A Blood Department Clinical Infection Source Detection System and Method Based on Big Data Analysis

Abstract

The present invention discloses a blood department clinical infection source detection system and method based on big data analysis, which is used in the field of medical detection. The blood department clinical infection source detection system includes: an image collection module, an image analysis module, an infection site prediction module, an infection source identification module, and a risk assessment module. By integrating advanced image analysis, disease identification, risk assessment, and data fusion technologies, the present invention realizes the accurate positioning of the patient's infection source and the division of risk levels, greatly improves the accuracy and efficiency of clinical detection, optimizes resource allocation, accelerates the treatment decision-making process, provides more personalized and timely medical intervention for patients, and significantly improves the treatment effect and safety of patients.

Description

Technical Field

[0001] The present invention relates to the field of medical detection. Specifically, it particularly relates to a system and method for detecting clinical infection sources in the hematology department based on big data analysis. Background Art

[0002] Clinics generally refer to medical practices directly related to patients, including the processes of diagnosing, treating, and preventing diseases. Clinical work is mainly carried out in hospitals or medical facilities, involving direct interactions between doctors, nurses, and other medical professionals and patients. Its goal is to apply medical theories and practices to solve specific health problems and improve the quality of life of patients.

[0003] Clinical hematology is a branch of clinical medicine that focuses on the diagnosis and treatment of diseases of the blood and its related organ systems (such as the bone marrow, spleen, lymphatic system, etc.). Problems dealt with by clinical hematologists include various types of anemia, blood coagulation disorders, leukemia, lymphoma, and other blood cancers. In addition, clinical hematology also involves the study of blood components such as red blood cells, white blood cells, and platelets.

[0004] An infection source refers to the origin of pathogens that can cause infections. These pathogens may include bacteria, viruses, fungi, parasites, etc. Infection sources can not only be diseased individuals, but may also be asymptomatic carriers, or certain objects and surfaces in the environment. Identifying infection sources is crucial for preventing and controlling the spread of infections, formulating treatment plans, and preventive measures.

[0005] For many patients with blood diseases, especially those undergoing chemotherapy or bone marrow transplantation, their immune system functions may be severely suppressed, making them more susceptible to infections. For hematology patients, common infections may rapidly develop into life-threatening situations. Therefore, early diagnosis and treatment are crucial. By accurately identifying the infection source, the most effective antibiotics or antiviral treatments can be selected for specific pathogens, thereby improving the effectiveness and efficiency of treatment. In hospitals and medical treatment environments, correctly identifying the infection source helps to take appropriate isolation and disinfection measures to prevent the spread of pathogens between patients and medical staff.

[0006] Traditional clinical diagnostic methods rely on manual interpretation of images and symptoms, which is not only time-consuming but also easily affected by doctors' experience and subjective judgment, resulting in limited diagnostic accuracy and efficiency. In the absence of rapid and accurate diagnostic support, formulating effective treatment plans may be delayed, affecting the treatment timing and treatment effect of patients. Traditional methods usually adopt a "one-size-fits-all" treatment plan, making it difficult to develop personalized treatment plans according to the specific conditions of patients. In the existing clinical decision-making process, rich medical knowledge bases are rarely utilized, resulting in treatment plans that may not be comprehensive enough and unable to fully utilize the latest medical research results.

[0007] In response to the problems in the related art, no effective solution has been proposed yet. Summary of the Invention

[0008] In order to overcome the above problems, the present invention aims to provide a blood department clinical infection source detection system and method based on big data analysis, aiming to solve the problem that traditional clinical diagnosis methods rely on manual interpretation of images and symptoms, which is not only time-consuming but also easily affected by doctors' experience and subjective judgment, resulting in limited diagnostic accuracy and efficiency. In the absence of rapid and accurate diagnostic support, the formulation of effective treatment plans may be delayed, affecting the treatment timing and treatment effect of patients.

[0009] To this end, the specific technical solutions adopted by the present invention are as follows:

[0010] According to one aspect of the present invention, there is provided a blood department clinical infection source detection system based on big data analysis. The blood department clinical infection source detection system includes: an image collection module, an image analysis module, an infection site prediction module, an infection source identification module, and a risk assessment module;

[0011] The image collection module, the image analysis module, the infection site prediction module, the infection source identification module, and the risk assessment module are sequentially connected.

[0012] The image collection module is used to collect clinical medical image and clinical medical infection image of patients from the blood department;

[0013] The image analysis module is used to analyze the collected clinical medical images and mark the infected image area based on the analysis results;

[0014] The infection site prediction module predicts the infection image site based on the clinical medical infection image;

[0015] The infection source identification module is used to combine the image features of the infected image area with the predicted infection image site and identify the patient's infection source image using the knowledge graph in the clinical medical field;

[0016] The risk assessment module evaluates the infection risk of the infection source image through big data analysis method in the clinical environment and classifies the infection risk into different levels;

[0017] Among them, when the infection site prediction module predicts the infection image site based on the clinical medical infection image, it includes:

[0018] Collect the clinical medical infection image of the patient from the blood department, extract the image features and patient information features, and combine the image features and patient information features to form a comprehensive feature set;

[0019] Screen the feature set, construct a classification model based on the screened features, introduce a shared weight matrix into the classification model, and design a loss function at the same time;

[0020] Use the gradient descent strategy to optimize the loss function and obtain the optimal classification model parameters;

[0021] Construct a multi-layer classification model based on the optimal classification model parameters;

[0022] Use the cross-validation method to evaluate the multi-layer classification model and verify the prediction accuracy of the model on the test set;

[0023] Input the new clinical medical infection image into the multi-layer classification model to predict the infection image location of the patient.

[0024] Optionally, when the image analysis module analyzes the collected clinical medical image and marks the infection image area based on the analysis result, it includes:

[0025] Obtain the clinical medical image of the patient collected from the hematology department and input the clinical medical image into a pre-configured convolutional neural network model;

[0026] The convolutional layer and pooling layer of the convolutional neural network model extract comprehensive image features associated with the infection image area;

[0027] Based on the comprehensive image features, the convolutional neural network model outputs a segmentation mask and visually displays the infection image area on the clinical medical image through color coding;

[0028] According to the bounding box coordinates and class labels of the infection image area, perform localization and classification to identify and mark the infection image area.

[0029] Optionally, screening the feature set, constructing a classification model based on the screened features, introducing a shared weight matrix into the classification model, and designing a loss function includes:

[0030] Preset the target variable, and use information gain to evaluate the association between each feature in the feature set and the target variable, and screen the feature subset most relevant to the target variable;

[0031] Use the screened feature subset to construct a classification model and adjust the parameters of the classification model;

[0032] Create a shared weight matrix and capture the interdependence of the infection image locations based on the shared weight matrix:

[0033] Integrate the shared weight matrix into the classification model and design a loss function including a regularization term.

[0034] Optionally, preset a target variable, and evaluate the association between each feature in the feature set and the target variable using information gain, and filter the feature subset most relevant to the target variable, including:

[0035] Preset the target variable for the infected image part, and convert all features in the feature set into a unified format;

[0036] Calculate the entropy of the target variable;

[0037] For each feature, calculate the entropy of the target variable under the condition of a given feature value;

[0038] Subtract the conditional entropy of the target variable after the given feature from the entropy of the target variable, and calculate the information gain of each feature;

[0039] Select the features with the information gain of each feature higher than the preset threshold, and combine the filtered features into a new feature subset;

[0040] Among them, the formula for calculating the entropy of the target variable is:

[0041] ;

[0042] In the formula, represents the probability that the target variable Y takes a specific value ;

[0043] represents the entropy of the target variable Y ;

[0044] Y represents the target variable.

[0045] Optionally, use the gradient descent strategy to optimize the loss function and obtain the best classification model parameters, including:

[0046] Initialize the parameters of the classification model;

[0047] Select an optimizer and set the learning rate;

[0048] Calculate the gradient of the loss function under the current parameters through the backpropagation algorithm;

[0049] Update the parameters of the classification model according to the calculated gradient and the set learning rate, and perform iteration in the direction of minimizing the loss;

[0050] Repeat the gradient calculation and parameter update until the parameters of the classification model reach the predetermined number of iterations.

[0051] Optionally, when the infection source identification module combines the imaging features of the infected image area with the predicted infected image part and uses the knowledge graph in the clinical medical field to identify the infected source image of the patient, it includes:

[0052] Obtain the imaging features of the infected image region and the predicted infected image location;

[0053] Collect knowledge in the medical field and construct a clinical medical knowledge graph;

[0054] Use entity extraction and entity linking techniques to connect the imaging features of the infected image region and the predicted infected image location with relevant entities in the knowledge graph;

[0055] Based on the imaging features of the infected image region, the predicted infected image location, and the information in the knowledge graph, construct a hybrid prediction model, introduce an attention mechanism, and identify the source image of the patient's infection.

[0056] Optionally, using entity extraction and entity linking techniques to connect the imaging features of the infected image region and the predicted infected image location with relevant entities in the knowledge graph includes:

[0057] Use natural language processing tools to extract medical entities from medical documents and image descriptions, where the medical entities include infection types, lesion sites, and symptom manifestations;

[0058] Compare and match the names of infection types, lesion sites, and symptom manifestations with a medical term library, and based on the matching results, convert the names of infection types, lesion sites, and symptom manifestations into the standardized forms in the standard term library;

[0059] Use graph embedding technology to accurately link the standardized medical entities with the corresponding entities in the constructed clinical medical knowledge graph and establish the association relationships between entities;

[0060] Among them, the use of graph embedding technology to accurately link the standardized medical entities with the corresponding entities in the constructed clinical medical knowledge graph and establish the association relationships between entities includes the following steps:

[0061] In the clinical medical knowledge graph, update or add corresponding nodes and the edges between them according to the standardized infection types, lesion sites, and symptom manifestations;

[0062] Select a graph embedding model according to the data scale, sparsity, and actual analysis requirements of infection types, lesion sites, and symptom manifestations;

[0063] Use the current graph structure to train the selected graph embedding model and encode the nodes in the clinical medical knowledge graph into vector representations;

[0064] Use the trained node vectors to calculate the similarity between the vector of the target medical entity and the existing node vectors in the graph;

[0065] Based on the vector similarity results, identify and link the most relevant nodes to ensure an accurate match between the standardized medical entities and the corresponding entities in the knowledge graph;

[0066] Based on the linking results, update or define new relationships between entities to strengthen and clarify the interactions and associations between nodes in the clinical medical knowledge graph.

[0067] Optionally, construct a hybrid prediction model based on the imaging features of the infected image region, the predicted infected image location, and the information in the knowledge graph, introduce an attention mechanism, and identify the source image of the patient's infection, including:

[0068] Integrate and normalize the imaging features of the infected image region, the predicted infected image location, and the information in the knowledge graph;

[0069] Through the fusion layer technology in the deep learning network, create a comprehensive hybrid prediction model by combining imaging and semantic features;

[0070] Design and apply a multi-head attention mechanism to evaluate and optimize the importance of each feature in identifying the source of infection;

[0071] Train the model using the labeled dataset and optimize the model parameters through cross-validation techniques;

[0072] Test the model on an independent validation set and conduct a comprehensive evaluation using multiple performance metrics.

[0073] Optionally, the risk assessment module evaluates the infection risk of the source image of the infection through big data analysis methods in a clinical environment, and includes the following when classifying the infection risk into risk levels:

[0074] Determine the key indicators for measuring the infection risk;

[0075] Analyze the source of infection using statistical methods and evaluate the infection risk of the source of infection under various conditions based on the key indicators;

[0076] Preset level thresholds at each stage of the infection risk, and classify the infection risk based on the preset level thresholds.

[0077] According to another aspect of the present invention, there is also provided a method for detecting the source of clinical infection in the hematology department based on big data analysis. The method for detecting the source of clinical infection in the hematology department includes:

[0078] S1. Collect the clinical medical imaging images and clinical medical infection images of patients from the hematology department;

[0079] S2. Analyze the collected clinical medical imaging images and mark the infected image regions based on the analysis results;

[0080] S3. Predict the location of the infection image based on the clinical medical infection image;

[0081] S4. Combine the imaging features of the infection image area with the predicted location of the infection image, and use the knowledge graph in the clinical medical field to identify the infection source image of the patient;

[0082] S5. Evaluate the infection risk of the infection source image through big data analysis in the clinical environment, and classify the infection risk into different levels.

[0083] Compared with the prior art, the present application has the following beneficial effects:

[0084] 1. By integrating advanced imaging analysis, disease identification, risk assessment, and data fusion technologies, the present invention realizes the accurate positioning and risk level classification of the patient's infection source, greatly improves the accuracy and efficiency of clinical detection, optimizes resource allocation, accelerates the treatment decision-making process, provides more personalized and timely medical intervention for patients, and significantly improves the treatment effect and safety of patients.

[0085] 2. The present invention integrates an infection site prediction module that combines clinical medical infection image features and patient information. Through fine data processing and advanced machine learning technologies (such as XGBoost and shared weight matrices), it effectively improves the accuracy of identifying and predicting the infection site; it can not only accurately capture the specific location and nature of the infection, but also further refine the feature selection and model training process through information gain and optimized loss functions. Finally, it realizes efficient clinical decision support. Using cross-validation and multi-layer classification technologies, it ensures the robustness and wide applicability of the model, significantly improves the speed and accuracy of medical detection, and provides a more accurate and personalized treatment plan for patients.

[0086] 3. Through the comprehensive application of advanced imaging feature extraction and knowledge graph, the present invention effectively identifies and predicts the infection source image. Using a hybrid prediction model and an attention mechanism, it accurately analyzes and identifies the infection source. At the same time, combined with the extensive knowledge base in the clinical medical field, it enhances the accuracy and reliability of detection. In addition, through entity linking technology, it tightly combines the imaging features of the infection area with relevant clinical data, making the model not just a simple image recognition, but an intelligent system that comprehensively considers the clinical context, greatly improving the doctor's understanding and handling ability of infectious diseases, optimizing the treatment strategy, and improving the treatment effect of patients. BRIEF DESCRIPTION OF THE DRAWINGS

[0087] With the following description of the embodiments in conjunction with the accompanying drawings, the above characteristics, features, and advantages of the present invention and the implementation methods and means thereof become more clearly understandable. The embodiments are described in detail in conjunction with the accompanying drawings. Shown herein in schematic diagrams:

[0088] Figure 1It is a schematic block diagram of a blood department clinical infection source detection system based on big data analysis according to an embodiment of the present invention;

[0089] Figure 2 It is a flowchart of a blood department clinical infection source detection method based on big data analysis according to an embodiment of the present invention.

[0090] In the figure:

[0091] 1. Image collection module; 2. Image analysis module; 3. Infection site prediction module; 4. Infection source identification module; 5. Risk assessment module. Detailed implementation manners

[0092] In order to enable those skilled in the art of the present technology to better understand the solution of this application, the technical solutions in the embodiments of this application will be clearly and completely described below in conjunction with the accompanying drawings in the embodiments of this application. Obviously, the described embodiments are only a part of the embodiments of this application, rather than all the embodiments. Based on the embodiments in this application, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of this application.

[0093] According to an embodiment of the present invention, a blood department clinical infection source detection system and method based on big data analysis are provided.

[0094] Now, the present invention will be further described in conjunction with the accompanying drawings and specific implementation manners. As Figure 1 shown, according to an embodiment of the present invention, a blood department clinical infection source detection system is provided. The blood department clinical infection source detection system includes: an image collection module 1, an image analysis module 2, an infection site prediction module 3, an infection source identification module 4, and a risk assessment module 5;

[0095] The image collection module 1, the image analysis module 2, the infection site prediction module 3, the infection source identification module 4, and the risk assessment module 5 are sequentially connected.

[0096] The image collection module 1 is used to collect clinical medical image and clinical medical infection image of patients from the blood department.

[0097] It should be explained that the clinical medical images include the following types:

[0098] 1) Ray: used to detect fractures, lung diseases, etc.

[0099] 2) CT scan (Computed Tomography): provides detailed cross-sectional images of the internal structure of the body, and is often used to detect tumors, bleeding, trauma, etc.

[0100] 3) MRI scan (Magnetic Resonance Imaging): Used to show in detail the soft tissues in the body, such as the brain, spinal cord, internal organs, etc.

[0101] 4) Ultrasound examination: Often used to observe pregnancy, blood vessel conditions, abdominal organs, etc.

[0102] Clinical medical infection images include the following types:

[0103] 1) X-ray or CT scan of the infected area: For example, pneumonia may show abnormal shadow areas in X-ray images.

[0104] 2) MRI or ultrasound: Used to examine soft tissue infections or infections of internal organs (such as the liver, kidneys).

[0105] 3) PET scan (Positron Emission Tomography): Sometimes used to find the source of infection, especially in complex cases, which can help locate potential infection areas.

[0106] 4) Angiography: If the infection involves the vascular system, such as sepsis or vasculitis.

[0107] Through the image collection module, these clinical medical images can be collected and sorted out. For infection images, doctors and researchers can analyze more deeply the type, severity and spread of the infection, as well as its correlation with other medical conditions of the patient.

[0108] The image analysis module 2 is used to analyze the collected clinical medical images and mark the infection image area based on the analysis results.

[0109] Preferably, when the image analysis module 2 analyzes the collected clinical medical images and marks the infection image area based on the analysis results, it includes:

[0110] Obtain the clinical medical images of the patient collected from the hematology department and input the clinical medical images into a pre-configured convolutional neural network model;

[0111] The convolutional layer and pooling layer of the convolutional neural network model extract the comprehensive image features associated with the infection image area;

[0112] Based on the comprehensive image features, the convolutional neural network model outputs a segmentation mask and visually shows the infection image area on the clinical medical images through color coding;

[0113] According to the bounding box coordinates and class labels of the infection image area, perform localization and classification to identify and mark the infection image area.

[0114] It should be noted that clinical medical images of patients are collected from the hematology department and input into a pre-configured convolutional neural network model to ensure that all images to be analyzed can be correctly processed by the model;

[0115] The convolutional layer and pooling layer in the convolutional neural network are used to extract comprehensive image features related to infection from the images. The convolutional layer is responsible for capturing local features in the image, while the pooling layer is used to reduce the feature dimension, thereby extracting higher-level abstract features.

[0116] Based on the comprehensive image features, the network outputs a segmentation mask to distinguish the infected area from the non-infected area, and visually displays the infected area on the original image through color coding, enhancing the visual expression of the results.

[0117] According to the marked infected image areas in the segmentation mask, precise localization and classification are performed using bounding box coordinates and class labels. This ensures that not only the infected areas are identified, but also correctly classified, providing important information for clinical detection.

[0118] The infection site prediction module 3 predicts the infection image site based on clinical medical infection images.

[0119] When the infection site prediction module 3 predicts the infection image site based on clinical medical infection images, it includes:

[0120] Clinical medical infection images of patients are collected from the hematology department, and image features and patient information features are extracted. The image features and patient information features are combined to form a comprehensive feature set.

[0121] It should be noted that clinical medical infection images of patients are collected from the hematology department. Clinical medical infection images include images generated by imaging techniques such as X-rays, CTs, and MRIs, which are specifically used to show the details of the infected area. Image processing techniques (such as edge detection, texture analysis, and morphological operations) are used to extract key features in the images. The key features can depict information such as the shape, size, location of the infected area, and the contrast with surrounding tissues. At the same time, the basic information and relevant medical data of the patients, such as age, gender, medical history, and laboratory test results, are collected. These information can be obtained from the electronic medical record (EMR) and are the key to understanding the overall health status of the patients. The comprehensive feature set can be used to support clinical decisions, such as infection severity assessment and treatment plan selection.

[0122] The feature set is screened, a classification model is constructed based on the screened features, a shared weight matrix is introduced into the classification model, and a loss function is designed at the same time.

[0123] Preferably, screen the feature set, construct a classification model based on the screened features, and introduce a shared weight matrix into the classification model. At the same time, design the loss function to include:

[0124] Preset the target variable, and use information gain to evaluate the association between each feature in the feature set and the target variable, and screen the feature subset that is most relevant to the target variable;

[0125] Use the screened feature subset to construct a classification model and adjust the parameters of the classification model;

[0126] Create a shared weight matrix and capture the interdependencies of the infected image parts based on the shared weight matrix:

[0127] Integrate the shared weight matrix into the classification model and design a loss function that includes a regularization term.

[0128] Preferably, preset the target variable, and use information gain to evaluate the association between each feature in the feature set and the target variable. The feature subset that is most relevant to the target variable includes:

[0129] Preset the target variable of the infected image part and convert all features in the feature set into a unified format;

[0130] Calculate the entropy of the target variable;

[0131] For each feature, calculate the entropy of the target variable under the condition of a given feature value;

[0132] Subtract the conditional entropy of the target variable after a given feature from the entropy of the target variable to calculate the information gain of each feature;

[0133] Select the features whose information gain of each feature is higher than the preset threshold, and combine the screened features into a new feature subset;

[0134] Among them, the formula for calculating the entropy of the target variable is:

[0135] ;

[0136] In the formula, represents the probability that the target variable Y takes a specific value ;

[0137] represents the target variable Y entropy;

[0138] Y represents the target variable.

[0139] The formula for calculating the entropy of the target variable under the condition of a given feature value is:

[0140] ;

[0141] In the formula, represents the conditional entropy of the target variable X after the value of the known variable Y is known;

[0142] represents the probability that the variable X takes a specific value x j ;

[0143] represents the entropy of the target variable X when the variable x j is equal to a certain value; Y is the entropy of the target variable

[0144] x j represents a specific value of the variable X ;

[0145] i , y both represent indices.

[0146] It should be noted that first, the infected image part is set as the target variable, which will guide the training process of the entire model, ensuring that all features have a unified format before analysis. This is an important step in data preprocessing, which helps to avoid biases in subsequent analysis. Calculate the entropy of the target variable to obtain its initial information state. For each feature, calculate the conditional entropy of the target variable under the condition of the feature value. Calculate the information gain by subtracting the conditional entropy from the entropy of the target variable to measure the amount of information provided by each feature for the target variable. Select the features with information gain higher than the preset threshold and form a new feature subset. The features selected in this way are more critical for predicting the target variable.

[0147] Construct a classification model using the selected feature subset, and the classification model is an XGBoost classification model. XGBoost is an efficient gradient boosting algorithm widely used in classification tasks. Its advantage lies in the speed and effect when dealing with large-scale data sets. Adjust the parameters of the XGBoost model through methods such as cross-validation to find the optimal parameter configuration to improve the model performance; design a shared weight matrix to capture the interdependencies between different infected image parts, which can help the model be more accurate when dealing with multi-label classification problems. Integrate the shared weight matrix into the XGBoost model so that the model can consider the correlations between different labels during the training process. Design a loss function containing a regularization term to optimize the XGBoost model. The regularization term helps to control the model complexity and avoid overfitting, so that the model performs better on unseen data.

[0148] A method that combines information gain, XGBoost, a shared weight matrix, and a regularization loss function provides a powerful framework to enhance the performance and robustness of the model when dealing with complex clinical imaging data.

[0149] Use the gradient descent strategy to optimize the loss function and obtain the optimal classification model parameters.

[0150] Preferably, using the gradient descent strategy to optimize the loss function and obtain the optimal classification model parameters includes:

[0151] Initialize the parameters of the classification model;

[0152] Select an optimizer and set the learning rate;

[0153] Calculate the gradient of the loss function under the current parameters through the backpropagation algorithm;

[0154] Update the parameters of the classification model according to the calculated gradient and the set learning rate, and iterate in the direction of minimizing the loss;

[0155] Repeat the gradient calculation and parameter update until the parameters of the classification model reach a predetermined number of iterations.

[0156] It should be explained that model parameters (such as weights and biases) are usually initialized as small random numbers or using specific initialization methods (such as He or Xavier initialization) to help the algorithm converge more effectively; common optimizers include SGD (stochastic gradient descent), Adam, RMSprop, etc. Different optimizers are suitable for different application scenarios. Among them, Adam is popular due to its adaptive learning rate adjustment. The learning rate is a hyperparameter that controls the step size of parameter updates. Setting an appropriate learning rate is crucial. Too high may lead to instability during training, and too low may lead to slow training speed. Use the backpropagation algorithm to calculate the gradient of the loss function with respect to the model parameters. This process involves a forward pass of the model to calculate the loss and then a backward pass of the gradient to obtain the gradient of the parameters. According to the calculated gradient and the set learning rate, update the model parameters. The update rule is usually that the new value of the parameter is equal to the original value minus the learning rate multiplied by the gradient. This step is performed iteratively with the goal of gradually reducing the loss of the model. Repeat the gradient calculation and parameter update process until a predetermined number of iterations is reached, or stop iterating when the performance of the model no longer improves significantly. Usually, the performance on the validation set is also monitored to avoid overfitting of the model.

[0157] Construct a multi-layer classification model based on the optimal classification model parameters.

[0158] Use the cross-validation method to evaluate the multi-layer classification model and verify the prediction accuracy of the model on the test set.

[0159] Input the new clinical medical infection image into the multi-layer classification model to predict the infection image location of the patient.

[0160] It should be noted that based on the best classification model parameters obtained from previous optimization, a multi-layer classification model is constructed; the multi-layer classification model includes multiple hidden layers, and each layer uses activation functions such as ReLU or Sigmoid to increase the non-linear processing ability of the model; the cross-validation method is used to evaluate the performance of the multi-layer classification model; this usually includes dividing the data set into several subsets (such as five-fold or ten-fold cross-validation), where each subset is used as the test set in turn, and the rest are used as the training set to ensure that the model can be effectively verified on different data subsets, select appropriate evaluation metrics such as accuracy, recall rate, F1 score, etc. to comprehensively evaluate the performance of the model in all aspects, verify the prediction accuracy of the multi-layer classification model on an independent test set, input the newly collected clinical medical infection image into the trained multi-layer classification model to predict the infection image location of the patient, which can help doctors quickly identify the infection area and thus start treatment earlier.

[0161] The infection source identification module 4 is used to combine the image features of the infection image area with the predicted infection image location and use the knowledge graph in the clinical medical field to identify the infection source image of the patient.

[0162] Preferably, when the infection source identification module 4 combines the image features of the infection image area with the predicted infection image location and uses the knowledge graph in the clinical medical field to identify the infection source image of the patient, it includes:

[0163] Obtain the image features of the infection image area and the predicted infection image location.

[0164] Collect knowledge in the medical field and construct a clinical medical knowledge graph.

[0165] It should be noted that, first, the infected image area is preprocessed, including denoising, enhancing contrast, etc., to improve the image quality. Image analysis techniques, such as edge detection, texture analysis, color analysis, etc., are used to extract the key image features of the infected area. The key image features include area size, shape, boundary clarity, texture uniformity, etc. The extracted image features are input into the previously trained classification model to predict the specific location of the infected image; Information about diseases, symptoms, treatment methods, etc. is collected from sources such as medical literature, clinical trial records, online medical databases, etc. The collected information is sorted out, entities (such as diseases, symptoms, drugs, treatment methods) and the relationships between entities (such as the association between symptoms and diseases) are defined, and knowledge graph technology is used to convert the structured data into a graph database, where nodes represent entities and edges represent the relationships between entities. In this way, a complex network can be formed to support clinical decision-making and research. The knowledge graph is applied to support clinical decision-making, such as personalized medical advice, etc. At the same time, the knowledge graph can also be used for education and research to help medical professionals and scholars better understand complex medical information and knowledge structures.

[0166] The image features of the infected image area and the predicted location of the infected image are connected to the relevant entities in the knowledge graph by using entity extraction and entity linking techniques.

[0167] Preferably, connecting the image features of the infected image area and the predicted location of the infected image to the relevant entities in the knowledge graph by using entity extraction and entity linking techniques includes:

[0168] Using natural language processing tools to extract medical entities from medical documents and image descriptions. Among them, medical entities include infection types, lesion sites, and symptom manifestations.

[0169] It should be noted that before entity extraction, the medical documents and image descriptions need to be preprocessed first, including text cleaning (removing irrelevant characters, punctuation marks, etc.), format unification, encoding conversion, etc. Regular expressions are used to remove or replace special characters and noises in the text. For example, the re library in Python can be used for this purpose. Natural language processing tools are used to identify medical entities in the text, such as infection types, lesion sites, and symptom manifestations. Entities are usually proper nouns or technical terms. The identified medical entities may have multiple expressions, and these entities need to be converted into standardized medical terms, usually by matching to an authoritative medical term library (such as ICD-10, SNOMED CT, etc.), and mapping the extracted entities to the standard terms.

[0170] The names of the infection types, lesion sites, and symptom manifestations are compared and matched with the medical term library. Based on the matching results, the names of the infection types, lesion sites, and symptom manifestations are converted into the normative forms in the standard term library.

[0171] It should be noted that a suitable medical terminology library is selected, such as ICD-10, SNOMED CT or UMLS. These medical terminology libraries contain a wide range of medical terms and their standardized forms. The most appropriate terminology library is selected according to the data requirements and project goals. The extracted infection types, lesion sites, and symptom names are matched with the terms in the selected medical terminology library. String matching techniques or more advanced semantic matching tools, such as natural language processing algorithms, are used to find the closest terms. Once the matching terms are found, the extracted non-standard medical entity names are converted into the canonical forms in the terminology library.

[0172] Using graph embedding technology, the standardized medical entities are accurately linked to the corresponding entities in the constructed clinical medical knowledge graph, and the association relationships between entities are established;

[0173] Among them, the use of graph embedding technology to accurately link the standardized medical entities to the corresponding entities in the constructed clinical medical knowledge graph and establish the association relationships between entities includes the following steps:

[0174] In the clinical medical knowledge graph, according to the standardized infection types, lesion sites, and symptom manifestations, update or add the corresponding nodes and the edges between them;

[0175] According to the data scale, sparsity of the infection types, lesion sites, and symptom manifestations, and the actual analysis requirements, select a graph embedding model;

[0176] Using the current graph structure, train the selected graph embedding model to encode the nodes in the clinical medical knowledge graph into vector representations;

[0177] Using the trained node vectors, calculate the similarity between the vector of the target medical entity and the existing node vectors in the graph;

[0178] According to the vector similarity results, identify and link the most relevant nodes to ensure the accurate matching of the standardized medical entities with the corresponding entities in the knowledge graph;

[0179] Based on the link results, update or define new entity relationships to strengthen and clarify the interactions and associations between the nodes in the clinical medical knowledge graph.

[0180] It should be noted that according to the infection types, lesion sites, and symptom manifestations extracted and standardized from the documents, the corresponding nodes and edges are updated or added to the clinical medical knowledge graph. A graph database management system, such as Neo4j or ArangoDB, is used to manage and update the graph; nodes represent medical entities, and edges represent the relationships between entities, such as correlation or causal relationships. In the graph, the accurate representation of nodes and edges is crucial for subsequent analysis and linking.

[0181] Select a suitable graph embedding model according to the scale, sparsity, and analysis requirements of the entity data. Evaluate different graph embedding models such as Node2Vec, GraphSAGE, or DeepWalk. The graph embedding model can transform the nodes in the graph structure into vectors in a low-dimensional space. The vectors can capture the topological relationships and node attributes between nodes. Use the current structure of the clinical medical knowledge graph to train the selected graph embedding model. Input the graph data into the model for training, and use the node and edge information in the graph structure to generate the vector representation of each node. During the training process, the model learns the optimal vector representation of the nodes by optimizing an objective function, which usually involves maximizing the similarity between neighboring nodes. Use the trained node vectors to calculate the similarity between the vector of the target medical entity and the existing node vectors in the graph. According to the vector similarity results, identify and link the most relevant nodes to ensure the accurate matching of the medical entity with the corresponding entity in the knowledge graph. Select the node with the highest similarity score as the matching node, and create or update the links between nodes in the graph. Based on the link results, update or define new relationships between entities, strengthen and clarify the interactions and correlations between nodes in the clinical medical knowledge graph. According to the newly discovered links between entities, update the attributes or types of the edges in the graph, such as adding new causal relationship or correlation labels.

[0182] Suppose data on the infection type "pneumonia" is extracted from a batch of medical documents, including the common lesion site "lower respiratory tract" and symptom manifestations "cough" and "fever".

[0183] The specific example steps are as follows:

[0184] 1) Update or add nodes and edges:

[0185] In the clinical medical knowledge graph, add nodes such as "pneumonia", "lower respiratory tract", "cough", and "fever".

[0186] Establish edges between nodes. For example, connect "pneumonia" and "lower respiratory tract" to represent the lesion site; connect "pneumonia" with "cough" and "fever" to represent the symptom manifestations.

[0187] 2) Select a graph embedding model:

[0188] Given that the entities involved in the dataset are relatively clear and closely related, the Node2Vec model is selected for graph embedding to optimize the structural and functional expressions between nodes.

[0189] 3) Train the graph embedding model:

[0190] Input the graph data constructed above into the Node2Vec model and train the model to obtain the vector representation of each node.

[0191] 4) Calculate vector similarity:

[0192] Use the cosine similarity calculation method to compare the similarity between the "pneumonia" node vector and the node vectors of other diseases or symptoms in the knowledge graph.

[0193] 5) Identify and link the most relevant nodes:

[0194] Identify other respiratory infection disease nodes that are highly similar to "pneumonia", such as "bronchitis", and link these nodes to indicate that they have similar symptoms or lesion sites.

[0195] 6) Update or define new relationships between entities:

[0196] In the knowledge graph, update the attributes of the edges based on the new link relationships, such as adding labels of "possible complications" or "common treatment methods" to strengthen the practical relevance between nodes. Construct a hybrid prediction model based on the imaging features of the infected image region, the predicted infected image site, and the information in the knowledge graph, introduce the attention mechanism, and identify the source image of the patient's infection.

[0197] Preferably, constructing a hybrid prediction model based on the imaging features of the infected image region, the predicted infected image site, and the information in the knowledge graph, introducing the attention mechanism, and identifying the source image of the patient's infection includes:

[0198] Integrate and normalize the imaging features of the infected image region, the predicted infected image site, and the information in the knowledge graph;

[0199] Create a comprehensive hybrid prediction model by combining imaging and semantic features through the fusion layer technology in the deep learning network;

[0200] Design and apply a multi-head attention mechanism to evaluate and optimize the importance of each feature in identifying the source of infection;

[0201] Train the model using the labeled dataset and optimize the model parameters through cross-validation techniques;

[0202] Test the model on an independent validation set and conduct a comprehensive evaluation using multiple performance metrics.

[0203] It should be noted that, first of all, the imaging features of the infected image area, the predicted infected image parts, and the information in the knowledge graph are integrated together. Since the data sources have different scales and formats, they need to be uniformly processed so that the model can effectively learn. The integrated data is normalized to ensure the stability and efficiency of model training. Normalization usually includes scaling the feature values to the same numerical range, such as between 0 and 1 or -1 and 1.

[0204] Use the fusion layer technology in the deep learning network to combine the imaging and semantic features. This can be achieved in various ways, such as concatenation fusion, weighted fusion, etc. The purpose is to create a comprehensive model that can simultaneously understand the imaging features and semantic information. Design the model architecture according to requirements, including selecting appropriate deep learning frameworks and network structures (such as CNN, RNN, Transformer, etc.).

[0205] Design and apply the multi-head attention mechanism, which enables the model to focus on the features that are particularly important for identifying the source image of the infection. The multi-head attention mechanism allows the model to analyze the input data from multiple perspectives, enhancing the model's ability to capture key information. Evaluate the importance of each feature in identifying the source of the infection through attention scores, thereby optimizing the interpretability and performance of the model.

[0206] The risk assessment module 5 evaluates the infection risk of the source image of the infection through big data analysis in a clinical environment and classifies the infection risk into different levels;

[0207] Preferably, when the risk assessment module evaluates the infection risk of the source image of the infection through big data analysis in a clinical environment and classifies the infection risk into different levels, it includes:

[0208] Determine the key indicators for measuring the infection risk;

[0209] Analyze the source of the infection using statistical methods and evaluate the infection risk of the source of the infection under various conditions based on the key indicators;

[0210] Preset level thresholds at each stage of the infection risk and classify the infection risk based on the preset level thresholds.

[0211] It should be noted that first, key indicators that can reflect the severity of infection and the potential for transmission are determined. The key indicators may include the size and depth of the infected area, the type of infection (such as bacteria, virus), known drug resistance characteristics, etc. Data on historical infection cases are collected, including imaging features, treatment outcomes, patient responses, etc. Statistical analysis methods, such as regression analysis and survival analysis, are applied to evaluate the infection risks of different infection sources under various conditions, helping to identify which factors are most likely to lead to severe infection outcomes. According to the severity of the infection risk, different risk level thresholds are preset; for example, low, medium, and high risk assessment criteria can be set. High risk may require immediate intervention, while low risk is subject to routine monitoring. Based on the preset level thresholds and the results of statistical analysis, the risks of infection sources are classified, which needs to be implemented on a big data platform to process and analyze a large amount of clinical data. The risk assessment results can directly support medical decision-making, helping doctors choose the most appropriate treatment plan. High-risk infections may require more medical resources and emergency responses, and through risk level classification, hospitals can allocate resources more effectively.

[0212] Actual application example: Pneumonia infection detection and risk assessment, steps are as follows:

[0213] The following is a hypothetical application scenario, the scenario is as follows: A patient is admitted to the hospital due to acute respiratory discomfort. The doctor suspects pneumonia and uses X-ray and CT scans to collect the patient's chest images;

[0214] Operation 1: The image collection module automatically obtains the latest patient images from the hospital's imaging system; analyzes the collected chest X-ray and CT scans, and analyzes the patient's images to determine whether there are signs of infection;

[0215] Operation 2: Through a pre-configured deep learning model (such as a convolutional neural network), analyze the images, identify and mark the areas showing abnormal shadows and typical infection signs; predict the specific infection site based on the marked imaging features;

[0216] Operation 3: The module analyzes the details in the images, such as the size, location, and shape of the shadows, combines the patient's clinical information (such as age, past medical history) to predict the specific site and severity of the infection, and determines the specific type of infection, such as bacterial pneumonia or viral pneumonia;

[0217] Operation 4: Use the features of the infected image area in combination with the knowledge graph to identify possible pathogens; for example, by comparing the image features with historical data, the module may predict that the infection is Streptococcus pneumoniae; evaluate the risk level of this pneumonia infection to provide a basis for treatment decisions;

[0218] Operation 5: Based on information such as the location, size, patient age, and comorbidities of the infection, evaluate the risk level of the infection; using big data analysis, determine that the infection is of medium risk, recommend hospitalization and the use of antibiotics.

[0219] According to another embodiment of the present invention, as Figure 2 shown, there is also provided a method for detecting the source of clinical infection in the hematology department based on big data analysis. This method for detecting the source of clinical infection in the hematology department includes:

[0220] S1. Collect the clinical medical imaging images and clinical medical infection images of patients from the hematology department;

[0221] S2. Analyze the collected clinical medical imaging images and mark the infected image area based on the analysis results;

[0222] S3. Predict the location of the infected image based on the clinical medical infection image;

[0223] S4. Combine the image features of the infected image area with the predicted location of the infected image, and use the knowledge graph in the clinical medical field to identify the source image of the patient's infection;

[0224] S5. Evaluate the infection risk of the source image of the infection through big data analysis in a clinical environment and classify the infection risk into risk levels.

[0225] In summary, by means of the above technical solutions of the present invention, the present invention integrates advanced imaging analysis, disease recognition, risk assessment and data fusion technologies, realizes the accurate positioning of the patient's infection source and the division of risk levels, greatly improves the accuracy and efficiency of clinical detection, optimizes resource allocation, accelerates the treatment decision-making process, provides more personalized and timely medical intervention for patients, and significantly improves the treatment effect and safety of patients; the present invention integrates an infection site prediction module that combines clinical medical infection image features and patient information, and through fine data processing and advanced machine learning technologies (such as XGBoost and shared weight matrices), effectively improves the accuracy of identifying and predicting the infection site; it can not only accurately capture the specific location and nature of the infection, but also further refine the feature selection and model training process through information gain and optimized loss functions, and finally achieve efficient clinical decision support. Using cross-validation and multi-layer classification technologies, it ensures the robustness and wide applicability of the model, significantly improves the speed and accuracy of medical detection, and provides a more accurate and personalized treatment plan for patients; the present invention effectively identifies and predicts the infection source image through the comprehensive application of advanced image feature extraction and knowledge graph, uses a hybrid prediction model and an attention mechanism to accurately analyze and identify the infection source, and at the same time combines a wide knowledge base in the clinical medical field to enhance the accuracy and reliability of detection. In addition, through entity linking technology, the image features of the infected area are closely combined with relevant clinical data, making the model not just a simple image recognition, but an intelligent system that comprehensively considers the clinical context, greatly improving the doctor's understanding and handling ability of infectious diseases, optimizing the treatment strategy, and improving the treatment effect of patients.

[0226] Although the present invention has been disclosed above with preferred embodiments, the said embodiments are only for the convenience of illustration and are given by way of example, and are not intended to limit the present invention. Those skilled in the art can make several modifications and refinements without departing from the spirit and scope of the present invention. The scope of protection claimed by the present invention shall be subject to what is described in the claims.

Claims

1. A hematology clinical infection source detection system based on big data analysis, characterized in that: include: Image collection module; An image analysis module, used to analyze the collected clinical medical images and mark the infected image areas based on the analysis results; Infection site prediction module, which predicts the infection site based on clinical medical infection images; The infection source identification module is used to combine the image features of the infected image area with the predicted infected image location, and use the knowledge graph in the clinical medical field to identify the patient's infection source image; Risk assessment module; Wherein, the prediction of infection image parts based on clinical medical infection images includes: Collect clinical medical infection images of patients from the hematology department, extract image features and patient information features, and combine the image features and patient information features to form a comprehensive feature set; Screening feature sets, building a classification model based on the screened features, introducing a shared weight matrix into the classification model, and designing a loss function; including: presetting the target variable, using information gain to evaluate the association between each feature in the feature set and the target variable, and screening the feature subset most relevant to the target variable; using the screened feature subset to build a classification model, and adjusting the parameters of the classification model; creating a shared weight matrix, and capturing the interdependence of infected image parts based on the shared weight matrix: integrating the shared weight matrix into the classification model, and designing a loss function containing a regularization term; Use the gradient descent strategy to optimize the loss function and obtain the optimal classification model parameters; Construct a multi-layer classification model based on the optimal classification model parameters; The multi-layer classification model was evaluated using the cross-validation method and the prediction accuracy of the model was verified on the test set; New clinical medical infection images are input into the multi-layer classification model to predict the patient's infection image location.

2. According to claim 1, a hematology clinical infection source detection system based on big data analysis is characterized in that: The image analysis module analyzes the collected clinical medical images and marks the infected image area based on the analysis results, including: Acquire clinical medical imaging images of patients collected from a hematology department, and input the clinical medical imaging images into a pre-configured convolutional neural network model; The convolutional and pooling layers of the convolutional neural network model extract comprehensive image features associated with infected image regions; Based on the comprehensive image features, the convolutional neural network model outputs a segmentation mask and intuitively displays the infected image area on the clinical medical image through color coding; According to the bounding box coordinates and category labels of the infected image area, positioning and classification are performed to identify and mark the infected image area.

3. A hematology clinical infection source detection system based on big data analysis according to claim 2, characterized in that: The preset target variable and the information gain are used to evaluate the association between each feature in the feature set and the target variable, and the feature subset most relevant to the target variable is selected, including: Preset the target variable of the infected image part and convert all features in the feature set into a unified format; Calculate the entropy of the target variable; For each feature, calculate the entropy of the target variable given the feature value; The information gain of each feature is calculated by subtracting the conditional entropy of the target variable after the given feature from the entropy of the target variable; Select the features whose information gain of each feature is higher than the preset threshold, and combine the screened features into a new feature subset; Wherein, the formula for calculating the entropy of the target variable is: ； In the formula, Represents the target variable Y Get a specific value probability; Represents the target variable Y Entropy of Y represents the target variable.

4. A hematology clinical infection source detection system based on big data analysis according to claim 3, characterized in that: The method of optimizing the loss function using the gradient descent strategy and obtaining the best classification model parameters includes: Initialize the parameters of the classification model; Select an optimizer and set the learning rate; Calculate the gradient of the loss function under the current parameters through the back-propagation algorithm; Update the parameters of the classification model according to the calculated gradient and the set learning rate, and iterate in the direction of minimizing the loss; Gradient calculation and parameter update are repeated until the parameters of the classification model reach the predetermined number of iterations.

5. According to the big data analysis-based hematology clinical infection source detection system of claim 1, it is characterized in that: The infection source identification module combines the image features of the infected image area with the predicted infected image location and uses the knowledge graph in the clinical medical field to identify the infection source image of the patient, including: Obtaining image features of the infected image area and predicted infected image parts; Collect knowledge in the medical field and build a clinical medical knowledge graph; Entity extraction and entity linking techniques are used to connect the image features of the infected image area with the predicted infected image parts and related entities in the knowledge graph; A hybrid prediction model is constructed based on the image features of the infected image area, the predicted infected image location and the information in the knowledge graph, and an attention mechanism is introduced to identify the patient's infection source image.

6. A hematology clinical infection source detection system based on big data analysis according to claim 5, characterized in that: The method of using entity extraction and entity linking technology to connect the image features of the infected image area and the predicted infected image part with the relevant entities in the knowledge graph includes: Using natural language processing tools, medical entities are extracted from medical documents and image descriptions, including infection type, lesion location, and symptom manifestation; Compare and match the names of infection types, lesion sites, and symptom manifestations with the medical terminology database, and based on the matching results, convert the names of infection types, lesion sites, and symptom manifestations into standardized forms in the standard terminology database; Using graph embedding technology, the standardized medical entities are accurately linked to the corresponding entities in the constructed clinical medical knowledge graph, and the association relationship between the entities is established; The method of using graph embedding technology to accurately link the standardized medical entities with the corresponding entities in the constructed clinical medical knowledge graph and establish the association relationship between the entities includes the following steps: In the clinical medical knowledge graph, according to the standardized infection type, lesion site, and symptom manifestation, corresponding nodes and their edges are updated or added; Select a graph embedding model based on the data size and sparsity of infection type, lesion site, symptom manifestation, and actual analysis needs; Using the current graph structure, train the selected graph embedding model to encode the nodes in the clinical medical knowledge graph into vector representations; Using the node vectors obtained through training, the similarity between the vector of the target medical entity and the existing node vectors in the graph is calculated; Based on the vector similarity results, the most relevant nodes are identified and linked to ensure that the standardized medical entities accurately match the corresponding entities in the knowledge graph; Based on the linking results, new relationships between entities are updated or defined to strengthen and clarify the interactions and associations between nodes in the clinical medical knowledge graph.

7. A hematology clinical infection source detection system based on big data analysis according to claim 6, characterized in that: The hybrid prediction model is constructed based on the image features of the infected image area, the predicted infected image location and the information in the knowledge graph, and the attention mechanism is introduced to identify the infection source image of the patient, including: Integrate and normalize the image features of the infected image area, the predicted infected image location, and the information in the knowledge graph; Through the fusion layer technology in the deep learning network, the image and semantic features are combined to create a comprehensive hybrid prediction model; Design and apply a multi-head attention mechanism to evaluate and optimize the importance of each feature in identifying the source of infection; Use labeled data sets to train the model and optimize the model parameters through cross-validation techniques; The models were tested on an independent validation set and thoroughly evaluated using multiple performance metrics.

8. The hematology clinical infection source detection system based on big data analysis according to claim 1 is characterized in that: The risk assessment module assesses the infection risk of the infection source image by big data analysis in a clinical environment and classifies the infection risk into risk levels, including: Identify key indicators for measuring infection risk; Analyze infection sources using statistical methods and assess infection risks of infection sources under various conditions based on key indicators; Level thresholds are preset at each stage of infection risk, and risk levels of infection are divided based on the preset level thresholds.

9. A method for detecting clinical infection sources in hematology based on big data analysis, used to implement the system for detecting clinical infection sources in hematology based on big data analysis as claimed in any one of claims 1 to 8, characterized in that: The hematology clinical infection source detection method includes: S1. Collect clinical medical imaging images and clinical medical infection images of patients from the Department of Hematology; S2, analyzing the collected clinical medical imaging images, and marking the infected image areas based on the analysis results; S3, predicting the infected image location based on clinical medical infection images; S4, combining the image features of the infected image area with the predicted infected image location, and using the knowledge graph in the clinical medical field to identify the patient's infection source image; S5. Use big data analysis to assess the infection risk of infection source images in a clinical setting and classify the infection risk into risk levels.

Images