Skin testing data optimization integration method and platform combined with multiple validation analysis

Abstract

The present invention discloses a method and platform for optimizing and integrating skin detection data combined with multiple verification and analysis, and relates to the technical field related to detection data processing. The method includes: obtaining a first spectral feature significantly correlated with a first blood component based on a correlation analysis plan; machine training a skin database established according to a first correspondence between the first blood component and the first spectral feature to obtain a skin detection model; obtaining predicted blood component information of the skin to be detected; verifying the predetermined blood component information of the skin to be detected to obtain a verification result; and integrating and archiving the predetermined blood component information. The method solves the technical problem that it is difficult to accurately capture the correspondence between skin spectral features and blood components in the body, which leads to insufficient accuracy and credibility of skin detection results and low efficiency of data integration and archiving, and achieves the technical effect of improving the accuracy and reliability of skin detection and the efficiency of detection data integration.

Description

Technical Field

[0001] The present application relates to the technical field related to detection data processing, and specifically to a method and platform for optimizing and integrating skin detection data combined with multiple verification analyses. Background Art

[0002] In today's medical cosmetology and health monitoring field, skin testing is an important means of evaluating individual health status and beauty needs. Its detection accuracy and data integration efficiency are increasingly valued. The health status of the skin not only reflects the external characteristics of the human body, but also indirectly reflects the internal health status. The body's health indicators are obtained through the information on the skin surface. When inflammatory indicators appear in the human blood, it may be reflected on the skin, such as redness, swelling, itching and pain. Conversely, abnormal conditions of human skin can be detected in the blood. Inflammatory factors, for example, blood routine tests will show an increase in the percentage of white blood cells and neutrophils, an increase in C-reactive protein, and an increase in erythrocyte sedimentation rate. These are all manifestations of inflammation caused by bacterial infection. Detecting whether the human body has inflammation through blood routine tests, C-reactive protein, erythrocyte sedimentation rate and other blood component-related indicators is an indispensable skin detection method. Traditional skin detection methods often rely on a single detection method, such as visual observation, chemical analysis or simple optical measurement, which is not only time-consuming, but also time-consuming. It is laborious and can only provide some physical information on the surface of the skin. It is easily affected by subjective factors and is difficult to fully and accurately reflect the physiological state of the deep skin. It is often affected by many factors, such as skin type, ambient light, individual differences, etc., which can easily lead to uncertainty and deviation in the test results. The single-dimensional test results cannot fully consider the complex relationship between the skin and the blood components in the body. For example, insufficient number of red blood cells and hemoglobin or decreased function may lead to pale and dull skin; abnormal platelets or coagulation dysfunction may cause bruises and bleeding spots on the skin; high blood sugar stimulates the peripheral nerves of the skin, causing itching, which in turn affects skin health; high blood lipids lead to poor blood circulation, which in turn leads to skin wrinkles; at the same time, it also does not consider the effects of gene expression (such as hereditary skin diseases, skin cancer, etc.) and microecology (such as acne, atopic dermatitis, eczema, etc.) on skin condition, making it difficult to comprehensively evaluate the skin condition and ensure that the skin test data is accurate and reliable enough.

[0003] Therefore, in the current technologies related to skin detection data optimization and integration, there is a technical problem that it is difficult to accurately capture the correspondence between skin spectral characteristics and blood components in the body, which in turn leads to insufficient accuracy and credibility of skin detection results and low efficiency of data integration and archiving. Summary of the invention

[0004] This application solves the technical problem that it is difficult to accurately capture the correspondence between skin spectral characteristics and blood components in the body, which leads to insufficient accuracy and credibility of skin detection results and low efficiency of data integration and archiving, by providing a method and platform for optimizing and integrating skin detection data combined with multiple verification analyses. This achieves the technical effect of improving the accuracy, reliability and integration efficiency of skin detection data.

[0005] The present application provides a method for optimizing and integrating skin detection data combined with multiple verification analyses, the method comprising: acquiring a first spectral feature that is significantly correlated with a first blood component based on a correlation analysis plan; performing machine training on a skin database established according to a first correspondence between the first blood component and the first spectral feature to obtain a skin detection model; performing predictive analysis on the skin to be detected through the skin detection model to obtain predicted blood component information of the skin to be detected; verifying predetermined blood component information of the skin to be detected through the predicted blood component information to obtain a verification result; and integrating and archiving the predetermined blood component information when the verification result meets the verification constraints.

[0006] In a possible implementation, the method for optimizing and integrating skin detection data in combination with multiple verification analyses also performs the following processing: establishing a first blood sample, wherein the first blood sample contains a first blood component identified with a first concentration; establishing a second blood sample, wherein the second blood sample contains the first blood component identified with a second concentration; sequentially constructing a first skin model of the first blood sample and a second skin model of the second blood sample, and performing diffuse reflection simulation detection on the first skin model and the second skin model respectively to obtain a diffuse reflection spectrum set; performing multidimensional feature comparison analysis on the diffuse reflection spectrum set according to a spectral analysis plan to obtain first spectral feature comparison information; and analyzing the first spectral feature comparison information to obtain the first spectral feature that is significantly correlated with the first blood component.

[0007] In a possible implementation, the skin detection data optimization and integration method combined with multiple verification analysis also performs the following processing: randomly obtaining a first spectrum in the diffuse reflectance spectrum set, the first blood component corresponding to the first spectrum being the first concentration; obtaining a second spectrum in the diffuse reflectance spectrum set according to the spectral analysis plan, the first blood component corresponding to the second spectrum being the second concentration, wherein the concentration difference between the second concentration and the first concentration is greater than a concentration threshold value; reading predetermined spectral features, and sequentially obtaining first spectral feature information of the first spectrum and second spectral feature information of the second spectrum based on the predetermined spectral features; comparing the first spectral feature information with the second spectral feature information to obtain first spectral feature comparison information.

[0008] In a possible implementation, the skin detection data optimization integration method combined with multiple verification analysis also performs the following processing: the predetermined spectral characteristics include wavelength range, absorption peak, reflectivity, spectral shape, spectral width, spectral contrast and spectral slope, wherein the absorption peak includes hemoglobin absorption peak, water absorption peak, collagen absorption peak and elastin absorption peak.

[0009] In a possible implementation, the skin detection data optimization and integration method combined with multiple verification analysis also performs the following processing: obtaining the diffuse reflectance spectrum of the skin to be tested, and collecting spectral feature information of the diffuse reflectance spectrum in combination with the predetermined spectral feature; matching the first feature information corresponding to the first spectral feature in the spectral feature information; inputting the first feature information into the first prediction channel in the skin detection model to obtain the first predicted concentration of the first blood component; and composing the predicted blood component information based on the one-to-one correspondence between the first blood component and the first predicted concentration.

[0010] In a possible implementation, the skin detection data optimization and integration method combined with multiple verification analysis also performs the following processing: traversing and matching the first spectral feature in the first spectral feature information to obtain a first data group; adding the first concentration to the first data group, and machine training the first data group to obtain the first prediction channel, wherein the first prediction channel is used to perform intelligent concentration prediction on the first blood component.

[0011] In a possible implementation, the skin detection data optimization and integration method combined with multiple verification analyses also performs the following processing: matching the most similar blood component information of the predetermined blood component information in the archived database; obtaining the most similar skin image corresponding to the most similar blood component information; obtaining the skin image of the skin to be tested; determining whether the image feature value deviation between the skin image and the most similar skin image is within the deviation threshold value; if so, integrating and archiving the predetermined blood component information.

[0012] The present application also provides a skin detection data optimization and integration platform combined with multiple verification and analysis, including: a first spectral feature acquisition module, used to acquire a first spectral feature that is significantly correlated with a first blood component based on a correlation analysis plan; a skin detection model training module, used to perform machine training on a skin database established according to a first correspondence between the first blood component and the first spectral feature, to obtain a skin detection model; a predicted blood component information acquisition module, used to perform predictive analysis on the skin to be tested through the skin detection model, to obtain predicted blood component information of the skin to be tested; a verification result acquisition module, used to verify the predetermined blood component information of the skin to be tested through the predicted blood component information, to obtain a verification result; an integration and archiving module, used to integrate and archive the predetermined blood component information when the verification result meets the verification constraints.

[0013] It is proposed that the method and platform for optimizing and integrating skin detection data combined with multiple verification and analysis proposed in this application will obtain a first spectral feature significantly correlated with a first blood component based on a correlation analysis plan; machine train the skin database established according to the first correspondence between the first blood component and the first spectral feature to obtain a skin detection model; predict and analyze the skin to be detected through the skin detection model to obtain predicted blood component information of the skin to be detected; verify the predetermined blood component information of the skin to be detected through the predicted blood component information to obtain a verification result; and integrate and archive the predetermined blood component information when the verification result meets the verification constraint. This solves the technical problem that it is difficult to accurately capture the correspondence between skin spectral features and blood components in the body, which leads to insufficient accuracy and credibility of skin detection results and low efficiency of data integration and archiving, and achieves the technical effect of improving the accuracy and reliability of skin detection and the efficiency of detection data integration. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] In order to more clearly illustrate the technical solution of the embodiment of the present disclosure, the accompanying drawings of the embodiment of the present disclosure will be briefly introduced below. A flow chart is used in the present application to illustrate the operations performed by the system according to the embodiment of the present application. It should be understood that the preceding or following operations are not necessarily performed accurately in order. On the contrary, various steps can be processed in reverse order or simultaneously as needed. At the same time, other operations can also be added to these processes, or one or more operations can be removed from these processes.

[0015] Figure 1 A schematic diagram of a process flow of a method for optimizing and integrating skin detection data combined with multiple validation analyses provided in an embodiment of the present application;

[0016] Figure 2Schematic diagram of the structure of the skin detection data optimization and integration platform combined with multiple verification analysis provided in an embodiment of the present application.

[0017] Explanation of the reference numerals: first spectral feature acquisition module 10 , skin detection model training module 20 , predicted blood component information acquisition module 30 , verification result acquisition module 40 , integration and archiving module 50 . DETAILED DESCRIPTION

[0018] The above description is only an overview of the technical solution of the present application. In order to more clearly understand the technical means of the present application, it can be implemented in accordance with the contents of the specification. In order to make the above and other purposes, features and advantages of the present application more obvious and easy to understand, the specific implementation methods of the present application are listed below.

[0019] In order to make the purpose, technical solutions and advantages of the present application clearer, the present application will be further described in detail below in conjunction with the accompanying drawings. The described embodiments should not be regarded as limiting the present application. All other embodiments obtained by ordinary technicians in the field without making creative work are within the scope of protection of this application.

[0020] In the following description, reference is made to "some embodiments", which describe a subset of all possible embodiments, but it is understood that "some embodiments" may be the same subset or different subsets of all possible embodiments, and may be combined with each other without conflict, and the terms "first\second" involved are merely to distinguish similar objects and do not represent a specific ordering of objects. The terms "including" and "having" and any variations are intended to cover non-exclusive inclusions, for example, a process, method, system, product, or server that includes a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or modules that are not clearly listed or inherent to these processes, methods, products, or devices. Unless otherwise defined, all technical and scientific terms used herein have the same meaning as those generally understood by technicians in the technical field of this application. The terms used herein are for the purpose of describing the embodiments of the present application only.

[0021] The present application embodiment provides a method for optimizing and integrating skin detection data combined with multiple validation analyses, such as Figure 1 As shown, the method includes:

[0022] Step S100, obtaining a first spectral feature significantly correlated with a first blood component based on a correlation analysis plan.

[0023] Preferably, the correlation analysis plan refers to a pre-designed plan for correlation analysis of specific blood components, which includes how to process data, select specific algorithms for correlation analysis (such as Pearson correlation coefficient, Spearman correlation coefficient, etc.), and how to screen out significantly correlated spectral features. Through the correlation analysis method, those spectral features that are closely related to the first blood component are screened out from the skin spectral data. Specifically, skin spectral data and corresponding blood component data are obtained from multiple samples. Skin spectral data are usually obtained through specific spectral technology (such as near-infrared spectroscopy technology NIRS). These spectral data contain characteristics such as light reflection and absorption at different wavelengths. Blood component data are obtained through traditional blood testing methods, such as routine blood tests, C-reactive protein, erythrocyte sedimentation rate and other indicators. Among them, C-reactive protein is a non-specific acute phase reactant synthesized by the liver and released into the blood, usually in infection. , inflammation, tissue damage, surgery, malignant tumors, etc., and as an indicator of these pathological conditions, it is mainly used to activate the complement system in the body, help clear necrotic or damaged cells, and participate in the body's immune defense response; the collected blood component data and skin spectral data are correlated to determine which spectral features (such as reflection intensity, absorbance, etc.) have a significant correlation with the target blood component under spectra of different wavelengths. For example, when testing blood sugar, it may be found that the reflectance or absorbance of light at certain specific wavelengths shows a significant linear or nonlinear correlation with the blood sugar level. These spectral features are the spectral features that are significantly correlated with the first blood component. Then, based on the results of the correlation analysis, those spectral features that have a high correlation with the target blood component (the first blood component) are screened out, called the first spectral features, which are used to predict and analyze blood components. In other words, changes in spectral features can reflect changes in blood components.

[0024] In a possible implementation, step S100 further includes step S110, establishing a first blood sample, wherein the first blood sample contains a first blood component identified with a first concentration; step S120, establishing a second blood sample, wherein the second blood sample contains the first blood component identified with a second concentration; step S130, sequentially constructing a first skin model of the first blood sample and a second skin model of the second blood sample, and performing diffuse reflection simulation detection on the first skin model and the second skin model respectively to obtain a diffuse reflection spectrum set; step S140, performing a multi-dimensional feature comparison analysis on the diffuse reflection spectrum set according to a spectral analysis plan to obtain first spectral feature comparison information; step S150, analyzing the first spectral feature comparison information to obtain the first spectral feature that is significantly correlated with the first blood component.

[0025] Preferably, a first and a second blood sample are established, each sample contains a specific blood component (such as C-reactive protein, etc.) and its concentration, the first blood sample has a first blood component of a first concentration, and the second blood sample has the same blood component of a second concentration, and then based on the first and second blood samples, a first skin model of the first blood sample and a second skin model of the second blood sample are respectively constructed, and these skin models simulate the skin characteristics under different blood component concentrations; then, diffuse reflection simulation detection is performed on the constructed first skin model and the second skin model to obtain a diffuse reflection spectrum set of each model at different wavelengths, and these spectral data reflect the skin characteristics under different blood component concentrations. reaction to light; then, according to the spectral analysis plan, a multi-dimensional feature comparative analysis is performed on the acquired diffuse reflectance spectrum set, that is, by comparing the spectral data corresponding to different blood samples, the respective spectral features are identified to obtain the first spectral feature comparison information, and the first spectral feature comparison information is analyzed by a statistical analysis method (such as correlation coefficient calculation, regression analysis, etc.) to determine which spectral features have a significant correlation with the first blood component, so as to evaluate the relationship between the spectral features and the blood component concentration. Finally, those spectral features that are significantly correlated with the first blood component are determined, which will serve as an important basis for subsequent modeling and analysis, and are used to develop a skin detection model to achieve the prediction of specific blood components.

[0026] In a possible implementation, step S140 further includes step S141, randomly acquiring a first spectrum in the diffuse reflectance spectrum set, wherein the first blood component corresponding to the first spectrum is the first concentration; step S142, acquiring a second spectrum in the diffuse reflectance spectrum set according to the spectral analysis plan, wherein the first blood component corresponding to the second spectrum is the second concentration, wherein the concentration difference between the second concentration and the first concentration is greater than a concentration threshold value; step S143, reading a predetermined spectral feature, and based on the predetermined spectral feature, sequentially acquiring first spectral feature information of the first spectrum and second spectral feature information of the second spectrum; step S144, comparing the first spectral feature information with the second spectral feature information to obtain comparison information of the first spectral feature.

[0027] Further, step S143 also includes that the predetermined spectral characteristics include wavelength range, absorption peak, reflectivity, spectral shape, spectral width, spectral contrast and spectral slope, wherein the absorption peak includes hemoglobin absorption peak, water absorption peak, collagen absorption peak and elastin absorption peak.

[0028] Preferably, a piece of spectral data is randomly selected from the diffuse reflectance spectrum set, referred to as the first spectrum, and the concentration of the first blood component corresponding to this spectrum is the first concentration, which means that it reflects the skin's response to the blood component under specific conditions; according to the spectral analysis plan, another piece of spectral data is obtained from the diffuse reflectance spectrum set, referred to as the second spectrum, and the concentration of the same blood component corresponding to this spectrum is the second concentration, wherein the concentration difference between the second concentration and the first concentration must be greater than a set concentration threshold value to ensure sufficient distinction between the two spectra and the effectiveness of comparative analysis; a set of predetermined spectral features is determined through the spectral features obtained through the previous analysis, including wavelength range, absorption peak, reflectivity, spectral shape, spectral width, spectral contrast and spectral slope, wherein the absorption peaks include hemoglobin absorption peak, water absorption peak, collagen absorption peak and elastin absorption peak, in order to describe The spectral responses of the blood components at different concentrations are analyzed, based on predetermined spectral features, the first spectrum and the second spectrum are analyzed in sequence to extract corresponding first spectral feature information and second spectral feature information, which may include processing of spectral data, such as spectral smoothing, feature extraction and other technologies, to obtain more accurate and reliable feature information; finally, the obtained first spectral feature information is compared and analyzed with the second spectral feature information, and through comparison, it is possible to identify which spectral features have changed significantly when the concentration changes, that is, to evaluate how the spectral features of the blood components change at different concentrations, and through comparative analysis, first spectral feature comparison information is generated, including the differences between the first spectrum and the second spectrum, such as the degree of change of spectral features, significance, etc., which helps to understand the impact of changes in blood component concentrations on skin spectral responses and provide data support for subsequent model training and health monitoring.

[0029] Preferably, the predetermined spectral characteristics refer to several important parameter indicators predetermined to describe and evaluate the spectral characteristics in skin spectral analysis; the wavelength range refers to the wavelength interval measured in the spectral analysis, usually in nanometers (nm), and different wavelengths correspond to different optical properties and absorption properties; the absorption peak is the peak portion of the spectrum that indicates the concentration of a specific substance. In skin testing, the hemoglobin absorption peak is related to the hemoglobin in the blood, and usually shows obvious absorption characteristics at specific wavelengths, which can be used to detect blood components; the moisture absorption peak reflects the moisture content of the skin, and moisture has obvious absorption characteristics in the near-infrared spectral region; the collagen absorption peak, collagen is an important component of the skin, and its absorption characteristics can be used to evaluate the structure and health of the skin; the elastin absorption peak It is related to skin elasticity, and its absorption characteristics also reflect the health and aging of the skin; reflectance refers to the ratio of the intensity of incident light reflected by the skin surface to the incident light intensity at a specific wavelength, which affects the shape and characteristics of the overall spectrum; spectral shape refers to the relative position and change mode of absorption peaks and valleys in the spectrum, reflecting the characteristics of different components; spectral width refers to the width of the absorption peak, usually expressed as full width half maximum (FWHM), and the change in width can indicate the concentration of the component and environmental conditions; spectral contrast refers to the relative intensity difference between the absorption peak and the background spectrum, and a higher contrast helps to identify the absorption peak more clearly; spectral slope refers to the rate of change of the spectral curve at a certain point, reflecting the change of certain characteristics in the spectrum (such as the steepness of the rise or fall of the absorption peak). By analyzing these characteristics, the health of the skin and related blood components can be predicted and evaluated more accurately.

[0030] Step S200, performing machine training on a skin database established according to a first correspondence between the first blood component and the first spectral feature to obtain a skin detection model.

[0031] Preferably, first, the first spectral feature determined by correlation analysis is associated with the corresponding first blood component (such as C-reactive protein, etc.) to construct a skin database. This skin database records the skin spectral data of a large number of samples and their corresponding blood component data. Each record in the database includes the spectral feature of a specific sample (reflection or absorption values ​​of multiple wavelengths) and the corresponding blood component value. If the inflammation is caused by a viral infection, it may lead to an increase in the percentage of lymphocytes, normal white blood cells, or even a slight decrease, and low C-reactive protein. If you want to detect whether inflammation occurs, you can draw blood for testing these indicators; if it is aseptic inflammation caused by rheumatic immune system diseases, it will also lead to an increase in C-reactive protein and erythrocyte sedimentation rate in the blood. For example, acute inflammation, tissue damage, myocardial infarction, surgical trauma, radiation damage, etc. may all lead to a rapid increase in CRP levels; such as rheumatism Chronic inflammatory diseases such as arthritis and systemic lupus erythematosus may also be accompanied by a continuous increase in CRP. If it is a bacterial infection, the CRP level usually increases significantly, and the magnitude of the increase is proportional to the severity of the infection. Then, training data is extracted from the established skin database, including taking the spectral features in the skin database as input variables and the corresponding blood components as target variables. A detection model is constructed based on machine learning algorithms (such as neural networks, decision trees, random forests), and the extracted training data is used to train the detection model to learn the relationship between spectral features and blood components. Finally, a skin detection model that can map new skin spectral data to predicted blood components is obtained. This model can be used to analyze the spectral features of the skin to be inspected in real time, predict its corresponding blood component information, and achieve rapid and non-invasive health monitoring, thereby achieving an assessment of skin health status.

[0032] Step S300: performing predictive analysis on the skin to be inspected by using the skin detection model to obtain predicted blood component information of the skin to be inspected.

[0033] Preferably, a spectrometer or other related equipment is first used to detect the skin to be tested, and spectral data of the skin is collected, usually including reflectance or absorbance information at multiple wavelengths. The collected spectral data of the skin to be tested is input into a trained skin detection model, and then the skin detection model predicts and calculates the blood component prediction values ​​corresponding to the spectral characteristics of the skin based on the input spectral data. The output result of the skin detection model is the predicted blood component information of the skin to be tested, including the concentrations or levels of various blood components, such as blood glucose, cholesterol, hemoglobin, etc., to achieve rapid and non-invasive health assessment, and the predicted blood component information can be used for health monitoring, disease warning or personalized care, etc.

[0034] In a possible implementation, step S300 further includes step S310, obtaining the diffuse reflectance spectrum of the skin to be tested, and collecting spectral feature information of the diffuse reflectance spectrum in combination with the predetermined spectral feature; step S320, matching the first feature information corresponding to the first spectral feature in the spectral feature information; step S330, inputting the first feature information into the first prediction channel in the skin detection model to obtain the first predicted concentration of the first blood component; step S340, forming the predicted blood component information based on the one-to-one correspondence between the first blood component and the first predicted concentration.

[0035] Preferably, a spectrometer or other device is used to detect the skin to be tested, and its diffuse reflectance spectrum is collected, which includes the reflection characteristics of the skin to light at different wavelengths. In the acquired diffuse reflectance spectrum data, combined with the previously defined predetermined spectral features (such as wavelength range, absorption peak, etc.), the corresponding spectral feature information is extracted, including analyzing the spectral data to identify specific feature points, such as the position and intensity of the absorption peak. In the extracted spectral feature information, the feature information corresponding to the first spectral feature is matched to find the spectral feature related to the blood component (such as hemoglobin, blood glucose, etc.) corresponding to the first spectral feature. Then, the matched first feature information is input into the first prediction channel of the skin detection model. The first prediction channel of the skin detection model outputs a first predicted concentration of the first blood component based on the input spectral feature information, representing the concentration level of the component in the skin to be tested. Finally, according to the one-to-one correspondence between the first blood component (such as hemoglobin, blood glucose, etc.) and the first predicted concentration, a complete predicted blood component information is formed, which represents the estimated concentration of the specific blood component in the skin to be tested, and provides a basis for further health monitoring and management.

[0036] In a possible implementation, step S330 further includes step S331, traversing and matching the first spectral feature in the first spectral feature information to obtain a first data group; step S332, adding the first concentration to the first data group, and performing machine training on the first data group to obtain the first prediction channel, wherein the first prediction channel is used to perform intelligent concentration prediction on the first blood component.

[0037] Preferably, in the collected first spectral feature information, all spectral features are traversed one by one to find features that match the first spectral feature, and spectral data related to the target blood component (such as hemoglobin, blood glucose, etc.) are identified through a set similarity threshold. By traversing and matching, all qualified spectral features are collected to form a data set, called a first data group, which includes multiple spectral data corresponding to the first spectral feature, and these data will be used for subsequent model training; the acquired first concentration (the concentration value corresponding to the first blood component) is added to the generated first data group, which means that the first data group contains not only spectral feature information, but also spectral data corresponding to these features. concentration information to form a data set of feature-concentration pairs, and then build a prediction model based on a machine learning model (such as linear regression, support vector machine, neural network, etc.), and use the first data group containing spectral features and corresponding concentrations to train the constructed prediction model to learn the relationship between spectral features and blood component concentrations. Through a large number of training samples, the prediction model can establish a mapping relationship between spectral features and concentrations, and finally generate a first prediction channel. The first prediction channel receives new spectral feature inputs and outputs a corresponding concentration prediction value of the first blood component, which can intelligently predict the blood component concentration in the sample to be tested and provide fast and accurate detection results.

[0038] Step S400, verifying the predetermined blood component information of the skin to be tested by using the predicted blood component information to obtain a verification result.

[0039] Preferably, predetermined blood component information of the skin to be tested is obtained through known medical test results or the patient's health records. The blood component information represents the normal blood component levels (such as blood glucose, cholesterol, hemoglobin, etc.) of the patient under specific circumstances. The predicted blood component information is the predicted information obtained by detecting and analyzing the spectral characteristics of the skin to be tested using a skin detection model. The predicted blood component information is compared with the predetermined blood component information to evaluate the accuracy and reliability of the blood component information prediction. Verification can adopt different comparison methods, such as absolute error calculation, calculating the absolute difference between the predicted value and the predetermined value, relative error calculation, calculating the gap between the predicted result and the predetermined result in percentage form, and statistical test, using statistical methods (such as t-test or variance analysis) to determine whether the difference between the two is significant; based on the results of comparison and analysis, a verification result is generated, including the difference between the predicted value and the predetermined value, the error range, and whether the set verification standard (such as error threshold) is met. If the verification result meets the preset standard, the prediction effect of the model is considered valid and can be used for clinical applications; if not, the model may need to be adjusted or retrained.

[0040] Step S500, when the verification result meets the verification constraint, the predetermined blood component information is integrated and archived.

[0041] Preferably, the obtained verification results are evaluated to confirm whether they meet the preset verification constraints. The verification constraints are usually set according to medical standards or research data, such as error range, accuracy requirements, etc. When the verification results meet the constraints, it means that the predicted blood component information is sufficiently consistent with the predetermined blood component information. This information can be integrated, including combining the predicted results with the predetermined results to form a comprehensive blood component information set. At the same time, relevant metadata may be recorded, such as the test date, sample source, spectral characteristics, predicted values, predetermined values ​​and verification results, etc., to ensure the traceability and integrity of the information; the integrated information will be archived in the platform database to ensure that all blood component data can be effectively managed and accessed, which is convenient for subsequent query and analysis. The archived integrated data will not only help future health management, but also provide valuable data support for medical services, helping doctors analyze patients' health status and changing trends, such as for clinical research, long-term health monitoring, disease warning or the formulation of personalized medical plans.

[0042] In a possible implementation, step S500 further includes step S510, matching the most similar blood component information of the predetermined blood component information in the archive database; step S520, obtaining the most similar skin image corresponding to the most similar blood component information; step S530, obtaining the skin image of the skin to be tested; step S540, determining whether the image feature value deviation between the skin image and the most similar skin image is within the deviation threshold value; step S550, if so, integrating and archiving the predetermined blood component information.

[0043] Preferably, a similarity algorithm or a correlation analysis method is used to determine the degree of similarity between the predetermined blood component information and a record in the database in the archived database, and the most similar blood component information is found, that is, historical data that is most similar to the predetermined blood component information of the skin to be tested is found, wherein the archived database contains a large number of archived historical records, including blood component information of different patients. If the most similar blood component information is matched, then skin images associated with the information are obtained. These skin images are collected at the same time as the blood component records, usually through a specific imaging device (such as a high-resolution camera or a skin scanning device), showing the appearance of the skin in a certain blood component state, such as skin texture, pigment, pores and other features, and then images of the skin to be tested are obtained from the sample to be tested. These images show the current state of the skin and will be used in the next step. In the step, the skin image to be tested is compared with the most similar skin image. Specifically, through image analysis technology, characteristic values ​​of the skin image to be tested and the most similar skin image are extracted. The characteristic values ​​may include skin pigment distribution, texture, reflection characteristics, pore size, etc., and it is determined whether the characteristic value deviation of the two images is within the set deviation threshold value. The deviation threshold value is used to determine the similarity between the two. If the difference in image features is small, it means that the similarity between the two skin images is high. Finally, if the image characteristic value of the skin to be tested deviates from the characteristic value of the most similar skin image within the threshold value, it is considered that the current predetermined blood component information has a high consistency with the historical records in the database. The predetermined blood component information of the skin to be tested is integrated with the matching historical records and archived in the database, ensuring the accuracy and consistency of the data, and providing valuable reference information for future detection.

[0044] In the above, refer to Figure 1 The skin detection data optimization integration method combined with multiple verification analysis according to an embodiment of the present invention is described in detail. Figure 2 A skin detection data optimization integration platform combined with multiple validation analysis according to an embodiment of the present invention is described.

[0045] The skin detection data optimization and integration platform combined with multiple verification analysis according to the embodiment of the present invention is used to solve the technical problem that it is difficult to accurately capture the correspondence between the skin spectral characteristics and the blood components in the body in the existing skin detection data optimization and integration, which leads to insufficient accuracy and credibility of skin detection results and low efficiency of data integration and archiving, and achieves the technical effect of improving the accuracy, reliability and integration efficiency of skin detection data. The skin detection data optimization and integration platform combined with multiple verification analysis includes: a first spectral feature acquisition module 10, a skin detection model training module 20, a predicted blood component information acquisition module 30, a verification result acquisition module 40, and an integration and archiving module 50.

[0046] A first spectral feature acquisition module 10, configured to acquire a first spectral feature significantly correlated with a first blood component based on a correlation analysis plan;

[0047] A skin detection model training module 20, configured to perform machine training on a skin database established according to a first correspondence between the first blood component and the first spectral feature, to obtain a skin detection model;

[0048] The predicted blood component information obtaining module 30 is used to perform predictive analysis on the skin to be inspected by using the skin detection model to obtain the predicted blood component information of the skin to be inspected;

[0049] A verification result obtaining module 40 is used to verify the predetermined blood component information of the skin to be inspected by using the predicted blood component information to obtain a verification result;

[0050] The integration and archiving module 50 is used to integrate and archive the predetermined blood component information when the verification result meets the verification constraint.

[0051] The specific configuration of the first spectral feature acquisition module 10 will be described in detail below. The first spectral feature acquisition module 10 further includes: establishing a first blood sample, the first blood sample contains a first blood component with a first concentration mark; establishing a second blood sample, the second blood sample contains the first blood component with a second concentration mark; sequentially constructing a first skin model of the first blood sample and a second skin model of the second blood sample, and performing diffuse reflection simulation detection on the first skin model and the second skin model respectively to obtain a diffuse reflection spectrum set; performing multi-dimensional feature comparison analysis on the diffuse reflection spectrum set according to the spectral analysis plan to obtain first spectral feature comparison information; analyzing the first spectral feature comparison information to obtain the first spectral feature that is significantly correlated with the first blood component.

[0052] The specific configuration of the first spectral feature acquisition module 10 will be described in detail below. The first spectral feature acquisition module 10 further includes: randomly acquiring a first spectrum in the diffuse reflectance spectrum set, the first blood component corresponding to the first spectrum being the first concentration; acquiring a second spectrum in the diffuse reflectance spectrum set according to the spectral analysis plan, the first blood component corresponding to the second spectrum being the second concentration, wherein the concentration difference between the second concentration and the first concentration is greater than a concentration threshold value; reading a predetermined spectral feature, and sequentially acquiring first spectral feature information of the first spectrum and second spectral feature information of the second spectrum based on the predetermined spectral feature; comparing the first spectral feature information with the second spectral feature information to obtain the first spectral feature comparison information.

[0053] The specific configuration of the first spectral feature acquisition module 10 will be described in detail below. The first spectral feature acquisition module 10 further includes: the predetermined spectral features include wavelength range, absorption peak, reflectivity, spectral shape, spectral width, spectral contrast and spectral slope, wherein the absorption peak includes hemoglobin absorption peak, water absorption peak, collagen absorption peak and elastin absorption peak.

[0054] The specific configuration of the predicted blood component information acquisition module 30 will be described in detail below. The predicted blood component information acquisition module 30 further includes: obtaining the diffuse reflectance spectrum of the skin to be tested, and collecting the spectral feature information of the diffuse reflectance spectrum in combination with the predetermined spectral feature; matching the first feature information corresponding to the first spectral feature in the spectral feature information; inputting the first feature information into the first prediction channel in the skin detection model to obtain the first predicted concentration of the first blood component; and forming the predicted blood component information according to the one-to-one correspondence between the first blood component and the first predicted concentration.

[0055] The specific configuration of the module 30 for obtaining the predicted blood component information will be described in detail below. The module 30 for obtaining the predicted blood component information further includes: traversing and matching the first spectral feature in the first spectral feature information to obtain a first data group; adding the first concentration to the first data group, and performing machine training on the first data group to obtain the first prediction channel, wherein the first prediction channel is used to perform intelligent concentration prediction on the first blood component.

[0056] The specific configuration of the integrated archiving module 50 will be described in detail below. The integrated archiving module 50 may further include: matching the most similar blood component information of the predetermined blood component information in the archiving database; obtaining the most similar skin image corresponding to the most similar blood component information; obtaining the skin image of the skin to be inspected; determining whether the image feature value deviation between the skin image and the most similar skin image is within the deviation threshold value; if so, integrating and archiving the predetermined blood component information.

[0057] The skin detection data optimization and integration platform combined with multiple verification analysis provided in the embodiment of the present invention can execute the skin detection data optimization and integration method combined with multiple verification analysis provided in any embodiment of the present invention, and has the corresponding functional modules and beneficial effects of the execution method.

[0058] Although the present application makes various references to certain modules in the system according to the embodiments of the present application, any number of different modules may be used and run on the user terminal and / or server, and the various units and modules included are only divided according to functional logic, but are not limited to the above division, as long as the corresponding functions can be achieved; in addition, the specific names of the functional units are only for the convenience of distinguishing each other, and are not used to limit the scope of protection of the present invention.

[0059] The above specific implementations do not constitute a limitation on the protection scope of this application. It should be understood by those skilled in the art that various modifications, combinations and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions and improvements made within the spirit and principles of this application should be included in the protection scope of this application.

Claims

1. A method for optimizing and integrating skin detection data combined with multiple validation analyses, characterized in that: include: Step S100: acquiring a first spectral feature significantly correlated with a first blood component based on a correlation analysis plan; Step S2 00: Perform machine training on a skin database established according to a first correspondence between the first blood component and the first spectral feature to obtain a skin detection model; Step S300: performing predictive analysis on the skin to be inspected by using the skin detection model to obtain predicted blood component information of the skin to be inspected; Step S400: verifying the predetermined blood component information of the skin to be tested by using the predicted blood component information to obtain a verification result, wherein the verification method includes absolute error calculation, relative error calculation and statistical test; Step S500: when the verification result meets the verification constraint, integrating and archiving the predetermined blood component information; Wherein, step S100 includes: Step S110: creating a first blood sample, wherein the first blood sample comprises a first blood component having a first concentration mark; Step S120: creating a second blood sample, wherein the second blood sample contains the first blood component with a second concentration marker; Step S130: sequentially constructing a first skin model of the first blood sample and a second skin model of the second blood sample, and performing diffuse reflection simulation detection on the first skin model and the second skin model respectively to obtain a diffuse reflection spectrum set; Step S140: performing multi-dimensional feature comparison analysis on the diffuse reflectance spectrum set according to the spectrum analysis plan to obtain first spectrum feature comparison information; Step S150: analyzing the first spectral feature comparison information to obtain the first spectral feature that is significantly correlated with the first blood component.

2. The method for optimizing and integrating skin detection data combined with multiple validation analysis according to claim 1, characterized in that: Step S140 includes: Step S141: randomly acquiring a first spectrum in the diffuse reflectance spectrum set, wherein the first blood component corresponding to the first spectrum has the first concentration; Step S142: acquiring a second spectrum in the diffuse reflectance spectrum set according to the spectrum analysis plan, wherein the first blood component corresponding to the second spectrum is the second concentration, wherein the concentration difference between the second concentration and the first concentration is greater than a concentration threshold value; Step S143: reading a predetermined spectral feature, and sequentially acquiring first spectral feature information of the first spectrum and second spectral feature information of the second spectrum based on the predetermined spectral feature; Step S144: comparing the first spectral feature information with the second spectral feature information to obtain the first spectral feature comparison information.

3. The method for optimizing and integrating skin detection data combined with multiple validation analysis according to claim 2, characterized in that: The predetermined spectral characteristics include wavelength range, absorption peak, reflectivity, spectral shape, spectral width, spectral contrast and spectral slope, wherein the absorption peak includes hemoglobin absorption peak, water absorption peak, collagen absorption peak and elastin absorption peak.

4. The method for optimizing and integrating skin detection data combined with multiple validation analysis according to claim 2, characterized in that: Step S300 includes: Step S310: acquiring the diffuse reflectance spectrum of the skin to be tested, and collecting spectral feature information of the diffuse reflectance spectrum in combination with the predetermined spectral feature; Step S320: matching first feature information corresponding to the first spectral feature in the spectral feature information; Step S330: inputting the first characteristic information into the first prediction channel in the skin detection model to obtain a first predicted concentration of the first blood component; Step S340: forming the predicted blood component information according to the one-to-one correspondence between the first blood component and the first predicted concentration.

5. The method for optimizing and integrating skin detection data combined with multiple validation analysis according to claim 4, characterized in that: Step S330 includes: Step S331: traverse and match the first spectral feature in the first spectral feature information to obtain a first data group; Step S332: adding the first concentration to the first data group, and performing machine training on the first data group to obtain the first prediction channel, wherein the first prediction channel is used to perform intelligent concentration prediction on the first blood component.

6. The method for optimizing and integrating skin detection data combined with multiple validation analysis according to claim 1, characterized in that: Step S400 further includes: Step S410: matching the most similar blood component information of the predetermined blood component information in the archive database; Step S420: obtaining the most similar skin image corresponding to the most similar blood component information; Step S430: Acquire a skin image of the skin to be inspected; Step S440: determining whether the image feature value deviation between the skin image and the most similar skin image is within a deviation threshold; Step S450: If yes, integrate and archive the predetermined blood component information.

7. A skin detection data optimization and integration platform combined with multiple validation analysis, characterized in that: The platform is used to implement the method for optimizing and integrating skin detection data combined with multiple validation analysis according to any one of claims 1 to 6, and the platform comprises: A first spectral feature acquisition module, used for acquiring a first spectral feature significantly correlated with a first blood component based on a correlation analysis plan; a skin detection model training module, configured to perform machine training on a skin database established according to a first correspondence between the first blood component and the first spectral feature, to obtain a skin detection model; A predicted blood component information obtaining module, used for performing a predictive analysis on the skin to be inspected by using the skin detection model to obtain the predicted blood component information of the skin to be inspected; A verification result obtaining module, used to verify the predetermined blood component information of the skin to be inspected by using the predicted blood component information to obtain a verification result; The integration and archiving module is used to integrate and archive the predetermined blood component information when the verification result meets the verification constraint.

Images