A brain lesion detection method and system based on multi-modal data

Abstract

The application discloses a brain lesion detection method and system based on multi-modal data, and relates to the technical field of brain disease detection. The brain lesion detection system based on multi-modal data comprises a brain lesion detection triggering module, a brain lesion detection preliminary screening module and a brain lesion detection re-screening module. The application can discover potential brain lesion clues earlier and improve the possibility of early diagnosis by screening current and historical medical records. The brain lesion feature extraction model, the disease-drug-parameter meta path and the three-attention channel feature extraction method can comprehensively and deeply analyze lesion characteristics and improve the accuracy of preliminary screening. The multi-level elimination mechanism is introduced by the brain lesion accurate re-screening model, and the parallel mixed feature analysis and space-time data are combined, so that the false judgment can be effectively reduced and the accuracy of the final diagnosis can be improved.

Description

Technical Field

[0001] The present invention relates to the technical field of brain disease detection, and in particular to a brain lesion detection method and system based on multimodal data. Background Art

[0002] Current brain diseases, such as Alzheimer's disease, Parkinson's disease, and stroke, develop insidiously and are difficult to detect in the early stages. They also have a high incidence rate and are prone to complications. However, existing detection methods have limitations: traditional MRI cannot accurately detect tiny neurodegenerative lesions, and by the time the lesions cause obvious symptoms, they have often entered an irreversible stage, reducing the effectiveness of treatment. At the same time, many patients also suffer from chronic diseases, but the potential association between these physiological data and brain lesions has not been fully explored and utilized. In addition, existing detection methods lack cross-system, multimodal linkage screening capabilities, making it difficult for high-risk groups to obtain effective intervention in a timely manner. Because additional examinations are usually cumbersome and costly, many patients choose to forgo early screening, further reducing the possibility of early intervention.

[0003] Therefore, there is an urgent need for a non-invasive detection method that reduces the need for additional medical examinations. Through correlation analysis and feature extraction of multimodal data, the early detection rate of brain lesions can be improved, and accurate detection and prediction solutions can be provided. Summary of the Invention

[0004] The present invention aims to provide a brain lesion detection method and system based on multimodal data to improve the accuracy of early brain lesion diagnosis.

[0005] A method for detecting brain lesions based on multimodal data comprises the following steps:

[0006] When receiving the current medical records of patients with potential brain changes, the current and historical medical records are screened to obtain clues to potential brain lesions; characteristic analysis is performed on the clues to potential brain lesions to obtain a preliminary screening label for brain changes; patients with potential brain changes have at least one potential brain change complication and meet the common brain change conditions;

[0007] A brain change feature extraction model is used to generate clues to potential brain lesions. These clues are then processed to generate preliminary screening labels for brain changes. The brain change feature extraction model is used to predict the spread of brain changes by combining the symptom-medication-parameter metapath and the three-attention channel feature extraction method to generate preliminary screening labels for brain changes.

[0008] Potential patients with brain degeneration whose initial screening labels meet the brain degeneration characteristic indicators are marked as patients with brain degeneration to be screened;

[0009] A secondary brain lesion test is performed on patients who are being screened for brain changes to obtain the results of the brain change re-screening. A precise re-screening model for brain changes is used to perform secondary brain lesion detection, introduce a multi-level elimination mechanism, and use parallel mixed features for judgment and screening, combining spatiotemporal data to improve the accuracy of the brain change re-screening results.

[0010] As a preferred technical solution of the present invention, the specific steps of screening current medical records and historical medical records include:

[0011] The brain change feature extraction model includes a data linkage layer, a clue extraction layer, and a label output layer;

[0012] The data linkage layer is used to extract the big data of symptoms of potential brain degeneration patients corresponding to potential brain degeneration complications to obtain brain degeneration induction data; the brain degeneration induction data and current medical records and historical medical records are screened to obtain brain degeneration induction multimodal data D n , n=1, 2, …, N; N is the total number of dimensions of brain change inducing factors;

[0013] The clue extraction layer is used to induce multimodal data D based on brain changes n Conduct feature analysis to obtain clues to potential brain lesions;

[0014] The label output layer is used to output features based on potential brain lesion clues to obtain preliminary screening labels for brain changes.

[0015] As a preferred technical solution of the present invention, the specific steps of performing feature analysis in the clue extraction layer include:

[0016] For brain change induced multimodal data D n Generate the disease-medication-parameter metapath to obtain the brain change clue extraction metapath L n ;

[0017] In the clue extraction layer, three attention channels are constructed. The three attention channels are pathological feature attention, treatment response attention, and social behavior attention. The internal structure is a parallel decoupling structure. The brain change clue extraction meta-path L n Input to the three attention channels for parallel analysis to obtain the brain change induced feature Z n ;

[0018] At the same time, all brain change-induced multimodal data D n The image data in the image data are used to obtain the brain variable image data;

[0019] Mapping brain variable image data into a density matrix in Hilbert space to obtain a brain variable analysis matrix;

[0020] Brain change analysis matrix and brain change induced feature Z nCross-modal superposition quantum gates are used to perform comprehensive clue extraction to obtain clues of potential brain lesions.

[0021] As a preferred technical solution of the present invention, the specific steps of performing feature output in the label output layer include:

[0022] Using potential brain lesion clues to predict brain lesion spread results, and obtaining brain lesion spread prediction results;

[0023] Perform quantitative label conversion on the prediction results of brain degeneration spread to obtain the initial screening label of brain degeneration;

[0024] Among them, using the formula The prediction results of brain lesion diffusion are obtained by calculation; X represents the potential clues of brain lesions, t represents the time axis of brain lesion development, and W represents the predicted brain lesion diffusion coefficient. Represents the result value of calculating potential brain lesion clues using the Laplace operator, Represents the nonlinear brain change reflection term.

[0025] As a preferred technical solution of the present invention, the specific steps of performing secondary brain lesion detection on patients with ethmoidal encephalopathy include:

[0026] Obtain brain degeneration-specific examination data for patients to be screened for brain degeneration;

[0027] The brain change special examination data is input into the brain change precision screening model for diagnosis. The brain change precision screening model includes a multi-level elimination mechanism layer and a brain change detection layer;

[0028] Perform parallel mixed feature judgment on several brain pathological features in the brain degeneration special examination data. If no less than m brain pathological features meet the brain lesion screening criteria, key images are collected for the patient to be screened for brain degeneration to obtain key images of brain degeneration;

[0029] Combining the brain degeneration special examination data and the key brain degeneration images, the quantum annealing machine is used to solve the optimal classification hyperplane to obtain the multidimensional brain degeneration risk vector R, R = (R1, R2, ..., R I ), I is the vector dimension in the multidimensional brain degeneration risk vector R;

[0030] The multidimensional brain degeneration risk vector R is input into the brain degeneration detection layer to obtain the brain degeneration re-screening result.

[0031] As a preferred technical solution of the present invention, the specific steps of obtaining the brain degeneration re-screening results in the brain degeneration detection layer include:

[0032] Extract the brain change spread prediction results of the patients to be screened for brain changes, and obtain the initial screening brain change spread prediction results;

[0033] Align the initial screening brain metastasis spread prediction results and the multidimensional brain metastasis risk vector R in time and space to obtain brain metastasis risk diagnosis data;

[0034] The brain change knowledge graph is used to encode the brain change attention bias matrix, and dynamic convolutional attention is introduced into the brain change attention bias matrix to perform brain change detection on the brain change risk diagnosis data to obtain the final brain change re-screening results.

[0035] A brain lesion detection system based on multimodal data, comprising:

[0036] The brain change detection trigger module includes a data screening unit; the data screening unit is used to receive the current medical records of patients with potential brain changes, screen the current medical records and historical medical records to obtain clues to potential brain lesions; perform feature analysis on the clues to potential brain lesions to obtain a preliminary screening label for brain changes; the patient with potential brain changes has at least one potential brain change complication and meets the common brain change conditions;

[0037] The brain change detection and screening module includes an anomaly detection unit. The anomaly detection unit is used to generate potential brain lesion clues using a brain change feature extraction model and process the potential brain lesion clues to generate brain change screening labels. The brain change feature extraction model is used to combine the symptom-medication-parameter meta-path and the three-attention channel feature extraction method to predict brain change spread and generate brain change screening labels. Potential brain change patients whose brain change screening labels meet the brain change feature indicators are marked as potential brain change patients for screening.

[0038] The brain degeneration detection and rescreening module includes a precise detection unit; the precise detection unit is used to perform secondary brain lesion detection on patients to be screened for brain degeneration and obtain brain degeneration rescreening results; the secondary brain lesion detection is performed using a precise rescreening model for brain degeneration, a multi-level elimination mechanism is introduced, and judgment and screening are performed through parallel mixed features, and spatiotemporal data is combined to improve the accuracy of brain degeneration rescreening results.

[0039] The present invention has the following advantages:

[0040] 1. The present invention combines current and historical medical records for screening, which can detect potential brain lesion clues earlier and increase the possibility of early diagnosis. The brain change feature extraction model is combined with the symptom-medication-parameter meta-path and the three-attention channel feature extraction method to more comprehensively and deeply analyze the lesion characteristics and improve the accuracy of the initial screening. By considering the potential complications and common brain change conditions that the patient may suffer from, the lesion risk is analyzed from multiple angles to improve the reliability of the screening. The multi-level elimination mechanism is introduced through the brain change precision re-screening model, and combined with parallel hybrid feature analysis and spatiotemporal data, it can effectively reduce misjudgments and improve the accuracy of the final diagnosis.

[0041] 2. The present invention establishes the association between the patient's pathological characteristics, treatment history and physiological parameters through the symptom-medication-parameter meta-path, so that screening is not limited to a single factor, but is analyzed based on complete medical data; it uses multi-source information such as imaging data, pathological data, and social behavior data to build a more complete patient health portrait and improve the accuracy of brain lesion clue extraction; pathological characteristic attention is used to focus on the biological characteristics of the disease, treatment response attention is used to analyze the patient's medication history and treatment effect, and social behavior attention is used to comprehensively consider the patient's impact on cognitive and neural functions. The introduction of a parallel decoupling structure can avoid information redundancy, ensure that each attention channel focuses on features of different dimensions, and improve the interpretability and adaptability of the model. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] Figure 1 This is a structural diagram of a brain lesion detection system based on multimodal data adopted in an embodiment of the present invention. DETAILED DESCRIPTION

[0043] In order to enable those skilled in the art to better understand the technical solutions in the present invention, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings in the embodiments of the present invention.

[0044] Example 1, a method for detecting brain lesions based on multimodal data, comprising the following steps:

[0045] When receiving the current medical records of patients with potential brain changes, the current and historical medical records are screened to obtain clues to potential brain lesions; characteristic analysis is performed on the clues to potential brain lesions to obtain a preliminary screening label for brain changes; patients with potential brain changes have at least one potential brain change complication and meet the common brain change conditions;

[0046] For example, the groups that meet the conditions of potential brain change patients mainly include the following groups of people: patients with a history of brain disease, long-term chronic diseases, long-term use of neurological drugs, or patients with high-risk lifestyle habits or environmental factors; potential brain change complications refer to complications that may cause or aggravate brain lesions, such as: cerebrovascular complications, neurological diseases, metabolic and systemic diseases, such as hypertension, diabetes, chronic renal insufficiency, hypothyroidism, etc.; common brain change conditions include imaging abnormalities, cognitive decline, motor dysfunction, neuropsychiatric symptoms, older age, etc., which will all become potential brain change patients;

[0047] By screening current and historical medical records, potential clues to brain lesions can be discovered earlier, preventing patients from missing the optimal opportunity for intervention due to the lack of obvious early symptoms. Traditional brain lesion screening relies on doctors' experience and manual analysis, or often shows more serious brain changes. This method, however, utilizes automated screening and feature analysis to significantly reduce the time cost of manual screening and improve the efficiency of medical resource utilization. By comprehensively analyzing medical history, medication use, and related parameters, further testing is only performed on patients who meet the criteria for brain lesion complications and common brain changes, avoiding unnecessary screening and improving screening accuracy.

[0048] A brain change feature extraction model is used to generate clues to potential brain lesions. These clues are then processed to generate preliminary screening labels for brain changes. The brain change feature extraction model is used to predict the spread of brain changes by combining the symptom-medication-parameter metapath and the three-attention channel feature extraction method to generate preliminary screening labels for brain changes.

[0049] Specific steps for screening current and historical medical records include:

[0050] The brain change feature extraction model includes a data linkage layer, a clue extraction layer, and a label output layer;

[0051] The data linkage layer is used to extract the big data of symptoms of potential brain degeneration patients corresponding to potential brain degeneration complications to obtain brain degeneration induction data; the brain degeneration induction data and current medical records and historical medical records are screened to obtain brain degeneration induction multimodal data D n , n=1, 2, …, N; N is the total number of dimensions of brain change inducing factors;

[0052] The clue extraction layer is used to induce multimodal data D based on brain changes n Conduct feature analysis to obtain clues to potential brain lesions;

[0053] The label output layer is used to output features based on potential brain lesion clues to obtain preliminary screening labels for brain changes;

[0054] The data linkage layer integrates multimodal data such as the medical records, medical history, and medication status of patients with potential brain changes, helping to uncover hidden clues to brain changes and improve the early detection rate of brain lesions. Through big data analysis, it can extract information on symptoms related to potential brain change complications, avoid overlooking potential risks due to individual factors, and improve screening accuracy.

[0055] By inducing multimodal data on brain changes, we fully consider multidimensional factors, avoid misjudgments that may result from relying solely on a single indicator, and improve the comprehensiveness of disease identification. Combined with feature extraction using three attention channels, we can highlight key features, increase the model's sensitivity to brain lesion risks, and reduce the risk of misdiagnosis and missed diagnosis.

[0056] The clue extraction layer uses feature analysis technology to mine valuable potential lesion clues from multimodal data. Automated analysis can quickly locate high-risk patients, improve hospital screening efficiency, and reduce the workload of doctors. The label output layer intelligently outputs the extracted brain lesion features to generate preliminary screening labels for brain changes, providing a clear basis for further precise detection. By labeling screening results, risk stratification can be achieved, providing doctors with more intuitive diagnostic references, optimizing the screening process, and improving the rational allocation of medical resources.

[0057] For example, diabetes is a common potential complication of brain degeneration. Its long-term development may lead to cerebrovascular disease, cognitive decline, and neurodegenerative diseases. If diabetes is analyzed as a potential complication of brain degeneration, the pathological big data may include the following aspects: the association between blood sugar fluctuations and cognitive decline, the association between diabetes and cerebral small vessel disease, the increased risk of hippocampal atrophy in diabetic patients, the neuroinflammatory response caused by diabetes, and the correlation between the frequency of hypoglycemia and Parkinson's syndrome-like symptoms. These clues can help screen high-risk diabetic patients and determine whether they are likely to develop brain degeneration.

[0058] The specific steps of feature analysis in the clue extraction layer include:

[0059] For brain change induced multimodal data D n Generate the disease-medication-parameter metapath to obtain the brain change clue extraction metapath L n ;

[0060] In the clue extraction layer, three attention channels are constructed. The three attention channels are pathological feature attention, treatment response attention, and social behavior attention. The internal structure is a parallel decoupling structure. The brain change clue extraction meta-path L n Input to the three attention channels for parallel analysis to obtain the brain change induced feature Z n ;

[0061] At the same time, all brain change-induced multimodal data D n The image data in the image data are used to obtain the brain variable image data;

[0062] Mapping brain variable image data into a density matrix in Hilbert space to obtain a brain variable analysis matrix;

[0063] Brain change analysis matrix and brain change induced feature Z n Using cross-modal superposition quantum gates to extract comprehensive clues and obtain clues of potential brain lesions;

[0064] Through the symptom-medication-parameter meta-path, the association between the patient's pathological characteristics, treatment history and physiological parameters is established, so that screening is not limited to a single factor, but is analyzed based on complete medical data; using multi-source information such as imaging data, pathological data, and social behavior data, a more complete patient health portrait is constructed to improve the accuracy of brain lesion clue extraction; pathological characteristic attention is used to focus on the biological characteristics of the disease, such as imaging manifestations, clinical symptoms and disease progression trends; treatment response attention is used to analyze the patient's medication history and treatment effects to determine whether the treatment affects brain health; social behavior attention is used to comprehensively consider the impact of external factors such as the patient's lifestyle, sleep patterns, social interactions, etc. on cognitive and neural functions; the introduction of a parallel decoupling structure can avoid information redundancy, ensure that each attention channel focuses on features of different dimensions, and improve the interpretability and adaptability of the model;

[0065] Extracting all imaging data from multimodal data induced by brain changes and converting it into a density matrix in Hilbert space allows for better mathematical representation of the data's features. Quantized mapping of imaging data helps improve image recognition accuracy and enhances the screening system's sensitivity to subtle lesions, such as white matter lesions and brain atrophy. By superimposing quantum gates across modalities, deep fusion of imaging data with disease-drug-parameter data is achieved, reducing deviations between cross-modal data and improving data complementarity. Quantum computing methods make data mapping more efficient, enabling faster calculation of potential brain lesion clues and improving computational efficiency in screening.

[0066] The specific steps for feature output in the label output layer include:

[0067] Using potential brain lesion clues to predict brain lesion spread results, and obtaining brain lesion spread prediction results;

[0068] Perform quantitative label conversion on the prediction results of brain degeneration spread to obtain the initial screening label of brain degeneration;

[0069] Among them, using the formula The predicted results of brain lesion diffusion are calculated; X represents the potential clues of brain lesions, t represents the timeline of brain lesion development, and W represents the predicted brain lesion diffusion coefficient, which is set by professional technicians according to actual conditions; Represents the result value of calculating potential brain lesion clues using the Laplace operator, represents the nonlinear brain change reflection term;

[0070] By using a formula to calculate brain lesion spread predictions and using the Laplace operator to calculate clues to potential brain lesions, the system can quantify the spread of brain lesions and improve the ability to predict disease progression. The predictions are combined with a timeline, enabling the screening system to dynamically monitor disease progression rather than statically analyze it, improving the timeliness of detection. The predictions are converted into initial brain lesion screening labels, enabling stratified screening of patients, allowing doctors to quickly identify high-risk patients and optimize medical decision-making. Nonlinear brain lesion reflection items are used to capture the complex dynamic changes in brain lesions. Their training primarily involves data preprocessing, feature extraction, mathematical modeling, parameter optimization, and verification.

[0071] Potential patients with brain degeneration whose initial screening labels meet the brain degeneration characteristic indicators are marked as patients with brain degeneration to be screened;

[0072] Brain degeneration characteristic indicators are set by professional technicians based on actual conditions; professional technicians can set screening thresholds based on specific needs. Brain degeneration characteristic indicators should be combined with multiple factors such as imaging, cognitive function, blood biomarkers, medical history, physical signs, lifestyle, etc. to ensure the accuracy and comprehensiveness of screening and improve the ability to identify patients with potential brain diseases at an early stage.

[0073] Perform secondary brain lesion detection on patients who are being screened for brain changes to obtain the results of the brain change rescreening. Use the brain change precision rescreening model to perform secondary brain lesion detection, introduce a multi-level elimination mechanism, and use parallel hybrid features for judgment and screening, combining spatiotemporal data to improve the accuracy of the brain change rescreening results.

[0074] The specific steps for secondary brain lesion testing in patients with ethmoidal encephalopathy include:

[0075] Obtaining brain change-specific examination data for patients to be screened for brain changes; the screening center or hospital information system will screen eligible patients for screening based on the initial brain change screening results, provide reasons for requiring further special examinations, and conduct further screening based on the patient's wishes to obtain brain change-specific examination data;

[0076] The brain change special examination data is input into the brain change precision screening model for diagnosis. The brain change precision screening model includes a multi-level elimination mechanism layer and a brain change detection layer;

[0077] Perform parallel mixed feature judgment on several brain pathological features in the brain degeneration special examination data. If no less than m brain pathological features meet the brain lesion screening criteria, key images are collected for the patient to be screened for brain degeneration to obtain key images of brain degeneration;

[0078] Among them, for patients with less than m brain pathological features that meet the brain lesion screening criteria, that is, individuals who do not meet the clear brain lesion screening criteria but may have a certain risk, hospitals and medical institutions can adopt tiered management or personalized intervention measures;

[0079] Combining the brain degeneration special examination data and the key brain degeneration images, the quantum annealing machine is used to solve the optimal classification hyperplane to obtain the multidimensional brain degeneration risk vector R, R = (R1, R2, ..., R I ), I is the vector dimension in the multidimensional brain degeneration risk vector R;

[0080] Input the multidimensional brain change risk vector R into the brain change detection layer to obtain the brain change re-screening result;

[0081] By acquiring specialized examination data for brain changes and conducting secondary analysis, the accuracy of detection can be improved and misjudgments can be reduced. A multi-level elimination mechanism is adopted to ensure that only truly high-risk patients enter the next stage of screening, avoiding waste of medical resources and optimizing the screening process. In the specialized examination data for brain changes, only if no fewer than m brain lesion characteristics are met will further key image acquisition be carried out to reduce the possibility of misjudgment of a single feature. The weights of different features are dynamically adjusted, and based on comprehensive analysis of pathological, imaging, and cognitive data, the model's adaptability to different types of brain lesions is improved. Key image acquisition, such as MRI and PET-CT, is only carried out after preliminary feature screening, reducing unnecessary imaging examinations, improving the efficiency of medical resource utilization, and saving patients' diagnosis and treatment costs.

[0082] By solving the optimal classification hyperplane using a quantum annealing machine, the optimal risk assessment boundary can be quickly calculated under multi-dimensional data, improving the computational efficiency of screening. The generation of a multi-dimensional brain change risk vector enables a more detailed assessment of different types of brain change risks.

[0083] The specific steps for obtaining the brain change re-screening results in the brain change detection layer include:

[0084] Extract the brain change spread prediction results of patients to be screened for brain changes, and obtain the initial screening brain change spread prediction results;

[0085] Align the initial screening brain metastasis spread prediction results and the multidimensional brain metastasis risk vector R in time and space to obtain brain metastasis risk diagnosis data;

[0086] The brain change knowledge graph is used to encode and generate a brain change attention bias matrix. Dynamic convolutional attention is introduced into the brain change attention bias matrix to perform brain change detection on brain change risk diagnosis data and obtain the final brain change re-screening results.

[0087] The brain change detection layer uses advanced technologies such as brain change spread prediction, multi-dimensional risk vector alignment, knowledge graph encoding, and dynamic convolutional attention mechanism to achieve more accurate, efficient, and intelligent re-screening of brain lesions. The initial screening results provide early lesion trend prediction based on the lesion spread trend, and the multi-dimensional brain change risk vector provides a more detailed individual risk score, which is combined with multimodal data such as imaging, biomarkers, and cognitive tests to enhance prediction accuracy. Based on a large-scale medical database containing clinical cases, medical research, molecular biology information, etc., a brain lesion causal relationship map is established. Through the knowledge graph reasoning association, high-risk factor combinations, such as certain specific biological factors, are automatically identified. Biomarkers plus imaging abnormalities may be highly correlated with specific types of neurodegenerative diseases; combining them with individual patient data for matching can improve the accuracy of disease classification and reduce misdiagnosis and missed diagnoses; traditional neural networks have fixed weights and may not be able to adapt to the brain lesion characteristics of different individuals, while dynamic convolutional attention can adjust feature weights for different patients to ensure that the most relevant lesion characteristics receive higher attention; for example, for Alzheimer's disease, the model may pay more attention to hippocampal atrophy and β-amyloid protein levels, while for stroke, it may pay more attention to white matter lesions and vascular factors; the attention mechanism filters out irrelevant information and reduces misjudgments caused by factors such as data errors and imaging artifacts;

[0088] During the brain lesion detection process, the present invention performs intelligent screening through early calculation, which can effectively reduce subsequent unnecessary high-cost examinations and optimize the allocation of medical resources. At the same time, misdiagnosis will cause patients to receive unnecessary treatment, increasing medical expenses; missed diagnosis will delay the best time for intervention, causing the patient's condition to worsen, and the cost of later treatment will far exceed the cost of early calculation. Early calculation helps to quickly lock in high-risk patients, which can optimize hospital resources, improve screening efficiency, reduce misdiagnosis rate, reduce patients' high later treatment costs, and improve the overall operational efficiency of the hospital and the level of patient health management.

[0089] Example 2, a brain lesion detection system based on multimodal data, see Figure 1 Shown, including:

[0090] The brain change detection trigger module includes a data screening unit; the data screening unit is used to receive the current medical records of patients with potential brain changes, screen the current medical records and historical medical records to obtain clues to potential brain lesions; perform feature analysis on the clues to potential brain lesions to obtain a preliminary screening label for brain changes; the patient with potential brain changes has at least one potential brain change complication and meets the common brain change conditions;

[0091] The brain change detection and screening module includes an anomaly detection unit. The anomaly detection unit is used to generate potential brain lesion clues using a brain change feature extraction model and process the potential brain lesion clues to generate brain change screening labels. The brain change feature extraction model is used to combine the symptom-medication-parameter meta-path and the three-attention channel feature extraction method to predict brain change spread and generate brain change screening labels. Potential brain change patients whose brain change screening labels meet the brain change feature indicators are marked as potential brain change patients for screening.

[0092] The brain degeneration detection and rescreening module includes a precise detection unit; the precise detection unit is used to perform secondary brain lesion detection on patients to be screened for brain degeneration and obtain brain degeneration rescreening results; the secondary brain lesion detection is performed using a precise rescreening model for brain degeneration, a multi-level elimination mechanism is introduced, and judgment and screening are performed through parallel mixed features, and spatiotemporal data is combined to improve the accuracy of brain degeneration rescreening results.

[0093] It should be understood that those skilled in the art may make improvements or modifications based on the above description, and all such improvements and modifications shall fall within the scope of protection of the appended claims. Any portion of this specification not described in detail is prior art known to those skilled in the art.

Claims

1. A brain lesion detection method based on multimodal data, characterized in that: The following steps are involved: When receiving the current medical records of patients with potential brain changes, the current medical records and historical medical records are screened to ensure that the patients with potential brain changes have at least one potential brain change complication and meet the common brain change conditions; The brain change feature extraction model is used to generate potential brain lesion clues, and the potential brain lesion clues are processed to generate brain change initial screening labels, including: The brain change feature extraction model includes a data linkage layer, a clue extraction layer, and a label output layer; The data linkage layer is used to extract the big data of symptoms of potential brain degeneration patients corresponding to potential brain degeneration complications to obtain brain degeneration induction data; the brain degeneration induction data and current medical records and historical medical records are screened to obtain brain degeneration induction multimodal data D n , n = 1, 2, ..., N; N is the total number of dimensions of brain change inducing factors; For brain change induced multimodal data D n Generate the disease-medication-parameter metapath to obtain the brain change clue extraction metapath L n ; In the clue extraction layer, three attention channels are constructed. The three attention channels are pathological feature attention, treatment response attention, and social behavior attention. The internal structure is a parallel decoupling structure. The brain change clue extraction meta-path L n Input to the three attention channels for parallel analysis to obtain the brain change induced feature Z n ; At the same time, all brain change-induced multimodal data D n The image data in the image data are used to obtain the brain variable image data; Mapping brain variable image data into a density matrix in Hilbert space to obtain a brain variable analysis matrix; Brain change analysis matrix and brain change induced feature Z n Using cross-modal superposition quantum gates to extract comprehensive clues and obtain clues of potential brain lesions; In the label output layer, potential brain lesion clues are used to predict the spread of brain changes and obtain the prediction results of brain change spread. Specifically, the formula The brain change diffusion prediction results are calculated; X represents the potential brain lesion clues, t represents the timeline of brain change development, W represents the predicted brain change diffusion coefficient, ΔX represents the result value calculated using the Laplace operator for the potential brain lesion clues, and f(X) represents the nonlinear brain change reflection term; Perform quantitative label conversion on the prediction results of brain degeneration spread to obtain the initial screening label of brain degeneration; Potential patients with brain degeneration whose initial screening labels meet the brain degeneration characteristic indicators are marked as patients with brain degeneration to be screened; A precise re-screening model for brain changes is used to conduct secondary brain lesion detection on patients to be screened for brain changes. A multi-level elimination mechanism is introduced and parallel mixed feature judgment of brain pathological characteristics is performed to achieve judgment screening. At the same time, the re-screening results of brain changes are obtained by combining spatiotemporal data.

2. The method for detecting brain lesions based on multimodal data according to claim 1, characterized in that: The specific steps for secondary brain lesion testing in patients with ethmoidal encephalopathy include: Obtain brain degeneration-specific examination data for patients to be screened for brain degeneration; The brain change special examination data is input into the brain change precision screening model for diagnosis. The brain change precision screening model includes a multi-level elimination mechanism layer and a brain change detection layer; Perform parallel mixed feature judgment on several brain pathological features in the brain degeneration special examination data. If no less than m brain pathological features meet the brain lesion screening criteria, key images are collected for the patient to be screened for brain degeneration to obtain key images of brain degeneration; Combining the brain degeneration special examination data and the key brain degeneration images, the quantum annealing machine is used to solve the optimal classification hyperplane to obtain the multidimensional brain degeneration risk vector R, R = (R1, R2, ..., R I ), I is the vector dimension in the multidimensional brain degeneration risk vector R; The multidimensional brain degeneration risk vector R is input into the brain degeneration detection layer to obtain the brain degeneration re-screening result.

3. The method for detecting brain lesions based on multimodal data according to claim 2, wherein: The specific steps for obtaining the brain change re-screening results in the brain change detection layer include: Extract the brain change spread prediction results of the patients to be screened for brain changes, and obtain the initial screening brain change spread prediction results; Align the initial screening brain metastasis spread prediction results and the multidimensional brain metastasis risk vector R in time and space to obtain brain metastasis risk diagnosis data; The brain change knowledge graph is used to encode the brain change attention bias matrix, and dynamic convolutional attention is introduced into the brain change attention bias matrix to perform brain change detection on the brain change risk diagnosis data to obtain the final brain change re-screening results.

4. A brain lesion detection system based on multimodal data, characterized in that: The system applies the brain lesion detection method based on multimodal data according to any one of claims 1 to 3, comprising: The brain change detection trigger module includes a data screening unit; the data screening unit is used to receive the current medical records of patients with potential brain changes, screen the current medical records and historical medical records to obtain clues to potential brain lesions; perform feature analysis on the clues to potential brain lesions to obtain a preliminary screening label for brain changes; the patient with potential brain changes has at least one potential brain change complication and meets the common brain change conditions; The brain change detection and screening module includes an anomaly detection unit. The anomaly detection unit is used to generate potential brain lesion clues using a brain change feature extraction model and process the potential brain lesion clues to generate brain change screening labels. The brain change feature extraction model is used to combine the symptom-medication-parameter meta-path and the three-attention channel feature extraction method to predict brain change spread and generate brain change screening labels. Potential brain change patients whose brain change screening labels meet the brain change feature indicators are marked as potential brain change patients for screening. The brain degeneration detection and rescreening module includes a precise detection unit; the precise detection unit is used to perform secondary brain lesion detection on patients to be screened for brain degeneration and obtain brain degeneration rescreening results; the secondary brain lesion detection is performed using a precise rescreening model for brain degeneration, a multi-level elimination mechanism is introduced, and judgment and screening are performed through parallel mixed features, and spatiotemporal data is combined to improve the accuracy of brain degeneration rescreening results.

Images