Physical examination package generation method and device based on dynamic reasoning, storage medium and computer equipment

Abstract

The application relates to the technical field of data processing, and discloses a physical examination package generation method and device based on dynamic reasoning, a storage medium and computer equipment. The method comprises the following steps: acquiring static feature data and dynamic health data of a target user; determining a single-level abnormal index, a composite abnormal index and a hidden risk index based on the data through a preset risk reasoning model; determining a first correlation degree of a first associated disease corresponding to the single-level abnormal index, determining a second correlation degree of a second associated disease corresponding to the composite abnormal index, and determining a third correlation degree of a third associated disease corresponding to the hidden risk index; determining a first examination item according to the associated diseases, and screening the first examination item based on the correlation degrees to obtain a second examination item; and generating a target physical examination package according to the second examination item. The application can be applied to the physical examination package generation scene in the field of intelligent medical treatment, and the matching degree between the physical examination package and the target user is improved.

Description

Technical Field

[0001] This application relates to the field of data processing technology, and in particular to a method and apparatus for generating physical examination packages based on dynamic reasoning, a storage medium, and a computer device. Background Technology

[0002] In traditional health checkup services, the creation of checkup packages often relies on fixed templates or doctors' experience-based judgment, lacking precise analysis and dynamic adjustments to individual health conditions. For example, routine checkup items are usually set based on general health risks, making it difficult to personalize them for different users' static characteristics such as age, gender, family medical history, and recent changes in health status. This "one-size-fits-all" model may not only lead to users undergoing unnecessary examinations and increasing their financial burden, but may also overlook potential health risks, reducing the actual value of the checkup.

[0003] With the rapid development of smart healthcare, the methods for collecting health data are becoming increasingly diverse, including wearable devices, mobile medical applications, and electronic health records. These methods can acquire users' dynamic health data (such as heart rate, blood pressure, blood sugar, and exercise levels) in real time or periodically. This data makes more accurate health risk assessments possible. However, effectively integrating static characteristic data (such as age, gender, and genetic information) with dynamic health data and extracting clinically significant health risk indicators remains a key issue that urgently needs to be addressed in the field of smart healthcare. Furthermore, users' demands for physical examination services are gradually shifting from "standardized" to "personalized," expecting customized examination plans based on their own health conditions to achieve early prevention and intervention of diseases. Summary of the Invention

[0004] In view of this, this application provides a method, apparatus, storage medium, and computer equipment for generating health checkup packages based on dynamic reasoning. Through the constructed multi-level risk reasoning network, it achieves deep fusion analysis of user static characteristic data and dynamic health data. This not only identifies directly correlated single-level abnormal indicators but also derives composite abnormal indicators from multi-factor interactions, further uncovering uncommon but highly dangerous hidden risk indicators. Thus, it systematically extracts health risk levels with clear clinical guidance significance from multi-source heterogeneous data. Based on this, the necessity of examination items is screened and personalized according to the quantitative correlation between each risk indicator and disease. The resulting health checkup package truly achieves a leap from standardization to personalization, accurately matching the target user's real-time health status and long-term risk characteristics.

[0005] According to one aspect of this application, a method for generating health checkup packages based on dynamic reasoning is provided, comprising: Acquire static feature data and dynamic health data of the target user; Based on the static feature data and the dynamic health data, the single-level abnormality indicators, composite abnormality indicators and hidden risk indicators of the target user are determined by the three-level risk inference network in the preset risk inference model. Determine the first degree of correlation between the single-level abnormal indicator and the first associated disease corresponding to the single-level abnormal indicator, determine the second degree of correlation between the composite abnormal indicator and the second associated disease corresponding to the composite abnormal indicator, and determine the third degree of correlation between the hidden risk indicator and the third associated disease corresponding to the hidden risk indicator; Based on the first associated disease, the second associated disease, and the third associated disease, multiple first items to be examined are determined, and based on the first correlation degree, the second correlation degree, and the third correlation degree, the multiple first items to be examined are screened to obtain multiple second items to be examined. Based on the multiple second items to be checked, a target health check package corresponding to the target user is generated.

[0006] According to another aspect of this application, a device for generating a health checkup package based on dynamic reasoning is provided, comprising: The data acquisition module is used to acquire static feature data and dynamic health data of the target user; The indicator determination module is used to determine the single-level abnormal indicator, composite abnormal indicator and hidden risk indicator of the target user based on the static feature data and the dynamic health data, through the three-level risk inference network in the preset risk inference model. The correlation determination module is used to determine the first correlation degree between the single-level abnormal indicator and the first associated disease corresponding to the single-level abnormal indicator, to determine the second correlation degree between the composite abnormal indicator and the second associated disease corresponding to the composite abnormal indicator, and to determine the third correlation degree between the hidden risk indicator and the third associated disease corresponding to the hidden risk indicator. The examination item determination module is used to determine multiple first examination items based on the first associated disease, the second associated disease, and the third associated disease, and to filter the multiple first examination items based on the first correlation degree, the second correlation degree, and the third correlation degree to obtain multiple second examination items. The health checkup package generation module is used to generate a target health checkup package corresponding to the target user based on the multiple second items to be checked.

[0007] According to another aspect of this application, a storage medium is provided that stores a computer program thereon, which, when executed by a processor, implements the above-described method for generating physical examination packages based on dynamic reasoning.

[0008] According to another aspect of this application, a computer device is provided, including a storage medium, a processor, and a computer program stored on the storage medium and executable on the processor, wherein the processor executes the program to implement the above-described method for generating a health check package based on dynamic reasoning.

[0009] Using the above technical solution, this application provides a method, apparatus, storage medium, and computer equipment for generating health checkup packages based on dynamic reasoning. First, static characteristic data and dynamic health data of the target user are acquired. Then, based on the above data, different levels of risk indicators for the target user are identified through a three-level risk reasoning network in a preset risk reasoning model. On this basis, each level of the risk reasoning network not only identifies the risk indicators but also further determines their associated diseases and the degree of association. Specifically, the first-level risk reasoning network determines the first directly associated disease and the first degree of association based on a single-level abnormal indicator; the second-level risk reasoning network derives the second associated disease and the second degree of association based on composite abnormal indicators; and the third-level risk reasoning network derives the third associated disease and the third degree of association based on hidden risk indicators. Further, based on all the identified associated diseases, several possible examination items are initially determined, i.e., the first examination items to be examined. Subsequently, based on the degree of association between each disease and the risk indicator, the necessity of these first examination items is quantified and screened to select more necessary and personalized second examination items. Finally, based on the set of second examination items obtained after screening, the final target health checkup package for the target user is generated. This application's embodiments, through the constructed multi-level risk reasoning network, achieve deep fusion analysis of user static feature data and dynamic health data. It can not only identify directly correlated single-level abnormal indicators, but also derive composite abnormal indicators from multi-factor interactions, and further uncover uncommon but highly dangerous hidden risk indicators. This systematically extracts health risk levels with clear clinical guidance from multi-source heterogeneous data. Based on this, the necessity of examination items is screened and personalized according to the quantitative correlation between each risk indicator and disease. The resulting health check package truly achieves a leap from standardization to personalization, accurately matching the target user's real-time health status and long-term risk characteristics.

[0010] The above description is only an overview of the technical solution of this application. In order to better understand the technical means of this application and to implement it in accordance with the contents of the specification, and to make the above and other objects, features and advantages of this application more obvious and understandable, the following are specific embodiments of this application. Attached Figure Description

[0011] The accompanying drawings, which are included to provide a further understanding of this application and form part of this application, illustrate exemplary embodiments and are used to explain this application, but do not constitute an undue limitation of this application. In the drawings: Figure 1 A flowchart illustrating a method for generating a health checkup package based on dynamic reasoning, as provided in an embodiment of this application, is shown. Figure 2 A flowchart illustrating a screening method for a second item to be inspected, provided in an embodiment of this application, is shown. Figure 3 A flowchart illustrating a method for generating a target health check package according to an embodiment of this application is shown; Figure 4 This illustration shows a structural schematic diagram of a physical examination package generation device based on dynamic reasoning provided in an embodiment of this application; Figure 5 A schematic diagram of the device structure of a computer device provided in an embodiment of this application is shown. Detailed Implementation

[0012] The present application will be described in detail below with reference to the accompanying drawings and embodiments. It should be noted that, unless otherwise specified, the embodiments and features described in the embodiments of the present application can be combined with each other.

[0013] This embodiment provides a method for generating physical examination packages based on dynamic reasoning, such as... Figure 1 As shown, the method includes: Step 101: Obtain the static feature data and dynamic health data of the target user.

[0014] Step 102: Based on the static feature data and the dynamic health data, determine the single-level abnormality index, composite abnormality index and hidden risk index of the target user through the three-level risk inference network in the preset risk inference model.

[0015] Step 103: Determine the first degree of correlation between the single-level abnormal indicator and the first associated disease corresponding to the single-level abnormal indicator, determine the second degree of correlation between the composite abnormal indicator and the second associated disease corresponding to the composite abnormal indicator, and determine the third degree of correlation between the hidden risk indicator and the third associated disease corresponding to the hidden risk indicator.

[0016] Step 104: Based on the first associated disease, the second associated disease, and the third associated disease, determine multiple first items to be examined, and based on the first correlation degree, the second correlation degree, and the third correlation degree, filter the multiple first items to be examined to obtain multiple second items to be examined.

[0017] Step 105: Generate the target health check package corresponding to the target user based on the plurality of second items to be checked.

[0018] This application provides a method for generating health checkup packages based on dynamic reasoning. First, it acquires static feature data and dynamic health data of the target user. The static feature data may include basic information such as age, gender, occupation, past medical history, and family history of genetic diseases. The dynamic health data may cover indicators such as heart rate, sleep quality, and exercise volume monitored in real time by wearable devices, as well as structured health information extracted from historical health checkup reports.

[0019] Next, based on the above data, different levels of risk indicators for the target user are identified through a three-level risk inference network in the preset risk inference model. The first-level risk inference network is used to extract single-level abnormal indicators; if a single health indicator exceeds the normal range, it is identified as a single-level abnormal indicator. For example, a fasting blood glucose level higher than a threshold directly triggers a diabetes risk assessment, and in this case, the fasting blood glucose level is used as a single-level abnormal indicator. The second-level risk inference network is used to perform composite analysis on multiple associated risk factors. It calculates the combined probability of a specific health risk event through a causal inference model. For example, considering both a user's high body mass index (BMI>28) and sedentary occupation, the network uses a built-in medical risk model (such as a statistical model or knowledge graph trained on large-scale clinical data) to deduce the composite risk probability of the user developing metabolic syndrome (e.g., an increase of 72%). When the calculated composite risk probability exceeds a preset clinical or operational threshold, the combination of "high BMI - sedentary occupation" is formally identified as a composite abnormal indicator requiring attention. The third-level risk inference network is used to identify and analyze those low-frequency but potentially extremely harmful risks. High hidden risk indicators, such scenarios are often referred to as long-tail risks in medicine. For example, considering two independent risk factors simultaneously, such as a user's family history of gastric cancer and a positive Helicobacter pylori test in a recent physical examination, a deeper causal reasoning model (such as a Bayesian network based on medical knowledge graphs and epidemiological data) can be used to deduce the user's current relative risk level of gastric cancer. When the calculated relative risk level exceeds a specific threshold set according to clinical importance, the specific combination of "family history of gastric cancer - positive Helicobacter pylori" is officially identified as a key hidden risk indicator.

[0020] Building upon this foundation, the risk inference networks at each level not only identify risk indicators but also further determine their associated diseases and the degree of association. Specifically, the first-level risk inference network determines the first directly associated disease and the first degree of association based on single-level anomaly indicators; the second-level risk inference network derives the second associated disease and the second degree of association based on composite anomaly indicators; and the third-level risk inference network deduces the third associated disease and the third degree of association based on hidden risk indicators. This process can be achieved using technologies such as causal inference engines, thereby enabling dynamic, multi-level inference from data to disease risk.

[0021] Furthermore, based on all the identified related diseases, several possible examination items were initially determined, namely the first set of examination items. Subsequently, based on the correlation between each disease and risk indicator, the necessity of these first set of examination items was quantified and screened to select more necessary and personalized second set of examination items.

[0022] Finally, based on the second set of items to be checked obtained after screening, a final target health check package is generated for the target user, thus forming a highly personalized health check plan.

[0023] By applying the technical solution of this embodiment, firstly, static feature data and dynamic health data of the target user are acquired. Then, based on the above data, different levels of risk indicators for the target user are identified through a three-level risk inference network in a preset risk inference model. On this basis, each level of the risk inference network not only identifies the risk indicators but also further determines their associated diseases and the degree of association. Specifically, the first-level risk inference network determines the first directly associated disease and the first degree of association based on a single-level abnormal indicator; the second-level risk inference network derives the second associated disease and the second degree of association based on composite abnormal indicators; and the third-level risk inference network derives the third associated disease and the third degree of association based on hidden risk indicators. Further, based on all the identified associated diseases, several possible examination items are initially determined, i.e., the first examination items. Subsequently, based on the degree of association between each disease and the risk indicator, the necessity of these first examination items is quantified and screened to select more necessary and personalized second examination items. Finally, based on the set of second examination items obtained after screening, a final target health check package for the target user is generated. This application's embodiments, through the constructed multi-level risk reasoning network, achieve deep fusion analysis of user static feature data and dynamic health data. It can not only identify directly correlated single-level abnormal indicators, but also derive composite abnormal indicators from multi-factor interactions, and further uncover uncommon but highly dangerous hidden risk indicators. This systematically extracts health risk levels with clear clinical guidance from multi-source heterogeneous data. Based on this, the necessity of examination items is screened and personalized according to the quantitative correlation between each risk indicator and disease. The resulting health check package truly achieves a leap from standardization to personalization, accurately matching the target user's real-time health status and long-term risk characteristics.

[0024] In the embodiments of this application, optionally, as shown... Figure 2 As shown, step 104 includes: Step 104-1: Determine the examination items corresponding to the first associated disease, the second associated disease, and the third associated disease respectively, and merge and deduplicate the determined examination items to obtain multiple first examination items.

[0025] Step 104-2: For each first item to be inspected, determine the target correlation degree corresponding to the first item to be inspected from the first correlation degree, the second correlation degree and the third correlation degree, and determine the target score corresponding to the first item to be inspected based on the target correlation degree.

[0026] Step 104-3: Based on the relationship between the target score and the preset score threshold corresponding to each first item to be inspected, select multiple second items to be inspected from the multiple first items to be inspected.

[0027] In this embodiment, firstly, recommended examination items are determined from the first, second, and third associated diseases, respectively. For example, glycated hemoglobin testing might be recommended for diabetes risk assessment, while an insulin release test might be recommended for metabolic syndrome. Subsequently, these examination items derived from different associated diseases are merged, and deduplication is performed, i.e., identical duplicate items are automatically identified and removed. This step ensures that all potentially relevant examination items are considered, while avoiding redundancy, ultimately forming a comprehensive and non-duplicative first set of examination items, laying the foundation for subsequent refined screening.

[0028] Next, a quantified target score is calculated for each first examination item. Specifically, for each first examination item, a target correlation degree can be determined. For example, a first examination item may be associated with multiple diseases simultaneously. In this case, the most representative or comprehensively calculated value can be selected from the first, second, or third correlation degrees corresponding to each disease associated with the first examination item as its target correlation degree. Then, based on this target correlation degree, the final target score of the first examination item is calculated. In one specific embodiment, the mapping relationship between the target correlation degree and the target score can be preset, or a formula for calculating the target score based on the target correlation degree can be preset. In another specific embodiment, the target score of the first examination item can also be calculated by combining dynamic rules such as the preset base score (from clinical guidelines), gain score (from individual risk adjustment), and inhibition score (from recent examination records) in algorithms such as gradient decision trees. This score comprehensively reflects the necessity, urgency, and cost-effectiveness of the first examination item for the current target user.

[0029] Finally, the target score calculated for each first item to be inspected is compared with a preset score threshold. All first items to be inspected whose target scores reach or exceed the preset score threshold are selected and identified as second items to be inspected; while first items to be inspected that do not reach the preset score threshold are considered insufficiently necessary or not cost-effective under the current circumstances and are excluded. Through this step, the convergence from a massive number of possible inspection items to a set of highly personalized, necessary, and focused core inspection items is achieved.

[0030] This application's embodiments systematically transform the potentially scattered and numerous disease risk alerts output by a multi-level risk inference network into a concise, efficient, and personalized list of items to be checked. This not only ensures that all significant risks are covered by corresponding items to be checked, eliminating omissions, but more importantly, through the dynamic calculation of target scores and the setting of preset score thresholds, it achieves a quantitative assessment of the necessity of the items to be checked and intelligent priority ranking. This effectively solves the problem that the recommended items to be checked in traditional solutions are either too general or redundant, thereby improving the ability to detect diseases early and optimizing the user experience.

[0031] In the embodiments of this application, optionally, as shown... Figure 3 As shown, step 105 includes: Step 105-1: Extract the age and gender of the target user from the static feature data of the target user. Based on the age and gender, determine the first basic inspection item from each of the second inspection items through a preset age-gender baseline model, and generate the first recommended data corresponding to the first basic inspection item.

[0032] Step 105-2: Extract the target user's region from the target user's static feature data; based on the region, determine the second basic examination items from each of the second examination items through a preset regional epidemic disease database; and generate the second recommended data corresponding to the second basic examination items.

[0033] Step 105-3: Generate a basic physical examination package based on the first basic examination items, the first recommended data, the second basic examination items, and the second recommended data.

[0034] Step 105-4: Remove the first basic examination item and the second basic examination item from each of the second examination items to obtain the examination items after removal. Based on the first target-related disease of the examination items after removal, determine the first target indicator corresponding to the first target-related disease. Based on the first target indicator, generate third recommendation data. Based on the examination items after removal and the third recommendation data, generate an additional physical examination package. The first target indicator includes at least one of the single-level abnormal indicator, the composite abnormal indicator, and the hidden risk indicator.

[0035] Step 105-5: Generate the target physical examination package for the target user based on the basic physical examination package and the supplementary physical examination package.

[0036] In this embodiment, firstly, the age and gender of the target user are extracted from their static characteristic data. Based on this information, a preset age-gender baseline model is invoked. This model can be a rule base or a lightweight model built based on large-scale clinical guidelines and epidemiological statistics, capable of mapping a set of recommended basic and universal examination items for different age groups and genders. For example, the preset age-gender baseline model may specify that women aged 30-40 should include breast ultrasound and cervical liquid-based cytology (TCT). According to this model, from the previously screened multiple second examination items, examination items that match the current user's age and gender characteristics are identified and determined as first basic examination items. Corresponding first recommendation data is generated for each first basic examination item, which explains the reason for recommending the first basic examination item.

[0037] Simultaneously, regional information can be extracted from the static characteristic data of target users. Based on this region, a preset regional prevalent disease database is queried. This database integrates the epidemiological characteristics of different regions, such as the high incidence of certain cancers or endemic diseases. According to the regional adaptation strategy in this database, examination items suitable for the common risks of the region are further filtered from the second set of examination items and determined as the second basic examination items. For example, for users in Guangdong, EB virus antibody testing may be automatically included to screen for nasopharyngeal carcinoma risk, and second recommendation data may be generated for these items to explain their regional relevance.

[0038] Subsequently, the first and second basic examination items, along with their respective recommended data, are combined to form the basic health check package for the user. This part ensures that the health check plan covers both general and regional risks.

[0039] Next, personalized supplementary packages will be constructed. Specifically, all items already allocated to the basic health check package (i.e., the first and second basic check items) can be removed from multiple second-stage check items. The remaining check items after removal reflect the unique, non-universal health risks of the target user. For these items, the core risk indicator behind them is determined based on which level of risk inference network initially triggered them, i.e., the first target indicator is found. This indicator may be one or more combinations of single-level abnormal indicators, composite abnormal indicators, or hidden risk indicators. Subsequently, third-level recommendation data is generated for each removed check item, which can clearly explain the reason for the recommendation, such as "Because your occupation is programmer and your BMI is high, it is recommended to add an insulin release test to assess the risk of metabolic syndrome." Based on these highly personalized check items and corresponding interpretive data, supplementary health check packages are generated.

[0040] Finally, the basic health check package and the supplementary health check package are integrated to generate the final target health check package delivered to the target users. This health check package is a well-structured and highly interpretable complete solution, which includes both basic protective examinations based on demographics and geography, as well as in-depth screening programs targeting individual-specific risks.

[0041] This application's embodiments achieve a clear balance and division between standardization and personalization in health checkup packages through a two-tiered "basic-additional" structure design. Specifically, a reliable basic package framework is efficiently constructed using a mature baseline model and regional database, ensuring the accuracy and broad applicability of the health checkup plan. By attributing and explaining the remaining personalized items after removing basic examination items, high-value-added and highly targeted additional packages are precisely created, fundamentally improving the accuracy of the health checkup packages and the user experience.

[0042] Optionally, after step 104, the method further includes: for each second item to be inspected, determining whether the second item to be inspected is a mandatory item, and calculating a basic score corresponding to the second item to be inspected based on the determination result; and determining a second target indicator corresponding to the second target-related disease based on the second target-related disease corresponding to the second item to be inspected, and calculating a gain score corresponding to the second item to be inspected based on the second target indicator, wherein the second target indicator includes at least one of the single-level abnormality indicator, the composite abnormality indicator, and the hidden risk indicator; and determining whether the second item to be inspected is a redundant item based on the historical inspection items within a preset time period of the target user, and calculating a suppression score corresponding to the second item to be inspected based on the determination result; and calculating an item score corresponding to the second item to be inspected based on the basic score, the gain score, and the suppression score. Accordingly, after step 105-5, the method further includes adding the item score corresponding to each second item to be examined to the target physical examination package.

[0043] In this embodiment, after obtaining multiple second items to be tested, for each second item, firstly, it is determined whether the second item is a mandatory item. Specifically, this can be determined based on relevant clinical guidelines, such as the requirement that men over 40 years of age must undergo prostate-specific antigen (PSA) testing. Based on this determination, a base score is calculated for the second item. Mandatory items can obtain a higher base score to ensure their inclusion, while non-mandatory items may obtain a default or lower base score.

[0044] Next, the gain score for the second examination item is calculated to reflect personalized adjustments. Specifically, the associated second-target disease can be traced back to the second examination item, and the root cause of the disease can be further identified. This indicator comes from one or more of the previously identified single-level abnormalities, composite abnormalities, or hidden risks. For example, for a coronary CTA examination, the complete second-target indicator is "age > 50 years, long-term history of hypertension, and persistently elevated LDL cholesterol (composite abnormality indicator)," and the second associated disease is high risk of coronary heart disease. When calculating the gain score for this CTA examination, the gain score can be calculated based on the comprehensive risk probability corresponding to the second-target indicator, ensuring a more comprehensive and accurate calculation. The calculation logic of the gain score is based on the severity and correlation strength of these specific risk indicators; the higher the risk and the stronger the correlation, the higher the gain score is assigned, thereby achieving a dynamic increase in the recommendation weight.

[0045] Additionally, a suppression score can be calculated to avoid unnecessary duplicate checks. Specifically, by querying the target user's historical check items over a preset time period (e.g., six months), it can be determined whether the current second item to be checked has already been completed recently. If it has been checked recently, it can be identified as a redundant item, and a higher suppression score can be calculated for it. This negative score will offset some of the recommendation strength during subsequent aggregation.

[0046] Then, the above base score, gain score and suppression score are combined using a predefined algorithm (such as weighted summation) to calculate the final project score of the second item to be inspected.

[0047] Accordingly, after generating the complete target health check package for the target user, an information enhancement step is performed: the calculated item score corresponding to each second item to be checked included in the target health check package is added to and associated with the package description information. In other words, what is finally delivered to the target user is not only a list of check items, but also a clear report with a detailed recommendation score for each item.

[0048] This application's embodiments introduce a structured, multi-factor fusion project scoring mechanism and present these project scores intuitively in the final output. This transforms the originally "black box" AI recommendation into a traceable and verifiable quantitative process derived from base scores, gain scores, and inhibition scores. This allows target users to clearly understand the specific reasons and urgency of each recommended check, thereby making more informed choices.

[0049] In this embodiment, optionally, the composite anomaly index, the second associated disease, and the second correlation degree are determined as follows: Structured data is extracted from the static feature data and the dynamic attribute data using the secondary risk inference network in the preset risk inference model, and the index name corresponding to each structured data is determined; for each index name corresponding to each structured data, whether the label corresponding to each index name is a composite label is determined using a preset label database, and structured data with composite labels is retained; for each preset composite index, according to the multiple index names included in the preset composite index and the index names corresponding to the retained structured data, the corresponding target structured data is matched from the retained structured data, and it is determined whether the target structured data meets the preset requirements corresponding to each index name in the preset composite index; when the target structured data meets the preset requirements corresponding to each index name in the preset composite index, the composite anomaly index is generated according to the target structured data and the index name corresponding to the target structured data, the associated disease corresponding to the preset composite index is taken as the second associated disease corresponding to the composite anomaly index, and the correlation degree corresponding to the associated disease is taken as the second correlation degree corresponding to the second associated disease.

[0050] In this embodiment, firstly, structured data is extracted from the target user's static characteristic data (such as occupation and family history) and dynamic attribute data (such as recent body mass index and blood pressure) through a two-level risk inference network in a pre-defined risk inference model, and the corresponding indicator name for each structured data is determined. For example, the target user's self-described occupation "programmer" can be structured and stored, and its indicator name can be "occupation type"; while a set of monthly average heart rate values ​​uploaded from a wearable device can be named "resting heart rate". This data standardization process lays the foundation for subsequent logical analysis.

[0051] Next, the system accesses a pre-defined label database, which defines attribute labels for each possible indicator name (such as BMI, occupation type, and blood pressure classification). A key category of these labels is the composite label. If an indicator name is marked as a composite label, it means that it is medically recognized as being able to work in conjunction with other indicators to influence the risk of specific diseases. For example, "BMI" and "occupation type" are typically marked as composite labels because they can jointly infer the risk of metabolic diseases; while purely identifying data such as "user ID" is not labeled. This allows the system to retain only structured data with composite labels, focusing on information that truly has value for combined analysis and improving the efficiency of subsequent processing.

[0052] The system also includes several pre-defined composite indicators, each representing a known health risk pattern resulting from the synergistic effect of multiple factors, and explicitly specifying the names of the included indicators. For example, a pre-defined composite indicator called "High Risk of Metabolic Syndrome" might require both the indicator names "BMI Value" (requirement: >28) and "Occupation Type" (requirement: belonging to the "Sedentary" list). Next, all the retained structured data of the current target user can be compared with these pre-defined composite indicators. A successful match occurs when a set of data can be successfully matched from the retained structured data, and every value in this set of data strictly meets all the requirements of the corresponding pre-defined composite indicator.

[0053] Once a pre-defined composite indicator is successfully matched, a composite abnormality indicator is formally generated, such as "BMI > 28 and sedentary occupation". Simultaneously, a pre-defined medically associated disease (such as "metabolic syndrome") is directly used as the second associated disease for this match. The fixed strength of the relationship between this disease and the composite abnormality indicator, i.e., the pre-defined correlation (e.g., a 72% increase in probability), is used as the second correlation. This mechanism ensures that the risk assessment is based on a validated medical knowledge model, rather than a simple ad-hoc calculation.

[0054] This application's embodiments achieve automated and standardized identification of complex health risks. Its core advantage lies in encapsulating complex multi-factor medical knowledge (i.e., "preset composite indicators") into computable rule templates. This enables the system to efficiently and accurately scan and locate known high-risk combinations in massive user data. This method not only significantly improves the efficiency and consistency of risk discovery and avoids the difficulty of manually formulating complex rules, but also yields second-related diseases and second-related degrees with better interpretability and clinical credibility, providing solid and reliable intermediate inference results for generating accurate health check recommendations.

[0055] In this embodiment of the application, optionally, the hidden risk indicator, the third associated disease, and the third correlation degree are determined based on the following method: Low-frequency data features and high-risk biomarker features are extracted from the static feature data and the dynamic attribute data through the three-level risk reasoning network in the preset risk reasoning model; the low-frequency data features and the high-risk biomarker features are then combined to obtain multiple feature combinations; for each feature combination, a preset knowledge base is invoked, and it is retrieved whether the feature combination is contained in the preset knowledge base; if the feature combination is contained in the preset knowledge base, the associated disease corresponding to the feature combination and the correlation degree corresponding to the associated disease are obtained from the preset knowledge base; the feature combination is used as the hidden risk indicator; and the associated disease corresponding to the feature combination and the correlation degree corresponding to the associated disease are respectively used as the third associated disease and the third correlation degree corresponding to the third associated disease.

[0056] In this embodiment, firstly, two types of special information are extracted from the user's static feature data and dynamic attribute data through a three-level risk inference network in a pre-defined risk inference model: low-frequency data features and high-risk biomarker features. Low-frequency data features refer to individualized information that is uncommon in the general population but has important indicative significance, such as a rare family history of genetic diseases, a specific occupational exposure history (such as asbestos exposure), or a specific regional infection history. High-risk biomarker features specifically refer to physiological, biochemical, or molecular indicators that have been confirmed by medical research to be strongly associated with serious diseases (such as cancer, autoimmune diseases), such as specific gene mutations (such as BRCA1), tumor markers (such as alpha-fetoprotein AFP), or chronic infection markers (such as Helicobacter pylori positivity). Subsequently, these extracted features are arranged and combined, that is, they are systematically attempted to be paired or grouped into small sets to generate a series of feature combinations that may represent complex risk situations, such as combining "family history of gastric cancer" with "Helicobacter pylori positivity".

[0057] Next, these automatically generated feature combinations are validated for medical effectiveness. Specifically, for each generated feature combination, a pre-defined knowledge base integrating authoritative medical literature, clinical guidelines, and expert knowledge is queried. This knowledge base is essentially a calibrated medical relationship graph, storing known, evidence-supported feature combinations-disease correspondences and their risk quantification data. Then, the existence of the current feature combination is checked in this knowledge base. If a record with a perfect or highly matching match exists in the pre-defined knowledge base, it indicates that the feature combination has been confirmed by medical knowledge as a meaningful hidden risk pattern.

[0058] Once a feature combination is successfully matched in the pre-defined knowledge base, the predefined associated disease (e.g., "gastric cancer") corresponding to that feature combination can be directly retrieved from the knowledge base, along with the strength of the association between that disease and the feature combination, i.e., the correlation degree (e.g., a risk enhancement factor of 3 times). Subsequently, the currently verified feature combination is identified as a hidden risk indicator, and the disease and correlation degree retrieved from the pre-defined knowledge base are used as the final third associated disease and third correlation degree, respectively. In this way, a vague feature combination derived from the user's original data is transformed into an executable judgment with clear clinical indication and a quantifiable risk level.

[0059] This application's embodiments achieve automated, evidence-driven mining and assessment of long-tail health risks. Its core advantage lies in breaking through the limitations of traditional methods that rely on common diseases and big data models. It can proactively discover and quantify atypical but highly dangerous risk intersections in individuals. This method greatly enhances the early identification capability of rare diseases and specific high-risk groups. Furthermore, because all judgments strictly follow a preset knowledge base, it ensures the reliability and traceability of the reasoning process.

[0060] Optionally, after step 105, the method further includes: obtaining the examination results of each second examination item in the target physical examination package; if the examination results indicate that the target user has a corresponding disease risk, then determining the target indicator corresponding to the second examination item, and increasing the target correlation degree of the target related disease corresponding to the target indicator.

[0061] In this embodiment, after generating a target health checkup package for the target user, the structured test results corresponding to each second test item in the target health checkup package can be obtained. These test results are no longer simple normal or abnormal indicators, but include specific quantitative values ​​(such as blood glucose levels) or qualitative descriptions (such as descriptions of imaging findings), providing a data foundation for subsequent in-depth analysis.

[0062] When analyzing the results of these tests, if a particular test result clearly indicates that the target user has a corresponding disease risk (e.g., polyps found during colonoscopy, or significantly elevated tumor markers), it is possible to trace back and identify the root cause of the initial recommendation for that second test. This root cause is the specific risk factor initially identified by the risk inference network that led to the recommendation of the test; it could be a single abnormal indicator, a combination of abnormal indicators, or a hidden risk indicator.

[0063] Next, the target correlation degree of the target associated disease corresponding to the target indicator is increased. For example, if an insulin release test is recommended based on the composite abnormal indicator of "programmer occupation + high BMI", and the target user is indeed found to have insulin resistance, then the correlation degree of "metabolic syndrome" as the target associated disease corresponding to this composite abnormal indicator is appropriately increased.

[0064] This application's embodiments introduce a reinforcement learning mechanism of "examination result feedback - correlation adjustment," which can continuously revise and calibrate its internal risk reasoning logic using real medical results. This makes the system no longer solely reliant on an initial medical knowledge base, but also learn from continuous service practice, thereby making future risk predictions and package recommendations increasingly accurate and in line with real disease occurrence patterns.

[0065] Furthermore, as Figure 1 In terms of specific implementation, this application provides a device for generating physical examination packages based on dynamic reasoning, such as... Figure 4 As shown, the device includes: The data acquisition module is used to acquire static feature data and dynamic health data of the target user; The indicator determination module is used to determine the single-level abnormal indicator, composite abnormal indicator and hidden risk indicator of the target user based on the static feature data and the dynamic health data, through the three-level risk inference network in the preset risk inference model. The correlation determination module is used to determine the first correlation degree between the single-level abnormal indicator and the first associated disease corresponding to the single-level abnormal indicator, to determine the second correlation degree between the composite abnormal indicator and the second associated disease corresponding to the composite abnormal indicator, and to determine the third correlation degree between the hidden risk indicator and the third associated disease corresponding to the hidden risk indicator. The examination item determination module is used to determine multiple first examination items based on the first associated disease, the second associated disease, and the third associated disease, and to filter the multiple first examination items based on the first correlation degree, the second correlation degree, and the third correlation degree to obtain multiple second examination items. The health checkup package generation module is used to generate a target health checkup package corresponding to the target user based on the multiple second items to be checked.

[0066] Optionally, the inspection item determination module is used for: The items to be examined for the first associated disease, the second associated disease and the third associated disease are determined respectively, and the determined items to be examined are merged and deduplicated to obtain multiple first items to be examined. For each first item to be inspected, the target correlation degree corresponding to the first item to be inspected is determined from the first correlation degree, the second correlation degree and the third correlation degree, and the target score corresponding to the first item to be inspected is determined based on the target correlation degree. Based on the relationship between the target score and the preset score threshold corresponding to each first item to be inspected, multiple second items to be inspected are selected from the multiple first items to be inspected.

[0067] Optionally, the physical examination package generation module is used for: The target user's age and gender are extracted from the static feature data of the target user. Based on the age and gender, a first basic inspection item is determined from each of the second inspection items through a preset age-gender baseline model, and first recommended data corresponding to the first basic inspection item is generated. Extract the target user's region from the target user's static feature data; based on the region, determine the second basic examination items from each of the second examination items through a preset regional epidemic disease database; and generate the second recommended data corresponding to the second basic examination items. A basic physical examination package is generated based on the first basic examination items, the first recommended data, the second basic examination items, and the second recommended data. The first basic examination item and the second basic examination item are removed from each of the second examination items to obtain the examination items after removal. Based on the first target associated disease of the examination items after removal, the first target indicator corresponding to the first target associated disease is determined. Based on the first target indicator, third recommendation data is generated. Based on the examination items after removal and the third recommendation data, an additional physical examination package is generated. The first target indicator includes at least one of the single-level abnormal indicator, the composite abnormal indicator and the hidden risk indicator. Based on the basic health check package and the supplementary health check package, the target health check package for the target user is generated.

[0068] Optionally, the device further includes a scoring calculation module; the scoring calculation module is used for: After obtaining multiple second items to be checked, for each second item to be checked, it is determined whether the second item to be checked is a mandatory item, and based on the determination result, the basic score corresponding to the second item to be checked is calculated; and, based on the second target-related disease corresponding to the second item to be checked, the second target-related disease is determined, and based on the second target indicator, the gain score corresponding to the second item to be checked is calculated, wherein the second target indicator includes at least one of the single-level abnormality indicator, the composite abnormality indicator, and the hidden risk indicator; and, based on the historical check items of the target user within a preset time period, it is determined whether the second item to be checked is a redundant item, and based on the determination result, the suppression score corresponding to the second item to be checked is calculated. Calculate the item score corresponding to the second item to be inspected based on the base score, the gain score, and the suppression score; Accordingly, the device further includes a rating addition module; the rating addition module is used for: After generating the target health check package for the target user, the project score corresponding to each second item to be checked is added to the target health check package.

[0069] Optionally, the composite abnormality index, the second associated disease, and the second correlation degree are determined based on the following method: Through the two-level risk reasoning network in the preset risk reasoning model, structured data is extracted from the static feature data and the dynamic attribute data, and the indicator name corresponding to each structured data is determined. For each structured data indicator name, a pre-set label database is used to determine whether the label corresponding to each indicator name is a composite label, and the structured data with composite labels is retained. For each preset composite indicator, based on the multiple indicator names included in the preset composite indicator and the indicator names corresponding to the retained structured data, the corresponding target structured data is matched from the retained structured data, and it is determined whether the target structured data meets the preset requirements corresponding to each indicator name in the preset composite indicator. When the target structured data meets the preset requirements corresponding to each indicator name in the preset composite indicator, the composite abnormal indicator is generated based on the target structured data and the indicator names corresponding to the target structured data. The associated disease corresponding to the preset composite indicator is taken as the second associated disease corresponding to the composite abnormal indicator, and the correlation degree corresponding to the associated disease is taken as the second correlation degree corresponding to the second associated disease.

[0070] Optionally, the hidden risk indicator, the third associated disease, and the third association degree are determined based on the following method: Through the three-level risk reasoning network in the preset risk reasoning model, low-frequency data features and high-risk biomarker features are extracted from the static feature data and the dynamic attribute data. The low-frequency data features and the high-risk biomarker features are then combined to obtain multiple feature combinations. For each feature combination, a preset knowledge base is invoked, and it is retrieved whether the feature combination is contained in the preset knowledge base. If the feature combination is contained in the preset knowledge base, the associated disease corresponding to the feature combination and the degree of association with the associated disease are obtained from the preset knowledge base. The feature combination is used as a hidden risk indicator, and the associated disease corresponding to the feature combination and the degree of association with the associated disease are used as the third associated disease and the third degree of association corresponding to the third associated disease, respectively.

[0071] Optionally, the device further includes a feedback module; the feedback module is used for: After generating the target health check package corresponding to the target user, the test results of each second test item in the target health check package are obtained. If the results of the project inspection indicate that the target user has a corresponding disease risk, then the target indicator corresponding to the second project to be inspected is determined, and the target correlation degree of the target-related disease corresponding to the target indicator is increased.

[0072] It should be noted that other corresponding descriptions of the functional units involved in the dynamic reasoning-based health checkup package generation device provided in this application embodiment can be found in the following references. Figures 1 to 3 The corresponding descriptions in the method will not be repeated here.

[0073] This application also provides a computer device, which may specifically be a personal computer, a server, a network device, etc. Figure 5 As shown, the computer device includes a bus, a processor, memory, and a communication interface, and may also include an input / output interface and a display device. The processor provides computing and control capabilities. The memory includes a non-volatile storage medium and internal memory. The non-volatile storage medium stores an operating system, computer programs, and a database. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage medium. The database stores location information. The network interface allows communication with external terminals via a network connection. When the computer program is executed by the processor, it implements the steps in the various method embodiments.

[0074] Those skilled in the art will understand that Figure 5The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0075] In one embodiment, a computer-readable storage medium is provided, which may be non-volatile or volatile, having stored thereon a computer program that, when executed by a processor, implements the steps in the above method embodiments.

[0076] In one embodiment, a computer program product is provided, including a computer program that, when executed by a processor, implements the steps in the above method embodiments.

[0077] It should be noted that any AI models, software tools, or components not belonging to this company appearing in the embodiments of this application are merely illustrative examples and do not represent actual use. All user personal information involved in the embodiments of this application has been authorized (with the knowledge and consent) by the relevant parties or has been fully authorized by all parties, and the executing entity may obtain it through various legal and compliant means. The collection, storage, use, processing, transmission, provision, and disclosure of the information, data, and signals involved all comply with relevant laws and regulations and do not violate public order and good morals.

[0078] Those skilled in the art will understand that all or part of the processes in the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium. When executed, the computer program can include the processes of the embodiments described above. Any references to memory, databases, or other media used in the embodiments provided in this application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM). The databases involved in the embodiments provided in this application may include at least one type of relational database and non-relational database. Non-relational databases may include, but are not limited to, blockchain-based distributed databases. The processors involved in the embodiments provided in this application may be general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic devices, quantum computing-based data processing logic devices, etc., and are not limited to these.

[0079] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0080] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of this patent application. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this application should be determined by the appended claims.

Claims

1. A method for generating health checkup packages based on dynamic reasoning, characterized in that, include: Acquire static feature data and dynamic health data of the target user; Based on the static feature data and the dynamic health data, the single-level abnormality indicators, composite abnormality indicators and hidden risk indicators of the target user are determined by the three-level risk inference network in the preset risk inference model. Determine the first degree of correlation between the single-level abnormal indicator and the first associated disease corresponding to the single-level abnormal indicator, determine the second degree of correlation between the composite abnormal indicator and the second associated disease corresponding to the composite abnormal indicator, and determine the third degree of correlation between the hidden risk indicator and the third associated disease corresponding to the hidden risk indicator; Based on the first associated disease, the second associated disease, and the third associated disease, multiple first items to be examined are determined, and based on the first correlation degree, the second correlation degree, and the third correlation degree, the multiple first items to be examined are screened to obtain multiple second items to be examined. Based on the multiple second items to be checked, a target health check package corresponding to the target user is generated.

2. The method according to claim 1, characterized in that, The process involves determining multiple first examination items based on the first associated disease, the second associated disease, and the third associated disease, and then filtering these multiple first examination items based on the first correlation, the second correlation, and the third correlation to obtain multiple second examination items, including: The items to be examined for the first associated disease, the second associated disease and the third associated disease are determined respectively, and the determined items to be examined are merged and deduplicated to obtain multiple first items to be examined. For each first item to be inspected, the target correlation degree corresponding to the first item to be inspected is determined from the first correlation degree, the second correlation degree and the third correlation degree, and the target score corresponding to the first item to be inspected is determined based on the target correlation degree. Based on the relationship between the target score and the preset score threshold corresponding to each first item to be inspected, multiple second items to be inspected are selected from the multiple first items to be inspected.

3. The method according to claim 1, characterized in that, The step of generating a target health check package for the target user based on the plurality of second items to be checked includes: The target user's age and gender are extracted from the static feature data of the target user. Based on the age and gender, a first basic inspection item is determined from each of the second inspection items through a preset age-gender baseline model, and first recommended data corresponding to the first basic inspection item is generated. Extract the target user's region from the target user's static feature data; based on the region, determine the second basic examination items from each of the second examination items through a preset regional epidemic disease database; and generate the second recommended data corresponding to the second basic examination items. A basic physical examination package is generated based on the first basic examination items, the first recommended data, the second basic examination items, and the second recommended data. The first basic examination item and the second basic examination item are removed from each of the second examination items to obtain the examination items after removal. Based on the first target associated disease of the examination items after removal, the first target indicator corresponding to the first target associated disease is determined. Based on the first target indicator, third recommendation data is generated. Based on the examination items after removal and the third recommendation data, an additional physical examination package is generated. The first target indicator includes at least one of the single-level abnormal indicator, the composite abnormal indicator and the hidden risk indicator. Based on the basic health check package and the supplementary health check package, the target health check package for the target user is generated.

4. The method according to claim 3, characterized in that, After obtaining multiple second items to be inspected, the method further includes: For each second item to be examined, determine whether the second item to be examined is a mandatory item, and calculate the basic score corresponding to the second item to be examined based on the determination result; and, based on the second target-related disease corresponding to the second item to be examined, determine the second target indicator corresponding to the second target-related disease, and calculate the gain score corresponding to the second item to be examined based on the second target indicator, wherein the second target indicator includes at least one of the single-level abnormality indicator, the composite abnormality indicator, and the hidden risk indicator; and, based on the historical examination items of the target user within a preset time period, determine whether the second item to be examined is a redundant item, and calculate the suppression score corresponding to the second item to be examined based on the determination result. Calculate the item score corresponding to the second item to be inspected based on the base score, the gain score, and the suppression score; Accordingly, after generating the target health check package for the target user, the method further includes: Add the score corresponding to each second item to be examined to the target health check package.

5. The method according to claim 1, characterized in that, The composite abnormality index, the second associated disease, and the second association degree are determined based on the following method: Through the two-level risk reasoning network in the preset risk reasoning model, structured data is extracted from the static feature data and the dynamic attribute data, and the indicator name corresponding to each structured data is determined. For each structured data indicator name, a pre-set label database is used to determine whether the label corresponding to each indicator name is a composite label, and the structured data with composite labels is retained. For each preset composite indicator, based on the multiple indicator names included in the preset composite indicator and the indicator names corresponding to the retained structured data, the corresponding target structured data is matched from the retained structured data, and it is determined whether the target structured data meets the preset requirements corresponding to each indicator name in the preset composite indicator. When the target structured data meets the preset requirements corresponding to each indicator name in the preset composite indicator, the composite abnormal indicator is generated based on the target structured data and the indicator names corresponding to the target structured data. The associated disease corresponding to the preset composite indicator is taken as the second associated disease corresponding to the composite abnormal indicator, and the correlation degree corresponding to the associated disease is taken as the second correlation degree corresponding to the second associated disease.

6. The method according to claim 1, characterized in that, The hidden risk indicator, the third associated disease, and the third association degree are determined based on the following method: Through the three-level risk reasoning network in the preset risk reasoning model, low-frequency data features and high-risk biomarker features are extracted from the static feature data and the dynamic attribute data. The low-frequency data features and the high-risk biomarker features are then combined to obtain multiple feature combinations. For each feature combination, a preset knowledge base is invoked, and it is retrieved whether the feature combination is contained in the preset knowledge base. If the feature combination is contained in the preset knowledge base, the associated disease corresponding to the feature combination and the degree of association with the associated disease are obtained from the preset knowledge base. The feature combination is used as a hidden risk indicator, and the associated disease corresponding to the feature combination and the degree of association with the associated disease are used as the third associated disease and the third degree of association corresponding to the third associated disease, respectively.

7. The method according to claim 1, characterized in that, After generating the target health check package corresponding to the target user, the method further includes: Obtain the examination results for each second item to be examined in the target health check package; If the results of the project inspection indicate that the target user has a corresponding disease risk, then the target indicator corresponding to the second project to be inspected is determined, and the target correlation degree of the target-related disease corresponding to the target indicator is increased.

8. A device for generating physical examination packages based on dynamic reasoning, characterized in that, include: The data acquisition module is used to acquire static feature data and dynamic health data of the target user; The indicator determination module is used to determine the single-level abnormal indicator, composite abnormal indicator and hidden risk indicator of the target user based on the static feature data and the dynamic health data, through the three-level risk inference network in the preset risk inference model. The correlation determination module is used to determine the first correlation degree between the single-level abnormal indicator and the first associated disease corresponding to the single-level abnormal indicator, to determine the second correlation degree between the composite abnormal indicator and the second associated disease corresponding to the composite abnormal indicator, and to determine the third correlation degree between the hidden risk indicator and the third associated disease corresponding to the hidden risk indicator. The examination item determination module is used to determine multiple first examination items based on the first associated disease, the second associated disease, and the third associated disease, and to filter the multiple first examination items based on the first correlation degree, the second correlation degree, and the third correlation degree to obtain multiple second examination items. The health checkup package generation module is used to generate a target health checkup package corresponding to the target user based on the multiple second items to be checked.

9. A storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the method of any one of claims 1 to 7.

10. A computer device, comprising a storage medium, a processor, and a computer program stored on the storage medium and executable on the processor, characterized in that, When the processor executes the computer program, it implements the method of any one of claims 1 to 7.

Images