Intelligent analysis system of mammary gland detector and mammary gland detector

Abstract

The invention relates to the technical field of medical instruments, and particularly discloses an intelligent analysis system of a mammary gland detector and the mammary gland detector, and the system comprises a mammary gland file generation module which records previous mammary gland detection results of a user and generates a mammary gland health file; the risk factor identification module is used for extracting all breast lesion risk factors from the archive; the lesion type prediction module analyzes a detection result according to the mammary gland lesion prediction model, predicts a potential abnormal lesion type and identifies corresponding potential decision data; the lesion possibility prediction module is used for predicting the lesion possibility of each potential abnormal lesion type based on the risk factors, the lesion types and the decision data; the early warning mechanism judgment module is used for immediately triggering an early warning mechanism once the lesion possibility exceeds an early warning threshold value; the system can effectively perform comprehensive management, prediction and early warning on the breast health of the user, and is helpful for improving the prevention and treatment effect of breast diseases.

Description

Technical Field

[0001] The present invention relates to the technical field of medical devices, and particularly relates to an intelligent analysis system for a breast detector and a breast detector. Background Art

[0002] Traditional breast examinations mainly rely on doctors to subjectively interpret imaging results (such as mammograms, ultrasound images, etc.) based on experience. With the development of information technology, some breast detectors have begun to have certain digital functions, capable of simply storing detection data and performing basic feature extraction, such as measuring basic parameters such as the size and shape of breast nodules. Some early intelligent auxiliary diagnosis systems have started to use machine learning algorithms to analyze breast detection data and attempt to predict the possibility of breast lesions. These systems can improve the accuracy of diagnosis to a certain extent.

[0003] Existing breast detectors lack systematic intelligent analysis. The previous breast detection results of patients are often stored separately, without forming a unified and complete breast health record. This makes it difficult for doctors to comprehensively understand the development process of patients' breast health and affects the accurate judgment of the condition. Although some systems can identify some common risk factors for breast lesions, they lack systematic extraction and analysis of all relevant risk factors. Many potential risk factors may be ignored, resulting in inaccurate assessment of the lesion risk for patients. When existing intelligent analysis systems predict the types of potential abnormal breast lesions, they often analyze based on only single detection data or a few features, without making full use of the patient's historical detection data for comprehensive judgment. This leads to inaccurate prediction results and the inability to comprehensively identify all potential lesion types. When predicting the possibility of lesions, the prior art does not fully consider the complex relationships among breast lesion risk factors, potential abnormal lesion types, and corresponding potential decision-making data. It simply makes judgments based on partial data, resulting in low accuracy of the predicted lesion possibility. Moreover, it lacks an effective early warning trigger mechanism and cannot timely remind patients and doctors to take corresponding measures. With the continuous improvement of people's attention to health management, the demand for long-term tracking and management of breast health is increasing. The traditional detection and analysis methods cannot meet the needs of users to establish a comprehensive breast health record and conduct dynamic risk assessment.

[0004] Therefore, the present invention proposes an intelligent analysis system for a breast detector and a breast detector. Summary of the Invention

[0005] The present invention provides an intelligent analysis system for a breast detector and a breast detector, including: A breast file generation module can record the previous detection results of a user, form a complete breast health file, and facilitate long-term tracking and comparative analysis of the user's breast health status. This enables doctors and patients to comprehensively understand the development process of breast health and provides complete data support for subsequent analysis and diagnosis. The risk factor identification module helps to comprehensively understand the factors that may lead to breast lesions and provides an important basis for subsequent prediction and early warning. It overcomes the problem of incomplete risk factor identification in the prior art and ensures more accurate assessment of the patient's lesion risk. The lesion type prediction module analyzes and predicts potential abnormal lesion types and decision-making data through model analysis, providing a direction for diagnosis and treatment in advance. By making full use of historical detection data, the accuracy and comprehensiveness of lesion type prediction are improved. The lesion probability prediction module comprehensively considers lesion risk factors, etc., and accurately predicts the probability of each potential lesion type, making the assessment more precise. By comprehensively considering the complex relationships between various factors, it accurately predicts the lesion probability of each potential abnormal breast lesion type of the user, solving the problem of inaccurate lesion probability prediction in the prior art. The early warning mechanism judgment module triggers an early warning in a timely manner when the lesion probability exceeds the threshold, which can remind the user and medical staff to take corresponding measures to achieve early intervention and treatment. It reminds patients and doctors to pay attention to potential health risks. This effective early warning trigger mechanism makes up for the defect of the imperfect early warning mechanism in the prior art. This system can effectively manage, predict and give early warning to the user's breast health, and helps to improve the prevention and treatment effect of breast diseases.

[0006] The present invention provides an intelligent analysis system for a breast detector, including:

[0007] A breast file generation module, configured to record the previous breast detection results of a user and generate a breast health file of the user;

[0008] A risk factor identification module, configured to extract all breast lesion risk factors from the breast health file of the user;

[0009] A lesion type prediction module, configured to analyze the previous breast detection results of a user based on a breast lesion prediction model, predict all potential abnormal breast lesion types of the user, and identify potential decision-making data for each potential abnormal breast lesion type;

[0010] A lesion probability prediction module, configured to predict the lesion probability of each potential abnormal breast lesion type of the user based on all breast lesion risk factors, the potential abnormal breast lesion types of the user, and the corresponding potential decision-making data;

[0011] An early warning mechanism judgment module, configured to trigger an early warning mechanism when the lesion probability of the user exceeds an early warning threshold.

[0012] Preferably, the breast file generation module includes:

[0013] A detection data acquisition sub-module, configured to acquire in real time the detection data generated by the breast detector for detecting the user's breast tissue as the user's single breast detection result;

[0014] A data processing and file generation sub-module, configured to normalize each breast detection result of the user, store it in a local storage device or a cloud service area according to a preset storage format and database architecture, and summarize the user's latest physiological signs information, personal medical history, and family medical history to construct the user's breast health file.

[0015] Preferably, the risk factor identification module includes:

[0016] A raw atlas acquisition sub-module, configured to acquire a preset breast lesion risk factor atlas, where the preset breast health atlas is used to define the ontology structure of breast lesion risk factors;

[0017] A file data mapping and extraction sub-module, configured to perform an association mapping between the user's breast health file and the preset breast health atlas, and use entity recognition and relationship extraction methods to identify information corresponding to the raw breast lesion risk factors in the preset breast health atlas from the user's breast health file, and establish the relationships between all the information corresponding to the raw breast lesion risk factors identified in the user's breast health file as the user's personalized breast lesion risk factor atlas;

[0018] A lesion risk factor extraction sub-module, configured to obtain all the breast lesion risk factors of the user based on the user's personalized breast lesion risk factor atlas.

[0019] Preferably, the lesion risk factor extraction sub-module includes:

[0020] An academic data mapping and extraction unit, configured to perform an association mapping between the preset breast diagnosis expert academic database and the preset breast health atlas, and use entity recognition and relationship extraction methods to identify information corresponding to the raw breast lesion risk factors in the preset breast health atlas from the preset breast diagnosis expert academic database, and establish the relationships between all the information corresponding to the raw breast lesion risk factors identified in the preset breast diagnosis expert academic database as the raw breast lesion risk factor atlas of the preset breast diagnosis expert academic database;

[0021] A lesion risk factor extraction unit, which is used to analyze the weight of each cross-class risk factor group in the preset academic library of breast diagnosis experts based on the original breast lesion risk factor atlas of the preset academic library of breast diagnosis experts, and combine the breast lesion diagnosis information corresponding to all cross-class risk factor information groups of each group of cross-class risk factors in the preset academic library of breast diagnosis experts to analyze all breast lesion risk factors of the user.

[0022] Preferably, the lesion risk factor extraction unit includes:

[0023] A cross-class factor information recognition subunit, which is used to recognize all cross-class risk factor information groups of each cross-class risk factor group in the preset academic library of breast diagnosis experts based on the original breast lesion risk factor atlas of the preset academic library of breast diagnosis experts;

[0024] A cross-class factor weight calculation subunit, which is used to calculate the weight of each cross-class risk factor group in the preset academic library of breast diagnosis experts based on all cross-class risk factor information groups of each cross-class risk factor group in the preset academic library of breast diagnosis experts;

[0025] An inferable factor screening subunit, which is used to regard each cross-class risk factor group with a weight exceeding the weight threshold as an inferable cross-class risk factor group;

[0026] A lesion risk factor recognition subunit, which is used to analyze all breast lesion risk factors of the user based on all cross-class risk factor information groups of all inferable cross-class risk factor groups, the corresponding breast lesion diagnosis information in the preset academic library of breast diagnosis experts, and the user's personalized breast lesion risk factor atlas.

[0027] Preferably, the lesion risk factor recognition subunit includes:

[0028] An inference original information extraction end, which is used to regard all cross-class risk factor information groups of each inferable cross-class risk factor group and the corresponding breast lesion diagnosis information in the preset academic library of breast diagnosis experts as the original information of the breast lesion risk inference factors of the corresponding inferable cross-class risk factor group;

[0029] A factor regression analysis end, which is used to perform a regression analysis on the original information of the breast lesion risk inference factors of each inferable cross-class risk factor group to obtain the inference relationship between each inferable cross-class risk factor group and the corresponding breast lesion risk inference factors;

[0030] The reasoning factor mining and verification end is used to verify and mine all cross - category risk factor information groups of each inferable cross - category risk factor group included in the user's personalized breast lesion risk factor map by using the reasoning relationship between all inferable cross - category risk factor groups and the corresponding breast lesion risk reasoning factors, so as to obtain all breast lesion risk reasoning factors in the user's personalized breast lesion risk factor map;

[0031] The lesion risk factor summary end is used to regard all breast lesion risk reasoning factors and original breast lesion risk factors in the user's personalized breast lesion risk factor map as all breast lesion risk factors of the user.

[0032] Preferably, the lesion type prediction module includes:

[0033] The first model building sub - module is used to build a breast lesion prediction model;

[0034] The lesion type prediction sub - module is used to input the user's previous breast detection results into the breast lesion prediction model to predict all potential abnormal breast lesion types of the user and the potential decision data of each potential abnormal breast lesion type.

[0035] Preferably, the lesion possibility prediction module includes:

[0036] The second model building sub - module is used to build a lesion possibility prediction model;

[0037] The lesion possibility prediction sub - module is used to input all breast lesion risk factors, the user's potential abnormal breast lesion types and the corresponding potential decision data into the lesion possibility prediction model to predict the lesion possibility of each potential abnormal breast lesion type of the user.

[0038] Preferably, it further includes:

[0039] The visualization display module is used to perform visualization processing on the user's breast health record and the lesion possibilities of all potential abnormal breast lesion types, generate a visualization chart and display it.

[0040] The present invention provides a breast detector, which is used to transmit the user's each detection result to the intelligent analysis system of any one of the above - mentioned breast detectors in real time based on a preset communication protocol or a preset technical interface.

[0041] The beneficial effects of the present invention compared with the prior art are as follows: The breast archive generation module can record the user's previous test results, form a complete breast health archive, and facilitate long-term tracking and comparative analysis of the user's breast health status. This enables doctors and patients to comprehensively understand the development process of breast health and provides complete data support for subsequent analysis and diagnosis. The risk factor identification module helps to comprehensively understand the factors that may lead to breast lesions and provides an important basis for subsequent prediction and early warning. It overcomes the problem of incomplete risk factor identification in the prior art and ensures more accurate assessment of the patient's lesion risk. The lesion type prediction module predicts potential abnormal lesion types and decision-making data through model analysis, providing a direction for diagnosis and treatment in advance. By making full use of historical test data, the accuracy and comprehensiveness of lesion type prediction are improved. The lesion probability prediction module comprehensively considers lesion risk factors, etc., and accurately predicts the probability of each potential lesion type, making the assessment more precise. By comprehensively considering the complex relationships between various factors, it accurately predicts the lesion probability of each potential abnormal breast lesion type of the user, solving the problem of inaccurate lesion probability prediction in the prior art. The early warning mechanism judgment module triggers an early warning in a timely manner when the lesion probability exceeds the threshold, which can remind users and medical staff to take corresponding measures and achieve early intervention and treatment. It reminds patients and doctors to pay attention to potential health risks. This effective early warning trigger mechanism makes up for the defect of the imperfect early warning mechanism in the prior art. This system can effectively manage, predict, and give early warnings to the user's breast health, which helps to improve the prevention and treatment effect of breast diseases.

[0042] Other features and advantages of the present invention will be described in the following specification, and in part, will be obvious from the specification, or can be understood by implementing the present invention. The objectives and other advantages of the present invention can be realized and obtained by the structure specifically pointed out in this application document.

[0043] The following will, through the drawings and embodiments, make a further detailed description of the technical solution of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] The drawings are used to provide a further understanding of the present invention and constitute a part of the specification. They are used together with the embodiments of the present invention to explain the present invention and do not constitute a limitation to the present invention. In the drawings:

[0045] Figure 1 is a schematic diagram of the functional modules of the intelligent analysis system of the breast detector in the embodiment of the present invention;

[0046] Figure 2 is a schematic diagram of the functional sub-modules of the breast archive generation module in the embodiment of the present invention;

[0047] Figure 3Schematic diagram of the functional sub - module of the risk factor identification module in the embodiment of the present invention;

[0048] Figure 4 Schematic diagram of the functional sub - module of the lesion type prediction module in the embodiment of the present invention;

[0049] Figure 5 Schematic diagram of the functional sub - module of the lesion possibility prediction module in the embodiment of the present invention. Detailed implementation manners

[0050] The preferred embodiments of the present invention will be described below with reference to the accompanying drawings. It should be understood that the preferred embodiments described herein are only for the purpose of illustrating and explaining the present invention, and are not used to limit the present invention.

[0051] Embodiment 1:

[0052] The present invention provides an intelligent analysis system for a breast detector. Referring to Figure 1 , including:

[0053] A breast file generation module, which is used to record the previous breast detection results of the user and generate a breast health file of the user;

[0054] A risk factor identification module, which is used to extract all breast lesion risk factors from the breast health file of the user;

[0055] A lesion type prediction module, which is used to analyze the previous breast detection results of the user based on a breast lesion prediction model, predict all potential abnormal breast lesion types of the user, and identify the potential decision data for each potential abnormal breast lesion type;

[0056] A lesion possibility prediction module, which is used to predict the lesion possibility of each potential abnormal breast lesion type of the user based on all breast lesion risk factors, the potential abnormal breast lesion types of the user, and the corresponding potential decision data;

[0057] An early warning mechanism judgment module, which is used to trigger the early warning mechanism when the lesion possibility of the user exceeds the early warning threshold.

[0058] In this embodiment, the previous breast detection results refer to various data and conclusions obtained from multiple previous breast detections of the user, including but not limited to the detection results of the breast tissue structure, physiological indicators, lesion conditions, etc.

[0059] In this embodiment, the breast health file is a comprehensive record of the user's breast health - related information, including previous breast detection results, latest physiological sign information, personal medical history, family medical history, etc., forming a document that comprehensively reflects the user's breast health status.

[0060] In this embodiment, all breast lesion risk factors cover various factors that may cause breast lesions, such as age, genetic factors, lifestyle habits, hormone levels, previous medical history, etc.

[0061] In this embodiment, the breast lesion prediction model is a mathematical model constructed through data analysis and algorithms, which is used to predict the possible types of breast lesions based on the user's detection data and relevant information.

[0062] In this embodiment, all potential abnormal breast lesion types of the user refer to various abnormal lesion categories that may exist in the user's breast predicted through analysis, such as breast hyperplasia, breast fibroadenoma, breast cancer, etc.

[0063] In this embodiment, the potential decision data for each potential abnormal breast lesion type is the decision information based on which each predicted potential abnormal lesion type is obtained, such as the size of the breast mass exceeding the preset size.

[0064] In this embodiment, the lesion possibility for each potential abnormal breast lesion type of the user is an assessment of the probability or likelihood of occurrence for each predicted potential abnormal lesion type.

[0065] In this embodiment, the warning threshold is a preset value or standard. When the lesion possibility exceeds this value, the warning mechanism is triggered.

[0066] In this embodiment, the warning mechanism is a notification and reminder system that is activated when specific conditions are met (such as the lesion possibility exceeding the warning threshold), so as to prompt patients and doctors to take corresponding measures.

[0067] The beneficial effects of the above technology are as follows: The breast health record generation module can record the user's previous test results, forming a complete breast health record, which is convenient for long-term tracking and comparative analysis of the user's breast health status. This enables doctors and patients to comprehensively understand the development process of breast health and provides complete data support for subsequent analysis and diagnosis. The risk factor identification module helps to comprehensively understand the factors that may lead to breast lesions and provides an important basis for subsequent prediction and early warning. It overcomes the problem of incomplete risk factor identification in the prior art and ensures more accurate assessment of the patient's lesion risk. The lesion type prediction module predicts potential abnormal lesion types and decision-making data through model analysis, providing a direction for diagnosis and treatment in advance. By making full use of historical test data, the accuracy and comprehensiveness of lesion type prediction are improved. The lesion probability prediction module comprehensively considers lesion risk factors, etc., and accurately predicts the probability of each potential lesion type, making the assessment more precise. By comprehensively considering the complex relationships between various factors, it accurately predicts the lesion probability of each potential breast abnormal lesion type of the user, solving the problem of inaccurate lesion probability prediction in the prior art. The early warning mechanism judgment module triggers an early warning in a timely manner when the lesion probability exceeds the threshold, which can remind the user and medical staff to take corresponding measures to achieve early intervention and treatment. It reminds patients and doctors to pay attention to potential health risks. This effective early warning trigger mechanism makes up for the defect of the imperfect early warning mechanism in the prior art. This system can effectively manage, predict, and give early warnings to the user's breast health, which helps to improve the prevention and treatment effect of breast diseases.

[0068] Embodiment 2:

[0069] Based on Embodiment 1, the breast health record generation module refers to Figure 2 , including:

[0070] The detection data acquisition sub-module is used to obtain in real time the detection data generated by the breast detector for the user's breast tissue as the user's single breast test result;

[0071] The data processing and record generation sub-module is used to normalize the user's each breast test result, store it in the local storage device or cloud service area according to the preset storage format and database architecture, and summarize the user's latest physiological signs information, personal medical history, and family medical history to construct the user's breast health record.

[0072] In this embodiment, the detection data generated by the breast detector for the user's breast tissue refers to the specific data obtained after using the breast detector to detect the user's breast, such as quantitative values or image information in aspects such as the shape, structure, and blood flow of the breast.

[0073] In this embodiment, normalizing the breast detection results of the user means organizing, converting, and unifying the original results obtained from each breast detection according to certain standards and rules to make them consistent and comparable. For example, unifying data in different units into standard units and performing standardization processing on images, etc.

[0074] In this embodiment, it is stored in a local storage device or a cloud service area according to a preset storage format and database architecture. According to a pre-set specific data organization form and structure, relevant data is saved in a local storage device (such as a computer hard drive) or a cloud server area for subsequent query, invocation, and analysis.

[0075] In this embodiment, the user's latest physiological sign information, personal medical history, and family medical history. The latest physiological sign information includes the user's current physical condition indicators, such as body temperature, blood pressure, etc.; the personal medical history is the disease situation that the user has suffered from himself; the family medical history is the record of relevant diseases suffered by relatives in the user's family.

[0076] The beneficial effects of the above technical solutions are as follows: The detection data acquisition sub-module can obtain the latest detection data in real time to ensure the timeliness and accuracy of the file information. The data processing and file generation sub-module normalizes the detection results to ensure the consistency and usability of the data. Stored locally or in the cloud according to a preset format, which is convenient for data management and invocation, and at the same time improves the security and reliability of the data. Summarize the user's latest physiological signs, personal medical history, family medical history and other information to make the breast health file more comprehensive and comprehensive. The breast file generation module can establish a complete, accurate and easy-to-manage breast health file for the user, providing strong support for subsequent analysis and prediction.

[0077] Embodiment 3:

[0078] On the basis of Embodiment 1, the risk factor identification module refers to Figure 3 , including:

[0079] The original atlas acquisition sub-module is used to acquire a preset breast lesion risk factor atlas, where the preset breast health atlas is used to define the ontology structure of breast lesion risk factors;

[0080] The file data mapping and extraction sub-module is used to perform an association mapping between the user's breast health file and the preset breast health atlas, and use entity recognition and relationship extraction methods to identify information corresponding to the original breast lesion risk factors in the preset breast health atlas from the user's breast health file, and establish the relationship between all the information corresponding to the original breast lesion risk factors identified in the user's breast health file as the user's personalized breast lesion risk factor atlas;

[0081] The lesion risk factor extraction sub-module is used to obtain all breast lesion risk factors of a user based on the user's personalized breast lesion risk factor map. This is a complex processing process. First, the user's breast health record is associated and corresponded with a pre-set standard map, and then specific technologies (such as entity recognition and relationship extraction) are used to find information related to the original risk factors in the pre-set map from the user record, and the associations between these information are sorted out, so as to construct a unique personalized map reflecting the user's breast lesion risk factors for this user.

[0082] In this embodiment, the pre-set breast lesion risk factor map is a chart or set that is pre-set and contains various risk factors that may cause breast lesions, providing standards and references for subsequent risk factor identification and analysis.

[0083] In this embodiment, the ontology structure of breast lesion risk factors refers to the internal basic composition and organizational form of breast lesion risk factors, including definitions and descriptions of aspects such as the classification, hierarchical relationship, and attributes of risk factors.

[0084] The beneficial effects of the above technical solutions are as follows: The original map acquisition sub-module provides a clearly defined breast lesion risk factor map, providing standards and references for subsequent identification work. The file data mapping and extraction sub-module can accurately extract relevant information from the user file through association mapping and advanced identification and extraction methods, and establish the relationships between the information to form a personalized map. The lesion risk factor extraction sub-module can comprehensively obtain the user's breast lesion risk factors based on the personalized map, improving the accuracy and pertinence of risk assessment. The risk factor identification module can effectively extract accurate and personalized breast lesion risk factors from the user's breast health record, providing an important basis for subsequent lesion prediction and early warning.

[0085] Embodiment 4:

[0086] Based on Embodiment 3, the lesion risk factor extraction sub-module includes:

[0087] The academic data mapping and extraction unit is used to associate and map the preset breast diagnosis expert academic library with the preset breast health map, and use entity recognition and relationship extraction methods to identify the information corresponding to the original breast lesion risk factors in the preset breast health map from the preset breast diagnosis expert academic library, and establish the relationship between the information corresponding to all the original breast lesion risk factors identified in the preset breast diagnosis expert academic library, as the original breast lesion risk factor map of the preset breast diagnosis expert academic library; this statement describes a comprehensive processing process. First, the preset breast diagnosis expert academic library and the preset breast health map are associated so that there is a corresponding relationship between them. Then, the entity recognition and relationship extraction methods are used to find out the information related to the original breast lesion risk factors in the preset breast health map from the breast diagnosis expert academic library. Finally, the relationship between the information corresponding to all these original breast lesion risk factors identified from the academic library is sorted out and established to form an original breast lesion risk factor map specifically for the preset breast diagnosis expert academic library. This map can clearly show the association and hierarchy between various original breast lesion risk factors in the academic library, providing a basis for subsequent analysis and application;

[0088] The lesion risk factor extraction unit is used to analyze the weight of each cross-class risk factor group in the preset breast diagnosis expert academic library based on the original breast lesion risk factor map of the preset breast diagnosis expert academic library, and combine the breast lesion diagnosis information corresponding to all groups of cross-class risk factor information in the preset breast diagnosis expert academic library of each group of cross-class risk factors to analyze all breast lesion risk factors of the user.

[0089] In this embodiment, the preset breast diagnosis expert academic database is a pre-established database that collects academic research results, experience summaries and related knowledge of experts in the field of breast diagnosis.

[0090] In this embodiment, the cross-category risk factor group refers to a set of multiple risk factors that are of different categories but are interrelated and jointly affect the risk of breast lesions.

[0091] In this embodiment, the breast lesion diagnosis information corresponding to all groups of cross-class risk factor information in the preset breast diagnosis expert academic library for each group of cross-class risk factors refers to all relevant information contained in each group of cross-class risk factors in the preset breast diagnosis expert academic library, as well as a description of the corresponding relationship between this information and the breast lesion diagnosis results.

[0092] The beneficial effects of the above technical solutions are as follows: The academic data mapping and extraction unit maps and associates the preset breast diagnosis expert academic database with the preset atlas, making full use of expert academic resources to provide a more authoritative and comprehensive reference for the extraction of risk factors. By using entity recognition and relationship extraction methods, relevant information can be accurately extracted from the academic database and relationships can be established to form an authoritative original breast lesion risk factor atlas. The lesion risk factor extraction unit can more scientifically and accurately extract the breast lesion risk factors of the user based on the atlas analysis weights and diagnostic information of the academic database. Combining the weights and diagnostic information of the expert academic database makes the extraction of risk factors more scientific and reliable, improving the accuracy of lesion prediction. The lesion risk factor extraction sub-module can make full use of academic resources to accurately extract the breast lesion risk factors of the user, providing strong support for the assessment and management of breast health.

[0093] Example 5:

[0094] Based on Example 4, the lesion risk factor extraction unit includes:

[0095] The cross-class factor information recognition sub-unit is used to identify all cross-class risk factor information groups of each cross-class risk factor group in the preset breast diagnosis expert academic database based on the original breast lesion risk factor atlas of the preset breast diagnosis expert academic database;

[0096] The cross-class factor weight calculation sub-unit is used to calculate the weight of each cross-class risk factor group in the preset breast diagnosis expert academic database based on all cross-class risk factor information groups of each cross-class risk factor group in the preset breast diagnosis expert academic database;

[0097] The inferable factor screening sub-unit is used to regard each cross-class risk factor group with a weight exceeding the weight threshold as an inferable cross-class risk factor group;

[0098] The lesion risk factor recognition sub-unit is used to analyze all breast lesion risk factors of the user based on all cross-class risk factor information groups of all inferable cross-class risk factor groups, the corresponding breast lesion diagnosis information in the preset breast diagnosis expert academic database, and the user's personalized breast lesion risk factor atlas.

[0099] In this embodiment, based on the original breast lesion risk factor atlas of the preset breast diagnosis expert academic database, all cross-class risk factor information groups of each cross-class risk factor group are identified in the preset breast diagnosis expert academic database: According to the existing original breast lesion risk factor atlas, all specific risk factor information included in each cross-class risk factor group is found from the preset breast diagnosis expert academic database.

[0100] In this embodiment, based on all the cross-category risk factor information groups of each cross-category risk factor group in the preset breast diagnosis expert academic database, the weight of each cross-category risk factor group in the preset breast diagnosis expert academic database is calculated: according to all the information obtained for each cross-category risk factor group, the relative importance or influence value of this cross-category risk factor group in the entire academic database is obtained through a specific calculation method. For example:

[0101] Suppose we have n cross-category potential factor groups. For the i-th cross-category potential factor group (A i ,B i ), the information set it contains is S i , where S i contains various data related to this group of factor pairs, such as the occurrence frequency in different cases, the degree of association with breast lesions, etc.;

[0102] Consider the influence of the occurrence frequency of each cross-category potential factor group in the academic database on its weight. Let f(A i ,B i ) represent the number of times the i-th cross-category potential factor group (A i ,B i ) appears in the preset breast diagnosis expert academic database, and N is the total number of times all cross-category potential factor pairs appear in the academic database. Then the preliminary weight ω f,i based on frequency can be calculated by the following formula:

[0103]

[0104] This formula simply and intuitively reflects the relative frequency of each group of factor pairs appearing. The higher the occurrence frequency, the greater the preliminary weight;

[0105] Next, in combination with the breast lesion diagnosis information in the preset breast diagnosis expert academic database, the weight is further adjusted. Suppose for the i-th cross-category potential factor group (A i ,B i ), among all the cases involving this group of factor pairs, the number of cases with a positive breast lesion diagnosis result (i.e., there is a lesion) is p(A i ,B i ), and the total number of cases is n(A i ,B i ). We introduce a lesion correlation coefficient r(A i ,B i ) to measure the degree of association between this group of factor pairs and breast lesions. The calculation formula is as follows:

[0106]

[0107] Then, considering the frequency weight ω f,iand the lesion correlation coefficient r(A i ,B i ), calculate the adjusted weight ω i :

[0108] ω i = ω f,i * r(A i ,B i )

[0109] This formula combines the occurrence frequency of the factor pair with the degree of lesion association, making the weight more accurately reflect the importance of this group of factor pairs in breast lesion risk judgment;

[0110] To more comprehensively consider the influence of the interaction between factors on the weight, we can introduce an interaction matrix M of factor pairs. For the i-th cross-category potential factor group (A i ,B i ) and the j-th cross-category potential factor group (A j ,B j ), the matrix element M ij represents the interaction strength between these two groups of factor pairs, and its value range can be [-1, 1]. A positive value indicates a positive correlation, a negative value indicates a negative correlation, and the larger the absolute value, the stronger the interaction. The final weight W i after considering the interaction between factors can be calculated by the following formula:

[0111]

[0112] This formula comprehensively considers the frequency and lesion association degree of each group of factor pairs itself, as well as the interaction with all other factor pairs, so as to more accurately calculate the weight of each cross-category potential factor group in the preset breast diagnosis expert academic library. Through the above steps and formulas, based on all the cross-category potential factor information of each group of cross-category potential factors in the preset breast diagnosis expert academic library, the weight of each cross-category potential factor group in the preset breast diagnosis expert academic library can be calculated more comprehensively and accurately, providing strong support for accurately analyzing the breast lesion risk factors of users subsequently.

[0113] In this embodiment, the weight threshold is a preset weight numerical standard for distinguishing cross-category risk factor groups of different importance levels.

[0114] In this embodiment, all the cross-category risk factor information groups of the inferable cross-category risk factor group: refer to all the risk factor information included in the cross-category risk factor group identified as having analyzable and inferable value.

[0115] In this embodiment, the breast lesion diagnosis information corresponding to the preset breast diagnosis expert academic library refers to the information regarding the diagnosis of breast lesions associated with specific risk factors or groups of risk factors in the preset breast diagnosis expert academic library, such as possible lesion types, severity levels, etc.

[0116] The beneficial effects of the above technical solutions are as follows: The cross-category factor information recognition subunit can comprehensively recognize the information of each cross-category risk factor group in the preset academic library, laying a foundation for subsequent weight calculation. The cross-category factor weight calculation subunit scientifically calculates the weights of each cross-category risk factor group, which helps to distinguish their importance levels. The inferable factor screening subunit screens out the inferable cross-category risk factor groups with higher weight values, focusing on key factors and improving the analysis efficiency. The lesion risk factor recognition subunit comprehensively analyzes the information of the inferable cross-category risk factor groups, the diagnosis information in the academic library, and the user's personalized atlas, accurately analyzes the breast lesion risk factors of the user, and improves the accuracy and reliability of risk assessment. The lesion risk factor extraction unit can accurately and effectively extract the key breast lesion risk factors related to the user through systematic analysis and screening, providing a strong basis for subsequent lesion prediction and prevention.

[0117] Embodiment 6:

[0118] Based on Embodiment 5, the lesion risk factor recognition subunit includes:

[0119] An inference original information extraction end, which is used to regard all cross-category risk factor information groups of each inferable cross-category risk factor group and the corresponding breast lesion diagnosis information in the preset breast diagnosis expert academic library as the original information of the breast lesion risk inference factors of the corresponding inferable cross-category risk factor group;

[0120] A factor regression analysis end, which is used to perform a regression analysis on the original information of the breast lesion risk inference factors of each inferable cross-category risk factor group to obtain the inference relationship between each inferable cross-category risk factor group and the corresponding breast lesion risk inference factors;

[0121] An inference factor mining and verification end, which is used to utilize the inference relationships between all inferable cross-category risk factor groups and the corresponding breast lesion risk inference factors to verify and mine all cross-category risk factor information groups of each inferable cross-category risk factor group included in the user's personalized breast lesion risk factor atlas, and obtain all breast lesion risk inference factors in the user's personalized breast lesion risk factor atlas;

[0122] A lesion risk factor summarization end, which is used to regard all breast lesion risk inference factors and the original breast lesion risk factors in the user's personalized breast lesion risk factor atlas as all breast lesion risk factors of the user.

[0123] In this embodiment, regression analysis is performed on the original information of the breast lesion risk inference factors for each inferable cross-class risk factor group to obtain the inference relationship between each inferable cross-class risk factor group and the corresponding breast lesion risk inference factor: By using the statistical method of regression analysis to process the initial data on the breast lesion risk inference factors in each inferable cross-class risk factor group, the inherent logic and association pattern existing between this risk factor group and the corresponding breast lesion risk inference factor can be obtained.

[0124] In this embodiment, by using the inference relationships between all inferable cross-class risk factor groups and the corresponding breast lesion risk inference factors, all cross-class risk factor information groups of each inferable cross-class risk factor group included in the user's personalized breast lesion risk factor map are verified and mined to obtain all breast lesion risk inference factors in the user's personalized breast lesion risk factor map: The obtained inference relationships are used to test and deeply explore all the information of each inferable cross-class risk factor group in the user's personalized breast lesion risk factor map, so as to find out all the inference factors related to breast lesion risk among them.

[0125] In this embodiment, all breast lesion risk inference factors in the user's personalized breast lesion risk factor map refer to the breast lesion risk factors that can be inferred based on all the existing original breast lesion risk factors in the breast lesion risk factor map constructed according to the user's individual situation.

[0126] In this embodiment, the original breast lesion risk factors refer to the basic risk factors that are initially set or have not been deeply analyzed and processed and may cause breast lesions.

[0127] The beneficial effects of the above technical solutions are as follows: The inference original information extraction end provides the necessary original data for subsequent analysis, ensuring that the basis of the analysis is comprehensive and accurate. The factor regression analysis end obtains the relationship between the inferable cross-class risk factor group and the breast lesion risk inference factor through regression analysis, making the analysis more scientific and logical. The inference factor mining and verification end uses the inference relationship to verify and mine the user's personalized map, and can accurately find out the lesion risk inference factors unique to the user. The lesion risk factor summary end integrates the mined risk inference factors and the original risk factors, comprehensively covering the user's breast lesion risk factors and improving the integrity and accuracy of risk assessment. The lesion risk factor identification subunit can accurately and comprehensively identify the user's breast lesion risk factors through a rigorous process and scientific method, providing strong support and guarantee for breast health management.

[0128] Example 7:

[0129] Based on Example 1, the lesion type prediction module refers toFigure 4 , including:

[0130] The first model building sub-module is used to build a breast lesion prediction model;

[0131] The lesion type prediction sub-module is used to input the user's previous breast detection results into the breast lesion prediction model, and predict all potential abnormal lesion types of the user's breast and the potential decision data of each potential abnormal lesion type of the breast.

[0132] In this embodiment, "building a breast lesion prediction model" refers to creating a mathematical model that can predict whether there are breast lesions and possible lesion types based on input breast-related data (such as detection results, user information, etc.) (such as the potential decision data of each potential abnormal lesion type of the breast output) through a series of operations such as selecting appropriate algorithms, collecting and organizing relevant data, and performing data preprocessing, feature engineering, model training and optimization. This model is usually based on machine learning or deep learning technologies, such as decision trees, neural networks, etc. For example, a large number of breast detection sample data can be used to extract key features such as the morphological features and physiological indicators of breast tissue, use these features as inputs, and the corresponding lesion results as outputs to train a neural network model to achieve the prediction of breast lesions for new input data.

[0133] The beneficial effects of the above technical solutions are as follows: The breast lesion prediction model built by the first model building sub-module provides an effective tool and method for the prediction of lesion types. The lesion type prediction sub-module uses this model to analyze the user's previous detection results, and can quickly and accurately predict the potential abnormal lesion types and related decision data. Through model prediction, the efficiency and accuracy of lesion type prediction are improved, and the subjectivity and error of manual judgment are reduced. It provides early warning and decision support for users and medical staff, and helps to take corresponding preventive and treatment measures in a timely manner. The lesion type prediction module can effectively predict the potential abnormal lesion types of the breast, providing an important reference basis for breast health management.

[0134] Embodiment 8:

[0135] Based on Embodiment 1, the lesion probability prediction module refers to Figure 5 , including:

[0136] The second model building sub-module is used to build a lesion probability prediction model;

[0137] The lesion probability prediction sub-module is used to input all breast lesion risk factors, the user's potential abnormal lesion types of the breast and the corresponding potential decision data into the lesion probability prediction model, and predict the lesion probability of each potential abnormal lesion type of the user's breast.

[0138] In this embodiment, "building a lesion possibility prediction model" means constructing a mathematical model through a series of technical means and data processing that can estimate the possibility of each potential abnormal lesion type based on given relevant factors (such as breast lesion risk factors, potential abnormal lesion types, potential decision-making data, etc.). This process involves the following steps: First, determine the variables to be included in the model, that is, those factors that may affect the lesion possibility; then collect a large amount of relevant data, clean, preprocess, and perform feature engineering on the data to convert the data into a format suitable for model learning; then select a suitable modeling method, such as logistic regression, random forest, etc.; then use the prepared data to train and optimize the model to improve the accuracy and generalization ability of the model; finally, obtain a model that can effectively predict the lesion possibility. For example, taking various breast lesion risk factors (such as age, family history, hormone levels, etc.), the already predicted potential abnormal lesion types (such as breast hyperplasia, breast fibroids, etc.), and the corresponding potential decision-making data as inputs, and taking the probability of occurrence of each lesion type as the output, train a prediction model based on random forest.

[0139] The beneficial effects of the above technical solutions are as follows: The lesion possibility prediction model built by the second model building sub-module provides technical support for accurately predicting the lesion possibility. The lesion possibility prediction sub-module can comprehensively analyze various factors, improving the comprehensiveness and accuracy of the prediction. By predicting the lesion possibility through the model, the limitations of relying solely on experience judgment are avoided, making the results more scientific and reliable. It can provide a quantitative assessment of the lesion possibility for users and medical staff in advance, which helps to formulate more targeted prevention and treatment plans. The lesion possibility prediction module can effectively predict the lesion possibility of each potential abnormal breast lesion type, providing an important decision-making basis for the prevention and control of breast diseases.

[0140] Embodiment 9:

[0141] Based on Embodiment 1, it further includes:

[0142] Visualization display module, which is used to visually process the user's breast health records and the lesion probabilities of all types of potential breast abnormal lesions, generate a visualization chart and display it. It can convert various record information about the user's breast health and data such as the probability of each potential abnormal lesion type into intuitive and easy-to-understand graphics, tables or other visual forms, and then display them. For example, the historical breast examination results, relevant physiological indicators and other health record data of the user can be presented in the form of a line chart or a bar chart to show their changing trends; for the lesion probabilities of different types of potential breast abnormal lesions, a pie chart can be used to show their proportions, or different colored areas can be used to mark the distribution of the lesion probabilities on the map. Through such visual processing and display, both patients and doctors can more quickly and clearly understand and grasp the user's breast health status and potential risks.

[0143] The beneficial effects of the above technical solutions are as follows: The visualization display module can convert complex breast health records and lesion probability data into intuitive visualization charts. This visual processing facilitates users and medical staff to quickly understand and obtain key information without spending a lot of time analyzing complex data. Through clear and intuitive display, it helps users to better understand their own breast health status and improve their self-health management awareness. For medical staff, the visualization chart can be more intuitively compared and analyzed, providing a more convenient reference for the formulation of diagnosis and treatment plans. The visualization display module enhances the effect of information transmission and improves the efficiency and quality of breast health management.

[0144] Embodiment 10:

[0145] The present invention provides a breast detector, which is used to transmit the user's each detection result to the intelligent analysis system of any one of the breast detectors in Embodiments 1 to 9 in real time based on a preset communication protocol or a preset technical interface.

[0146] In this embodiment, the preset communication protocol refers to the rules and standards preset for regulating data transmission between the breast detector and the intelligent analysis system. It includes regulations on aspects such as the data format, transmission rate, error checking method, etc., to ensure that data can be accurately, completely and orderly transmitted between the two. For example, common communication protocols such as TCP / IP, Bluetooth protocol, etc.

[0147] In this embodiment, the preset technical interface refers to the technical specifications and methods for connecting and interacting between the breast detector and the intelligent analysis system defined in advance. It stipulates the data input and output methods, the called function interfaces, the parameter passing, etc., so that the breast detector can smoothly transmit the detection results to the intelligent analysis system. For example, it may be a specific USB interface specification, an API of a network interface, etc.

[0148] The beneficial effects of the above technical solutions are as follows: It can transmit the detection results in real time, ensuring the timeliness and freshness of the data, and providing guarantee for subsequent rapid analysis and processing. Based on the preset communication protocol or technical interface, the stability and compatibility of the transmission are ensured, and the risk of data transmission errors and interruptions is reduced. The effective connection with the intelligent analysis system realizes the seamless docking of detection and analysis, improving the operation efficiency of the entire system. The real-time transmission function helps to detect problems in a timely manner and take corresponding measures, enhancing the timeliness and accuracy of breast detection and diagnosis. The real-time transmission function of this breast detector strengthens the coordination and effectiveness of the entire breast detection system, providing users with better services and more reliable diagnostic results.

[0149] Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalent technologies, the present invention is also intended to include these changes and modifications.

Claims

1. An intelligent analysis system for a breast detector, characterized in that: include: Breast record generation module, used to record the user's previous breast test results and generate the user's breast health record; The risk factor identification module is used to extract all breast disease risk factors from the user's breast health records; The lesion type prediction module is used to analyze the user's previous breast test results based on the breast lesion prediction model, predict all the user's potential breast abnormal lesion types, and identify the potential decision data for each potential breast abnormal lesion type; A lesion possibility prediction module is used to predict the lesion possibility of each type of potential abnormal breast lesion of the user based on all breast lesion risk factors and the user's potential abnormal breast lesion type and corresponding potential decision data; The early warning mechanism judgment module is used to trigger the early warning mechanism when the possibility of the user's lesion exceeds the early warning threshold.

2. The intelligent analysis system of the breast detector according to claim 1, characterized in that: Breast archive generation module, including: The detection data acquisition submodule is used to obtain in real time the detection data generated by the breast detector for detecting the breast tissue of the user as the single breast detection result of the user; The data processing and file generation submodule is used to standardize the user's breast test results each time, and store them in the local storage device or cloud service area according to the preset storage format and database architecture, and summarize the user's latest physiological signs information and personal medical history and family medical history to build the user's breast health file.

3. The intelligent analysis system of the breast detector according to claim 1, characterized in that: Risk factor identification module, including: The original atlas acquisition submodule is used to acquire a preset breast lesion risk factor atlas, wherein the preset breast health atlas is used to define the ontology structure of the breast lesion risk factor; The archive data mapping and extraction submodule is used to associate and map the user's breast health archive with the preset breast health map, and use entity recognition and relationship extraction methods to identify information corresponding to the original breast lesion risk factors in the preset breast health map from the user's breast health archive, and establish the relationship between the information corresponding to all the original breast lesion risk factors identified in the user's breast health archive as the user's personalized breast lesion risk factor map; The lesion risk factor extraction submodule is used to obtain all breast lesion risk factors of the user based on the user's personalized breast lesion risk factor map.

4. The intelligent analysis system of the breast detector according to claim 3, characterized in that: The lesion risk factor extraction submodule includes: The academic data mapping and extraction unit is used to associate and map the preset breast diagnosis expert academic library with the preset breast health atlas, and use entity recognition and relationship extraction methods to identify information corresponding to the original breast lesion risk factors in the preset breast health atlas from the preset breast diagnosis expert academic library, and establish the relationship between the information corresponding to all the original breast lesion risk factors identified in the preset breast diagnosis expert academic library as the original breast lesion risk factor atlas of the preset breast diagnosis expert academic library; The lesion risk factor extraction unit is used to analyze the weight of each cross-class risk factor group in the preset breast diagnosis expert academic library based on the original breast lesion risk factor map of the preset breast diagnosis expert academic library, and combine the breast lesion diagnosis information corresponding to all groups of cross-class risk factor information in the preset breast diagnosis expert academic library of each group of cross-class risk factors to analyze all breast lesion risk factors of the user.

5. The intelligent analysis system of the breast detector according to claim 4, characterized in that: Lesion risk factor extraction unit, including: A cross-category factor information identification subunit is used to identify all cross-category risk factor information groups of each cross-category risk factor group in the preset breast diagnosis expert academic database based on the original breast lesion risk factor atlas of the preset breast diagnosis expert academic database; The cross-category factor weight calculation subunit is used to calculate the weight of each cross-category risk factor group in the preset breast diagnosis expert academic library based on all cross-category risk factor information groups of each cross-category risk factor group in the preset breast diagnosis expert academic library; an inferable factor screening subunit, used for treating each cross-category risk factor group whose weight exceeds a weight threshold as an inferable cross-category risk factor group; The lesion risk factor identification subunit is used to analyze all breast lesion risk factors of the user based on all cross-category risk factor information groups of all inferable cross-category risk factor groups and the corresponding breast lesion diagnosis information in the preset breast diagnosis expert academic library and the user's personalized breast lesion risk factor map.

6. The intelligent analysis system of the breast detector according to claim 5, characterized in that: The subunit for identifying lesion risk factors includes: The inference original information extraction end is used to treat all cross-class risk factor information groups of each inferable cross-class risk factor group and the corresponding breast lesion diagnosis information in the preset breast diagnosis expert academic database as the original information of the breast lesion risk inference factor of the corresponding inferable cross-class risk factor group; The factor regression analysis end is used to perform regression analysis on the original information of the breast lesion risk inference factor of each inferable cross-class risk factor group, and obtain the inference relationship between each inferable cross-class risk factor group and the corresponding breast lesion risk inference factor; The inference factor mining verification end is used to verify and mine all cross-class risk factor information groups of each inferable cross-class risk factor group contained in the user's personalized breast lesion risk factor map by utilizing the inference relationship between all inferable cross-class risk factor groups and corresponding breast lesion risk inference factors, so as to obtain all breast lesion risk inference factors in the user's personalized breast lesion risk factor map; The lesion risk factor aggregation terminal is used to treat all breast lesion risk inference factors and original breast lesion risk factors in the user's personalized breast lesion risk factor map as all breast lesion risk factors of the user.

7. The intelligent analysis system of the breast detector according to claim 1, characterized in that: Lesion type prediction module, including: The first model building submodule is used to build a breast lesion prediction model; The lesion type prediction submodule is used to input the user's previous breast test results into the breast lesion prediction model to predict all the user's potential breast abnormal lesion types and potential decision data for each potential breast abnormal lesion type.

8. The intelligent analysis system of the breast detector according to claim 1, characterized in that: Lesion probability prediction module, including: The second model building submodule is used to build a lesion possibility prediction model; The lesion possibility prediction submodule is used to input all breast lesion risk factors and the user's potential abnormal breast lesion types and corresponding potential decision data into the lesion possibility prediction model to predict the lesion possibility of each potential abnormal breast lesion type of the user.

9. The intelligent analysis system of the breast detector according to claim 1, characterized in that: Also includes: The visualization display module is used to visualize the user's breast health record and the possibility of all types of potential abnormal breast lesions, generate visualization charts and display them.

10. A breast detector, characterized in that: An intelligent analysis system for transmitting each test result of the user in real time to the breast detector according to any one of claims 1 to 9 based on a preset communication protocol or a preset technical interface.

Images