111] In order to enrich and optimize the pre-diagnosis results of ophthalmic diseases, the present invention also uses the LLM large language model to construct a knowledge base related to ophthalmic diseases, so as to search for the onset and treatment strategies of patients' diseases such as cataracts for output fusion of the AI ​​model.

[0112] Optionally, the S2, collecting big data on clinical diagnosis of cataracts, extracting knowledge elements and associations related to cataracts based on the LLM large language model, and constructing a corresponding cataract LLM knowledge base, includes:

[0113] Collecting and preprocessing clinical diagnosis data of cataracts of multiple patients to obtain the clinical diagnosis big data of cataracts;

[0114] According to the symptom stage, the cataract clinical diagnosis big data is divided into a plurality of sub-datasets corresponding to different symptom stages;

[0115] Constructing entity keywords and onset association logic corresponding to different symptom stages respectively, and constructing large language model prompt words in combination with the entity keywords and the onset association logic, and inputting the large language model prompt words into a preset LLM large language model;

[0116] Traversing each of the sub-datasets through the LLM large language model, identifying and extracting cataract knowledge elements in the sub-datasets and association relationships between the cataract knowledge elements based on the large language model prompt words, wherein the association relationships include pathogenic factors and treatment strategies related to the symptom characteristics of cataract patients, and the treatment strategies include clinical drugs, surgery, or a combination thereof;

[0117] The cataract knowledge elements and their correlations in different symptom stages are sorted out through the LLM large language model to form a structured data set;

[0118] The structured data set is written into a preset knowledge base through the LLM large language model to obtain the cataract LLM knowledge base, and the cataract LLM knowledge base is deployed and applied.

[0119] It is possible to collect ophthalmic clinical diagnosis data of multiple patients, such as clinical diagnosis data of cataract patients, and obtain cataract clinical diagnosis big data after preprocessing. Preprocessing can be data classification, formatting and cleaning, for example, the cataract clinical diagnosis data of multiple patients can be divided into multiple sub-datasets corresponding to different symptom stages according to the symptom stage. Each sub-dataset, or subset, is a type of cataract clinical diagnosis data of the symptomatic stage, which contains the patient's basic information, symptom characteristics, pathogenic factors and treatment strategies. It can also include corresponding living habits, work, life, study and other data, so as to facilitate the use of data analysis and retrieval technology to find the pathogenic factors related to the patient's cataract disease.

[0120] This paper uses the LLM large language model to extract knowledge elements related to cataract disease from subsets of each symptom based on prompt words (consisting of entity keywords corresponding to the symptom period (such as early cataract, onset and other keywords) and onset association logic (early symptoms + etiology analysis and other logic)) and extract the association between each knowledge element, so as to capture the pathogenic factors of cataract disease and the treatment strategies for cataract disease. By using the large language model to perform intelligent mining and analyze the pathogenic factors and treatment strategies of cataract disease, and forming a structured data set based on the mining output results, the structured data set is written into the knowledge base, so as to obtain the cataract LLM knowledge base.

[0121] In the subsequent clinical diagnosis of cataract patients, the corresponding clinical consultation data of cataract patients are collected, and the corresponding search instructions are generated and imported into the knowledge base. In the knowledge base, based on the data structure formed in the cataract LLM knowledge base, the pathogenic factors and treatment strategies of different symptom stages are searched, matched with the patient's consultation information, and the relevant pathogenic factors that match the patient's consultation data are found and the corresponding treatment strategies are output. Therefore, it is possible to use the large language model to realize intelligent retrieval and analysis, and recommend corresponding clinical diagnosis and analysis results for cataract patients, including the pathogenic factors of the corresponding symptom stage and the corresponding treatment strategies.

[0122] Specific: ‌ Steps to use the large language model to extract the historical medical data of cataract patients and build a cataract LLM knowledge base ‌

[0123] ‌1. Data collection and preprocessing‌

[0124] ‌Collect historical medical data‌: Collect historical medical data of cataract patients from the hospital's electronic medical record system, medical record room or other data sources. These data should include the patient's basic information, symptom description, diagnosis results, treatment strategies, etc.‌

[0125] ‌Data preprocessing‌: Clean and normalize the collected data to ensure the accuracy and consistency of the data. For example, standardize the description of symptoms to eliminate the differences in synonyms or near-synonyms; classify and code the pathogenic factors and treatment strategies, etc.‌

[0126] ‌2. Extracting key information using large language models‌

[0127] ‌Text extraction and entity recognition‌: Use a large language model to extract key information from medical record text, including the patient's basic information (such as name, age, gender, etc.), symptom characteristics (such as blurred vision, glare, etc.), pathogenic factors (such as environmental factors, nutritional deficiencies, etc.) and treatment strategies (such as drug treatment, surgical treatment, etc.)‌.

[0128] ‌Relationship extraction and time processing‌: The large language model can identify relationships in the text, such as the relationship between symptoms and diseases, the relationship between drugs and treatment time, etc. At the same time, it processes time-related information in the text, such as medical history time, consultation time, drug use time, etc.‌

[0129] ‌3. Building a Cataract LLM Knowledge Base‌

[0130] ‌Data classification and organization‌: The extracted key information is classified and organized according to the patient's condition, symptom characteristics, pathogenic factors and treatment strategies to form a structured data set.

[0131] ‌Knowledge base construction‌: Based on the organized structured data, a cataract LLM knowledge base is constructed. This knowledge base should include basic information of patients, detailed description of symptom characteristics, analysis of pathogenic factors, and recommendations for treatment strategies.

[0132] ‌Model training and optimization‌: Use the built knowledge base to train the large language model to improve the model's ability to understand and process cataract-related information. At the same time, optimize the model according to actual application conditions to improve accuracy and efficiency.

[0133] ‌Continuous updating and maintenance‌: As new medical data is constantly generated, the knowledge base needs to be updated and maintained regularly to ensure the timeliness and accuracy of the information.

[0134] Through the above steps, the large language model can be used to extract key information from the historical medical data of cataract patients, and a cataract LLM knowledge base can be constructed to provide strong support for the diagnosis, treatment and research of cataracts. Please note that in the process of building and using the knowledge base, privacy and security regulations should be strictly followed to ensure the protection of sensitive patient information.

[0135] Traditional AI models intelligently output the predicted disease type or a single, small number of model prediction results, while the present invention integrates the AI ​​model prediction results with the search results of the cataract LLM knowledge base. LLM clusters and fuses the two output results based on clustering prompt words to obtain the final pre-diagnosis information, allowing doctors to obtain richer pre-diagnosis information on cataract diseases, comprehensively analyze the patient's disease condition, and provide reference for doctors to choose the best strategy.

[0136] Optionally, the step S3, obtaining the patient's cataract clinical examination data and importing the data into the cataract pre-diagnosis model and the cataract LLM knowledge base, respectively, predicting a corresponding first cataract pre-diagnosis strategy by the cataract pre-diagnosis model, and retrieving a corresponding second cataract pre-diagnosis strategy by the cataract LLM knowledge base, comprises:

[0137] Pre-configure parsing rules and input the LLM large language model, use the LLM large language model to read the cataract clinical examination data of the patient stored in the HIS system, and parse the cataract clinical examination data based on the parsing rules to obtain the cataract knowledge elements in the cataract CT image and the cataract clinical diagnosis data, and import the cataract knowledge elements in the cataract CT image and the cataract clinical diagnosis data into the cataract pre-diagnosis model and the cataract LLM knowledge base respectively;

[0138] Identifying cataract CT image features in the cataract CT image through the cataract pre-diagnosis model and predicting and outputting a first cataract pre-diagnosis strategy corresponding to the cataract CT image features, wherein the first cataract pre-diagnosis strategy includes the symptom stage of the cataract CT image features and the disease evolution trend thereof and the adopted clinical drug, surgery or a combination thereof treatment strategy;

[0139] The second cataract prognosis strategy associated with the cataract knowledge element is retrieved and output through the cataract LLM knowledge base, wherein the second cataract prognosis strategy includes the cataract knowledge element and the pathogenic factors and treatment strategies associated therewith, and the treatment strategies include clinical drugs, surgery or a combination thereof.

[0140] We use LLM to analyze patient examination data and obtain corresponding CT data and other consultation data, so as to obtain input information for the AI ​​model and cataract LLM knowledge base respectively, thereby replacing manual selection and improving efficiency.

[0141] ‌ The steps for using the LLM large language model to parse cataract clinical examination data can be referred to as follows:‌

[0142] 1. Pre-configure parsing rules:

[0143] Set specific parsing rules for cataract clinical examination data, including data format, key field identification, etc.

[0144] These rules should be able to accurately identify and extract key knowledge elements from cataract CT images and cataract clinical diagnosis data.

[0145] 2. Input the LLM large language model:

[0146] Input the pre-configured parsing rules into the LLM large language model.

[0147] Leverage the powerful natural language processing and deep learning capabilities of the LLM large language model to efficiently process input data.

[0148] 3. Read HIS system data:

[0149] Through the LLM large language model, connect and read the patient's cataract clinical examination data stored in the hospital information system (HIS).

[0150] Ensure the integrity and accuracy of the data and lay the foundation for subsequent analysis work.

[0151] 4.Analysis of cataract clinical examination data:

[0152] Image data analysis: The image processing technology in the LLM large language model is used to analyze cataract CT images and identify key features in the images, such as the degree and location of the lens opacity.

[0153] ‌Text data analysis‌: According to the pre-configured analysis rules, the text information in the cataract clinical diagnosis data is analyzed to extract key knowledge elements such as cataract diagnosis results, symptom descriptions, medical history, etc.‌

[0154] ‌Get cataract knowledge elements‌:

[0155] Integrate the analyzed cataract CT image features and key knowledge elements in clinical diagnosis data.

[0156] Form a structured cataract knowledge base to provide strong support for subsequent clinical decision-making, scientific research analysis, etc.

[0157] Through the above steps, the LLM large language model can be used to efficiently read and parse the cataract clinical examination data in the HIS system and extract valuable cataract knowledge elements.

[0158] The cataract pre-diagnosis model identifies and predicts the output of a first cataract pre-diagnosis strategy corresponding to the cataract CT image feature, and the cataract LLM knowledge base retrieves and outputs a second cataract pre-diagnosis strategy, which can be combined with the construction principle of the above model and knowledge base. No further details will be given here.

[0159] Optionally, the step S4, fusing the first cataract pre-diagnosis strategy and the second cataract pre-diagnosis strategy through an LLM large language model, generating and outputting a cataract treatment recommendation strategy, includes:

[0160] Pre-build the strategy clustering logic and input it into the LLM large language model;

[0161] The LLM large language model traverses the first cataract pre-diagnosis strategy and the second cataract pre-diagnosis strategy, clusters similar items in the first cataract pre-diagnosis strategy and the second cataract pre-diagnosis strategy based on the strategy clustering logic, and outputs corresponding clustering integration results, including: cataract symptom period and disease evolution trend, pathogenic factors and treatment strategies;

[0162] Writing the clustering integration results into the patient's electronic medical record in the HIS system, generating and saving the corresponding cataract treatment recommendation strategy for the patient;

[0163] Output and push the recommended cataract treatment strategy to the PC.

[0164] Here, LLM performs cluster analysis on the two results, allowing LLM to cluster the same items in the two strategies (the same or similar items under the same symptom period) based on the clustering strategy.

[0165] Strategy clustering logic can include, for example, drug strategy clustering, merging the same drug treatment methods in two strategies into one item. If the drugs in the two strategies are the same or belong to the same class of drugs, they can be merged. For example, the two strategies have the following topical drugs:

[0166] ‌Pirenoxine eye drops‌: can shorten the tear excretion time of corneal tissue, help relieve eye fatigue, and reduce eye discomfort caused by cataracts‌;

[0167] ‌Falconel Eye Drops‌: Helps improve the patient's vision by promoting protein decomposition. ‌

[0168] Then merge the two clusters into “eye drops”.

[0169] Please refer to the following steps for details:

[0170] 1. Pre-build strategy clustering logic

[0171] ‌Define clustering criteria‌: Clarify the definition of similar items in cataract prediagnosis strategies, such as similar symptom descriptions, the same diagnostic methods or treatment means, etc.

[0172] ‌Design clustering algorithm‌: Based on the defined clustering criteria, design a clustering algorithm suitable for the LLM large language model to ensure that the algorithm can accurately identify and cluster similar items.

[0173] 2. Input the LLM large language model and traverse the strategy

[0174] ‌Input clustering logic‌: Input the constructed strategy clustering logic into the LLM large language model.

[0175] ‌Traverse the prognosis strategies‌: Utilize the powerful text processing capabilities of the LLM large language model to traverse all the contents in the first cataract prognosis strategy and the second cataract prognosis strategy.

[0176] 3. Clustering based on clustering logic

[0177] ‌Identify similar items‌: During the traversal process, the LLM large language model automatically identifies similar items in the two pre-diagnosis strategies based on clustering logic.

[0178] ‌Perform clustering operations‌: Cluster the identified similar items to form cluster groups with common characteristics.

[0179] 4. Output clustering integration results

[0180] ‌Integrate cluster information‌: Integrate the information in the cluster group and extract key information, such as the theme of the cluster group, the included policy items, etc.

[0181] ‌Output results‌: Output the integrated clustering information in a clear and easy-to-read manner to facilitate subsequent analysis and application. ‌‌

[0182] Different items, such as the patient's disease trend and disease factors, can be output separately without being merged. Through the above clustering fusion, the pre-diagnosis prediction results of cataract patients are enriched for doctors' reference.

[0183] Figure 4 The block diagram of an intelligent pre-diagnosis system for ophthalmic diseases based on big data identification is shown according to an exemplary embodiment. The system is used for an intelligent pre-diagnosis method for ophthalmic diseases based on big data identification. Figure 3 The system includes an ophthalmic examination system 310, a backend server 320 and a PC 330. Among them:

[0184] Ophthalmological examination system, used to obtain patients' cataract clinical examination data and upload it to the backend server;

[0185] A backend server is used to import the patient's cataract clinical examination data into a preset cataract pre-diagnosis model and a preset cataract LLM knowledge base, respectively, predict the corresponding first cataract pre-diagnosis strategy by the cataract pre-diagnosis model, and retrieve the corresponding second cataract pre-diagnosis strategy by the cataract LLM knowledge base; perform strategy fusion on the first cataract pre-diagnosis strategy and the second cataract pre-diagnosis strategy through the LLM large language model, generate and output a cataract treatment recommendation strategy;

[0186] PC side, used for receiving and displaying the recommended cataract treatment strategy;

[0187] The ophthalmological examination system and the PC are respectively connected to the backend server for communication.

[0188] Please understand the interaction between the above systems in conjunction with the above method steps, and will not be repeated here.

[0189] Figure 5 is a schematic diagram of the structure of an electronic device provided by an embodiment of the present invention, such as Figure 5 As shown, the electronic device 410 may include a first processor 2001 .

[0190] Optionally, the electronic device 410 may further include a memory 2002 and a transceiver 2003 .

[0191] The first processor 2001, the memory 2002 and the transceiver 2003 may be connected via a communication bus.

[0192] Combine the following Figure 5 The components of the electronic device 410 are described in detail:

[0193] The first processor 2001 is the control center of the electronic device 410, and may be a processor or a general term for multiple processing elements. For example, the first processor 2001 is one or more central processing units (CPUs), or may be application specific integrated circuits (ASICs), or may be one or more integrated circuits configured to implement the embodiments of the present invention, such as one or more microprocessors (digital signal processors, DSPs), or one or more field programmable gate arrays (field programmable gate arrays, FPGAs).

[0194] Optionally, the first processor 2001 can execute various functions of the electronic device 410 by running or executing a software program stored in the memory 2002 and calling data stored in the memory 2002 .

[0195] In a specific implementation, as an embodiment, the first processor 2001 may include one or more CPUs, such as Figure 5 CPU0 and CPU1 are shown in FIG.

[0196] In a specific implementation, as an embodiment, the electronic device 410 may also include multiple processors, such as Figure 5 The first processor 2001 and the second processor 2004 are shown in FIG. Each of these processors may be a single-core processor (single-CPU) or a multi-core processor (multi-CPU). The processor here may refer to one or more devices, circuits, and / or processing cores for processing data (e.g., computer program instructions).

[0197] The memory 2002 is used to store the software program for executing the solution of the present invention, and is controlled to be executed by the first processor 2001. The specific implementation method can refer to the above method embodiment, which will not be repeated here.

[0198] Optionally, the memory 2002 may be a read-only memory (ROM) or other types of static storage devices that can store static information and instructions, a random access memory (RAM) or other types of dynamic storage devices that can store information and instructions, or an electrically erasable programmable read-only memory (EEPROM), a compact disc read-only memory (CD-ROM) or other optical disc storage, optical disc storage (including compressed optical disc, laser disc, optical disc, digital versatile disc, Blu-ray disc, etc.), a magnetic disk storage medium or other magnetic storage device, or any other medium that can be used to carry or store the desired program code in the form of instructions or data structures and can be accessed by a computer, but is not limited thereto. The memory 2002 may be integrated with the first processor 2001, or may exist independently and access the first processor 2001 through the interface circuit ( Figure 5 (not shown) is coupled to the first processor 2001, which is not specifically limited in this embodiment of the present invention.

[0199] The transceiver 2003 is used to communicate with a network device or a terminal device.

[0200] Optionally, the transceiver 2003 may include a receiver and a transmitter ( Figure 5 The receiver is used to implement a receiving function, and the transmitter is used to implement a sending function.

[0201] Optionally, the transceiver 2003 may be integrated with the first processor 2001, or may exist independently and communicate with the first processor 2001 through the interface circuit ( Figure 5 (not shown) is coupled to the first processor 2001, which is not specifically limited in this embodiment of the present invention.

[0202] It should be noted that Figure 5 The structure of the electronic device 410 shown in the figure does not constitute a limitation on the router, and the actual knowledge structure recognition device may include more or fewer components than those shown in the figure, or combine certain components, or arrange the components differently.

[0203] In addition, the technical effects of the electronic device 410 can refer to the technical effects of the intelligent pre-diagnosis method for ophthalmic diseases based on big data identification described in the above method embodiment, and will not be repeated here.

[0204] It should be understood that the first processor 2001 in the embodiment of the present invention may be a central processing unit (CPU), and the processor may also be other general-purpose processors, digital signal processors (DSP), application specific integrated circuits (ASIC), field programmable gate arrays (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor may be a microprocessor or the processor may also be any conventional processor, etc.

[0205] It should also be understood that the memory in the embodiments of the present invention may be a volatile memory or a non-volatile memory, or may include both volatile and non-volatile memories. Among them, the non-volatile memory may be a read-only memory (ROM), a programmable read-only memory (PROM), an erasable programmable read-only memory (EPROM), an electrically erasable programmable read-only memory (EEPROM), or a flash memory. The volatile memory may be a random access memory (RAM), which is used as an external cache. By way of example and not limitation, many forms of random access memory (RAM) are available, such as static RAM (SRAM), dynamic random access memory (DRAM), synchronous DRAM (SDRAM), double data rate synchronous dynamic random access memory (DDR SDRAM), enhanced synchronous dynamic random access memory (ESDRAM), synchronous link DRAM (SLDRAM), and direct rambus RAM (DR RAM).

[0206] The above embodiments can be implemented in whole or in part by software, hardware (such as circuits), firmware or any other combination. When implemented by software, the above embodiments can be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, the process or function described in the embodiment of the present invention is generated in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable system. The computer instructions can be stored in a computer-readable storage medium, or transmitted from one computer-readable storage medium to another computer-readable storage medium. For example, the computer instructions can be transmitted from one website, computer, server or data center to another website, computer, server or data center by wired (such as infrared, wireless, microwave, etc.). The computer-readable storage medium can be any available medium that can be accessed by a computer or a data storage device such as a server or data center that contains one or more available media sets. The available medium can be a magnetic medium (for example, a floppy disk, a hard disk, a tape), an optical medium (for example, a DVD), or a semiconductor medium. The semiconductor medium can be a solid-state hard disk.

[0207] It should be understood that the term "and / or" in this article is only a description of the association relationship of associated objects, indicating that there can be three relationships. For example, A and / or B can represent: A exists alone, A and B exist at the same time, and B exists alone. A and B can be singular or plural. In addition, the character " / " in this article generally indicates that the associated objects before and after are in an "or" relationship, but it may also indicate an "and / or" relationship. Please refer to the context for specific understanding.

[0208] In the present invention, "at least one" means one or more, and "plurality" means two or more. "At least one of the following" or similar expressions refers to any combination of these items, including any combination of single items or plural items. For example, at least one of a, b, or c can mean: a, b, c, ab, ac, bc, or abc, where a, b, and c can be single or multiple.

[0209] It should be understood that in various embodiments of the present invention, the size of the serial numbers of the above-mentioned processes does not mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present invention.

[0210] Those skilled in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of the present invention.

[0211] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the above-described devices, systems and units can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0212] In the several embodiments provided by the present invention, it should be understood that the disclosed devices, systems and methods can be implemented in other ways. For example, the system embodiments described above are only schematic. For example, the division of the units is only a logical function division. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another device, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of the system or unit, which can be electrical, mechanical or other forms.

[0213] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0214] In addition, each functional unit in each embodiment of the present invention may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit.

[0215] If the functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art or the part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium, including multiple instructions for enabling a computer device (which can be a personal computer, server, or network device, etc.) to perform all or part of the steps of the methods described in various embodiments of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM), random access memory (RAM), disk or optical disk, and other media that can store program codes.

[0216] The above is only a specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art can easily think of changes or substitutions within the technical scope disclosed by the present invention, which should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention should be based on the protection scope of the claims.

Claims

1. An intelligent pre-diagnosis method for ophthalmic diseases based on big data recognition, characterized in that: The method comprises: S1. Collect cataract CT image sets and build corresponding cataract pre-diagnosis models based on CNN training; S2. Collect big data on clinical diagnosis of cataracts, extract knowledge elements and associations related to cataracts based on the LLM large language model, and build a corresponding cataract LLM knowledge base; S3, obtaining the patient's cataract clinical examination data and importing them into the cataract pre-diagnosis model and the cataract LLM knowledge base respectively, predicting the corresponding first cataract pre-diagnosis strategy by the cataract pre-diagnosis model, and retrieving the corresponding second cataract pre-diagnosis strategy by the cataract LLM knowledge base; S4. Perform strategy fusion on the first cataract pre-diagnosis strategy and the second cataract pre-diagnosis strategy through the LLM large language model to generate and output a cataract treatment recommendation strategy.

2. The intelligent pre-diagnosis method for ophthalmic diseases based on big data recognition according to claim 1, characterized in that: S1, collecting cataract CT image sets, and building a corresponding cataract pre-diagnosis model based on CNN training, including: Collecting cataract CT images of multiple patients and performing preprocessing to obtain the cataract CT image set; Performing feature engineering on the cataract CT image set, and using CNN to extract cataract CT image features in each cataract CT image, including: orbital lateral wall / nodule soft tissue morphology, SUV value, and color and texture features; Performing feature annotation on the cataract CT image features, wherein the annotation information includes: the symptom stage corresponding to the cataract CT image features and the disease evolution trend thereof and the adopted clinical drug, surgery or a combination thereof treatment strategy; After labeling is completed, a feature set consisting of a plurality of cataract CT image features is obtained; Dividing the feature set into a training set and a validation set according to a preset ratio; Inputting the training set into a preset CNN model to perform feature training and learning to generate the cataract pre-diagnosis model; The prediction performance of the cataract pre-diagnosis model is verified using the validation set: If the verification is successful, the cataract pre-diagnosis model is deployed and applied; Otherwise, repeat the above steps to reconstruct the cataract pre-diagnosis model.

3. The intelligent pre-diagnosis method for ophthalmic diseases based on big data recognition according to claim 1, characterized in that: S2, collecting big data on clinical diagnosis of cataracts, extracting knowledge elements and associations related to cataracts based on the LLM large language model, and constructing a corresponding cataract LLM knowledge base, including: Collecting and preprocessing clinical diagnosis data of cataracts of multiple patients to obtain the clinical diagnosis big data of cataracts; According to the symptom stage, the cataract clinical diagnosis big data is divided into a plurality of sub-datasets corresponding to different symptom stages; Constructing entity keywords and onset association logic corresponding to different symptom stages respectively, and constructing large language model prompt words in combination with the entity keywords and the onset association logic, and inputting the large language model prompt words into a preset LLM large language model; Traversing each of the sub-datasets through the LLM large language model, identifying and extracting cataract knowledge elements in the sub-datasets and association relationships between the cataract knowledge elements based on the large language model prompt words, wherein the association relationships include pathogenic factors and treatment strategies related to the symptom characteristics of cataract patients, and the treatment strategies include clinical drugs, surgery, or a combination thereof; The cataract knowledge elements and their correlations in different symptom stages are sorted out through the LLM large language model to form a structured data set; The structured data set is written into a preset knowledge base through the LLM large language model to obtain the cataract LLM knowledge base, and the cataract LLM knowledge base is deployed and applied.

4. The intelligent pre-diagnosis method for ophthalmic diseases based on big data recognition according to claim 1, characterized in that: S3, obtaining the patient's cataract clinical examination data and importing them into the cataract pre-diagnosis model and the cataract LLM knowledge base respectively, predicting a corresponding first cataract pre-diagnosis strategy by the cataract pre-diagnosis model, and retrieving a corresponding second cataract pre-diagnosis strategy by the cataract LLM knowledge base, including: Pre-configure parsing rules and input the LLM large language model, use the LLM large language model to read the cataract clinical examination data of the patient stored in the HIS system, and parse the cataract clinical examination data based on the parsing rules to obtain the cataract knowledge elements in the cataract CT image and the cataract clinical diagnosis data, and import the cataract knowledge elements in the cataract CT image and the cataract clinical diagnosis data into the cataract pre-diagnosis model and the cataract LLM knowledge base respectively; Identifying cataract CT image features in the cataract CT image through the cataract pre-diagnosis model and predicting and outputting a first cataract pre-diagnosis strategy corresponding to the cataract CT image features, wherein the first cataract pre-diagnosis strategy includes the symptom stage of the cataract CT image features and the disease evolution trend thereof and the adopted clinical drug, surgery or a combination thereof treatment strategy; The second cataract prognosis strategy associated with the cataract knowledge element is retrieved and output through the cataract LLM knowledge base, wherein the second cataract prognosis strategy includes the cataract knowledge element and the pathogenic factors and treatment strategies associated therewith, and the treatment strategies include clinical drugs, surgery or a combination thereof.

5. The intelligent pre-diagnosis method for ophthalmic diseases based on big data recognition according to claim 4, characterized in that: The step S4, fusing the first cataract pre-diagnosis strategy and the second cataract pre-diagnosis strategy through the LLM large language model, generating and outputting a cataract treatment recommendation strategy, includes: Pre-build the strategy clustering logic and input it into the LLM large language model; The LLM large language model traverses the first cataract pre-diagnosis strategy and the second cataract pre-diagnosis strategy, clusters similar items in the first cataract pre-diagnosis strategy and the second cataract pre-diagnosis strategy based on the strategy clustering logic, and outputs corresponding clustering integration results, including: cataract symptom period and disease evolution trend, pathogenic factors and treatment strategies; Writing the clustering integration results into the patient's electronic medical record in the HIS system, generating and saving the corresponding cataract treatment recommendation strategy for the patient; Output and push the recommended cataract treatment strategy to the PC.

6. An intelligent pre-diagnosis system for ophthalmic diseases based on big data identification, the intelligent pre-diagnosis system for ophthalmic diseases based on big data identification is used to implement the intelligent pre-diagnosis method for ophthalmic diseases based on big data identification as claimed in any one of claims 1 to 5, characterized in that: The system comprises: Ophthalmological examination system, used to obtain patients' cataract clinical examination data and upload it to the backend server; A backend server is used to import the patient's cataract clinical examination data into a preset cataract pre-diagnosis model and a preset cataract LLM knowledge base, respectively, predict the corresponding first cataract pre-diagnosis strategy by the cataract pre-diagnosis model, and retrieve the corresponding second cataract pre-diagnosis strategy by the cataract LLM knowledge base; perform strategy fusion on the first cataract pre-diagnosis strategy and the second cataract pre-diagnosis strategy through the LLM large language model, generate and output a cataract treatment recommendation strategy; PC side, used for receiving and displaying the recommended cataract treatment strategy; The ophthalmological examination system and the PC are respectively connected to the backend server for communication.

7. An electronic device, characterized in that: The electronic device comprises: processor; A memory having computer-readable instructions stored thereon, wherein when the computer-readable instructions are executed by the processor, the method according to any one of claims 1 to 5 is implemented.

8. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores program codes, which can be called by a processor to execute the method according to any one of claims 1 to 5.

Images