A comprehensive data analysis and diagnosis assistance system for urticaria

Abstract

The present invention discloses a comprehensive data analysis and diagnosis assistance system for urticaria, which belongs to the technical field of skin diseases. A comprehensive data analysis and diagnosis assistance system for urticaria includes a data acquisition module, a data prediction module, an image diagnosis module and a result interaction module. The present invention solves the problem of reduced accuracy of existing single diagnostic methods. The present invention realizes accurate diagnosis of urticaria by collecting comprehensive data of patients and performing pre-processing, classification, feature extraction, model construction and optimization on the data. At the same time, combined with the identification and comparison of lesion areas, it can better provide auxiliary diagnosis services for doctors, and the prediction results and judgment results are displayed to doctors and patients and sent to the patient's mobile terminal at the same time, so that doctors and patients can understand changes in the condition and provide a scientific basis for disease prevention and treatment strategies.

Description

Technical Field

[0001] The present invention relates to the technical field of skin diseases, and in particular to an urticaria comprehensive data analysis and diagnosis auxiliary system. Background Art

[0002] Urticaria, also known as wheals or skin edema, is a skin disease caused by mast cell activation, which leads to dilation and increased permeability of small blood vessels in the skin and mucous membranes. Currently, most widely used methods for diagnosing skin diseases rely on extracting gross image features of the skin's appearance and learning the relationship between these features and the disease to provide a diagnosis. However, due to the complex causes of urticaria, extracting only surface pathological features provides only a single piece of information about the skin disease, reducing the accuracy of urticaria-assisted diagnosis. Therefore, this method does not meet existing needs. Therefore, we propose a comprehensive data analysis and diagnosis assistance system for urticaria. Summary of the Invention

[0003] The purpose of the present invention is to provide a comprehensive data analysis and diagnosis assistance system for urticaria. By collecting comprehensive data of patients and performing preprocessing, classification, feature extraction, model construction and optimization on the data, accurate diagnosis of urticaria is achieved. At the same time, combined with the identification and comparison of lesion areas, it may better provide auxiliary diagnosis services for doctors. The prediction results and judgment results are displayed to doctors and patients and sent to the patient's mobile terminal at the same time, so that doctors and patients can understand changes in the disease, thereby solving the problems raised in the above-mentioned background technology.

[0004] To achieve the above objectives, the present invention provides the following technical solutions: a comprehensive data analysis and diagnosis support system for urticaria, comprising:

[0005] Data acquisition module, used to:

[0006] Collect comprehensive data from urticaria patients through smart sensors, including medical history data, routine test data, and skin imaging data;

[0007] Pre-process, classify and archive the collected comprehensive data;

[0008] Data prediction module, used to:

[0009] Extract features from the preprocessed comprehensive data and fuse the extracted features;

[0010] Based on the integrated features, a prediction model based on machine learning is constructed, and the constructed prediction model is trained and optimized;

[0011] Deploy the trained prediction model into practical applications to provide auxiliary diagnosis services for doctors and patients;

[0012] Imaging diagnostic module for:

[0013] Build a lesion recognition model based on deep learning, and use skin image data from the comprehensive data to train and optimize the constructed lesion recognition model;

[0014] The trained model is deployed in real-world applications to identify the lesion areas in two periods of skin imaging data: the patient's last treatment period and the current treatment period.

[0015] Comparing the areas of the two lesion areas to determine whether the patient's urticaria has improved, wherein the two lesion areas are the lesion area during the last treatment period and the lesion area during the current treatment period;

[0016] Result interaction module, used to:

[0017] Present the prediction results of the data prediction module and the judgment results of the imaging diagnosis module to doctors and patients;

[0018] At the same time, the prediction results and judgment results are sent to the patient's mobile terminal through various means.

[0019] Furthermore, the data acquisition module includes:

[0020] Data collection module for:

[0021] Comprehensive data from urticaria patients is collected through smart sensors, including medical records, routine test data, and skin imaging data. Routine test data includes blood tests, allergen tests, autoantibody tests, and thyroid function tests.

[0022] Data processing module for:

[0023] Pre-process the collected comprehensive data, including data cleaning and data unification;

[0024] Through data cleaning, remove incomplete, inaccurate, duplicate, damaged or non-standard data in the comprehensive data, involving removing null values ​​and outliers, filling missing values ​​and removing duplicates;

[0025] Through data unification, comprehensive data from different sources and formats are converted into a unified format;

[0026] Among them, image enhancement is performed on the skin image data in the comprehensive data, including contrast enhancement, smoothing and edge detection;

[0027] Data classification module for:

[0028] The preprocessed comprehensive data are classified and numbered according to data type.

[0029] Furthermore, the contrast enhancement is performed on the skin image data in the comprehensive data, and the following steps are performed:

[0030] Extracting skin image data;

[0031] Performing grayscale processing on the skin image data to obtain a grayscale image corresponding to the skin image data;

[0032] Extracting the grayscale value of each pixel contained in the grayscale image;

[0033] Comparing the grayscale value of each pixel with a preset grayscale threshold;

[0034] Extracting pixels whose grayscale values ​​are lower than the grayscale threshold as target pixels;

[0035] Scanning the target pixel points, deleting target pixel points that are not connected to other target pixel points, and obtaining filtered target pixel points;

[0036] Acquire multiple target areas formed by the target pixel points connected according to the connection conditions between the filtered target pixel points;

[0037] Obtaining the grayscale coefficient value corresponding to each target area according to the grayscale value of the pixel points contained in each target area;

[0038] The grayscale coefficient value corresponding to each target area is obtained by the following formula:

[0039]

[0040] Among them, B represents the grayscale coefficient value corresponding to each target area; n represents the number of pixels contained in each target area; K i Indicates the grayscale value corresponding to the i-th pixel; K z Indicates the grayscale median value corresponding to each target area; K pi K represents the grayscale average of the pixels in the target area connected to the i-th pixel; mp Indicates the grayscale average value of n pixels corresponding to each target area; K fp K represents the grayscale average of the pixels on the boundary of the non-target area adjacent to the edge of the target area; b Indicates the standard deviation of the grayscale values ​​corresponding to n pixels;

[0041] The gamma value is used to adjust the contrast of the skin image data.

[0042] Furthermore, contrast adjustment of the skin image data is performed using the gamma coefficient value, including:

[0043] Extract the grayscale coefficient value corresponding to each target area;

[0044] Integrating the gamma coefficient values ​​corresponding to the target area to generate a comprehensive gamma coefficient;

[0045] The comprehensive gamma coefficient is obtained by the following formula:

[0046]

[0047] Among them, B z represents the comprehensive grayscale coefficient; m represents the number of target areas; B i Indicates the grayscale coefficient value corresponding to the i-th target area; B c Indicates the standard deviation of the grayscale coefficient corresponding to the m target areas; K zb represents the standard deviation of the grayscale median values ​​corresponding to the m target areas; P i represents the ratio between the area of ​​the i-th target region and the image area of ​​the skin image data;

[0048] Retrieve the initial contrast value corresponding to the skin image data;

[0049] Using the comprehensive gamma coefficient to enhance and adjust the initial contrast value of the skin image data to obtain skin image data after contrast adjustment;

[0050] The contrast value corresponding to the skin image data after contrast adjustment is obtained by the following formula:

[0051]

[0052] Among them, Q represents the contrast value corresponding to the skin image data after contrast adjustment; Q0 represents the initial contrast value; B z Indicates the comprehensive gamma.

[0053] Furthermore, the data classification module performs the following steps:

[0054] Classify the comprehensive data according to the data type into medical records, testing, and imaging categories;

[0055] Create corresponding folders according to the classified data types, and put medical record data, routine test data, and skin imaging data into corresponding folders respectively;

[0056] After the classification is completed, the created folders are numbered, including the file name, patient name, and ID number.

[0057] Furthermore, the data prediction module includes:

[0058] Feature extraction module, used to:

[0059] Extract features from the preprocessed comprehensive data and fuse the extracted features;

[0060] Divide the fused features into training set and test set with a ratio of 7:3 or 8:2;

[0061] Model building modules for:

[0062] Use machine learning to build prediction models based on the comprehensive data and fused features;

[0063] Training optimization module, used to:

[0064] Use the training set to train the constructed prediction model. During the training process, observe the performance indicators of the prediction model, including accuracy, recall rate, and F1 score, and continuously adjust the performance indicators until the prediction model reaches the optimal performance;

[0065] After completing the training of the prediction model, the performance of the prediction model is evaluated using the test set. If the prediction model performance is poor, the prediction model is optimized, including adding or deleting features and adjusting parameters.

[0066] Model deployment module, used to:

[0067] The trained prediction model is deployed in practical applications, and new comprehensive data of urticaria patients are input into the prediction model for prediction to obtain the prediction results of urticaria patients, and auxiliary diagnosis services are provided to doctors and patients based on the prediction results.

[0068] Furthermore, the feature extraction module is specifically:

[0069] Clustering technology or statistical methods are used to extract features from the pre-processed comprehensive data, where:

[0070] Characteristics in medical record data, including medical history, onset time, duration of illness, seasonal changes, and accompanying symptoms;

[0071] Features in routine test data, including laboratory test results for complete blood count, allergens, autoantibodies, and thyroid function;

[0072] Features in skin imaging data, including images of skin lesions such as wheals and angioedema;

[0073] After the feature extraction is completed, the extracted features are fused using the simple averaging method and the linear combination method.

[0074] Furthermore, the imaging diagnosis module includes:

[0075] Model creation module for:

[0076] Build a lesion recognition model based on deep learning, and use skin image data from the comprehensive data to train and optimize the constructed lesion recognition model;

[0077] Model application module, used for:

[0078] Deploy the trained model to real-world applications, identify the skin image data from the patient's last treatment period and the current treatment period, and identify the lesion areas in both periods;

[0079] Region division module, used for:

[0080] The threshold segmentation method is used to divide the lesion area of ​​the last treatment period and the lesion area of ​​the current treatment period respectively. After the division is completed, the areas of the two lesion areas are compared to determine whether the patient's urticaria has improved.

[0081] Furthermore, the model application module performs the following steps:

[0082] Collect skin image data of the same part of the patient in two periods, which are divided into first image data and second image data, wherein the first image data is the skin image data during the last diagnosis and treatment, and the second image data is the skin image data during the current diagnosis and treatment;

[0083] Using the trained lesion recognition model to respectively recognize the first image data and the second image data;

[0084] The lesion area of ​​the first image data and the lesion area of ​​the second image data are determined through recognition by the lesion recognition model.

[0085] Furthermore, the result interaction module includes:

[0086] Data display module, used for:

[0087] The prediction results of the data prediction module and the judgment results of the imaging diagnosis module are presented to doctors and patients, and export and download functions are provided;

[0088] The doctor and the patient understand the development of the disease through the displayed content, and the doctor formulates a subsequent treatment plan based on the displayed content;

[0089] Result transmission module, used to:

[0090] The prediction and judgment results are sent to the patient's mobile terminal via SMS and email.

[0091] Compared with the prior art, the present invention has the following beneficial effects:

[0092] The data acquisition module of the present invention collects comprehensive data of urticaria patients through intelligent sensors and performs preprocessing to improve the quality of comprehensive data. The data prediction module extracts and fuses features of the preprocessed comprehensive data, and then constructs a prediction model based on machine learning and performs training and optimization, so as to provide auxiliary diagnosis services for doctors and patients through the trained prediction model. The image diagnosis module constructs a lesion recognition model based on deep learning, and uses the lesion recognition model to identify the patient's skin image data in two periods, and compares the identified lesion areas in the two periods to determine whether the patient's urticaria has improved. By combining the predictive analysis of comprehensive data with image recognition, accurate auxiliary diagnosis of urticaria is achieved, avoiding the diagnosis based on single pathological information alone, and effectively improving the accuracy of auxiliary diagnosis of urticaria. BRIEF DESCRIPTION OF THE DRAWINGS

[0093] Figure 1 It is a structural diagram of the urticaria comprehensive data analysis and diagnosis auxiliary system of the present invention. DETAILED DESCRIPTION

[0094] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0095] In order to solve the technical problem that the existing technology can only obtain skin disease pathology information through surface pathology feature extraction, thereby reducing the accuracy of auxiliary diagnosis of urticaria, please refer to Figure 1 , this embodiment provides the following technical solutions:

[0096] An urticaria comprehensive data analysis and diagnosis auxiliary system, comprising:

[0097] Data acquisition module, used to:

[0098] Collect comprehensive data from urticaria patients through smart sensors, including medical history data, routine test data, and skin imaging data;

[0099] Pre-process, classify and archive the collected comprehensive data;

[0100] Data prediction module, used to:

[0101] Extract features from the preprocessed comprehensive data and fuse the extracted features;

[0102] Based on the integrated features, a prediction model based on machine learning is constructed, and the constructed prediction model is trained and optimized;

[0103] Deploy the trained prediction model into practical applications to provide auxiliary diagnosis services for doctors and patients;

[0104] Imaging diagnostic module for:

[0105] Build a lesion recognition model based on deep learning, and use skin image data from the comprehensive data to train and optimize the constructed lesion recognition model;

[0106] The trained model is deployed in real-world applications to identify the lesion areas in two periods of skin imaging data: the patient's last treatment period and the current treatment period.

[0107] Comparing the areas of the two lesion areas to determine whether the patient's urticaria has improved, wherein the two lesion areas are the lesion area during the last treatment period and the lesion area during the current treatment period;

[0108] Result interaction module, used to:

[0109] Present the prediction results of the data prediction module and the judgment results of the imaging diagnosis module to doctors and patients;

[0110] At the same time, the prediction results and judgment results are sent to the patient's mobile terminal through various means.

[0111] The technical effects of the above content are as follows: the data acquisition module collects comprehensive data of urticaria patients through intelligent sensors, such as medical records, routine test data and skin imaging data, and pre-processes and classifies and archives them for subsequent analysis. The data prediction module extracts and fuses features of the pre-processed comprehensive data, and builds a prediction model based on machine learning based on these features. The prediction model is trained and optimized to provide auxiliary diagnosis services for doctors and patients. The imaging diagnosis module builds a lesion recognition model based on deep learning, and uses skin imaging data for training and optimization to identify the lesion area. By identifying and comparing the lesion area between the patient's last treatment period and the current treatment period, it can be judged whether the patient's condition has improved. The result interaction module will display the results of data prediction and imaging diagnosis to doctors and patients, and send the results to the patient's mobile terminal in various ways. By combining the predictive analysis of comprehensive data with image recognition, doctors can analyze the patient's condition more comprehensively to achieve accurate auxiliary diagnosis of urticaria.

[0112] Data acquisition module, including:

[0113] Data collection module for:

[0114] Comprehensive data from urticaria patients is collected through smart sensors, including medical records, routine test data, and skin imaging data. Routine test data includes blood tests, allergen tests, autoantibody tests, and thyroid function tests.

[0115] Data processing module for:

[0116] Pre-process the collected comprehensive data, including data cleaning and data unification;

[0117] Through data cleaning, remove incomplete, inaccurate, duplicate, damaged or non-standard data in the comprehensive data, involving removing null values ​​and outliers, filling missing values ​​and removing duplicates;

[0118] Through data unification, comprehensive data from different sources and formats are converted into a unified format;

[0119] Among them, image enhancement is performed on the skin image data in the comprehensive data, including contrast enhancement, smoothing and edge detection;

[0120] Data classification module for:

[0121] The preprocessed comprehensive data are classified and numbered according to data type.

[0122] The technical effect of the above content is: the data collection module uses intelligent sensors to collect comprehensive data of urticaria patients, including medical history information, routine test data (such as blood routine, allergen detection, autoantibody detection, thyroid function test, etc.) and skin imaging data, so as to provide more comprehensive patient medical history and physiological status information, which is conducive to subsequent model construction and data analysis. The data processing module pre-processes the collected comprehensive data, including data cleaning and data unification, which can remove unnecessary information and simplify the data format, making the comprehensive data easier to understand and use, and improving the efficiency of data processing. The data classification module can classify and number the pre-processed data according to the type of data, which helps to better manage and analyze the data, and also provides convenience for subsequent analysis and prediction.

[0123] Specifically, the contrast enhancement is performed on the skin image data in the comprehensive data by performing the following steps:

[0124] Extracting skin image data;

[0125] Performing grayscale processing on the skin image data to obtain a grayscale image corresponding to the skin image data;

[0126] Extracting the grayscale value of each pixel contained in the grayscale image;

[0127] Comparing the grayscale value of each pixel with a preset grayscale threshold;

[0128] Extracting pixels whose grayscale values ​​are lower than the grayscale threshold as target pixels;

[0129] Scanning the target pixel points, deleting target pixel points that are not connected to other target pixel points, and obtaining filtered target pixel points;

[0130] Acquire multiple target areas formed by the target pixel points connected according to the connection conditions between the filtered target pixel points;

[0131] Obtaining the grayscale coefficient value corresponding to each target area according to the grayscale value of the pixel points contained in each target area;

[0132] The grayscale coefficient value corresponding to each target area is obtained by the following formula:

[0133]

[0134] Among them, B represents the grayscale coefficient value corresponding to each target area; n represents the number of pixels contained in each target area; K i Indicates the grayscale value corresponding to the i-th pixel; K z Indicates the grayscale median value corresponding to each target area; K pi K represents the grayscale average of the pixels in the target area connected to the i-th pixel; mp Indicates the grayscale average value of n pixels corresponding to each target area; K fp K represents the grayscale average of the pixels on the boundary of the non-target area adjacent to the edge of the target area; b Indicates the standard deviation of the grayscale values ​​corresponding to n pixels;

[0135] The gamma value is used to adjust the contrast of the skin image data.

[0136] The technical effect of the above-mentioned technical solution is that by grayscale processing skin image data and extracting, analyzing, and adjusting grayscale values, specifically enhancing pixels with low grayscale values ​​(i.e., darker areas), the visibility of dark details in the image can be effectively improved. This is particularly important for identifying and analyzing key information such as skin lesions and vascular structures. By screening areas connected to target pixels and calculating the gamma coefficient value of each target area, the technical solution helps to more accurately identify and distinguish different characteristic areas in the image. This assists in subsequent decision-making for skin disease diagnosis and treatment planning. Using the calculated gamma coefficient value to adjust the contrast of the skin image data further improves the overall visual quality of the image, making the boundaries between areas of different grayscale levels clearer, thereby enhancing image readability and analysis efficiency. During the process of extracting target pixels and target areas, by removing isolated pixels that are not connected to other target pixels, the impact of noise on image quality can be reduced to a certain extent, improving image purity and analysis accuracy. By calculating the gamma coefficient value of each target area and adjusting the contrast of the skin image data accordingly, this technical solution achieves adaptive contrast enhancement. This means that for different types of skin image data, the scheme can provide appropriate contrast enhancement effects to a certain extent.

[0137] In summary, this technical solution effectively improves the contrast of skin image data through a series of sophisticated processing steps, enhances the visibility and readability of image details, and provides strong support for the accurate diagnosis and analysis of skin diseases.

[0138] Specifically, using the gamma value to adjust the contrast of the skin image data includes:

[0139] Extract the grayscale coefficient value corresponding to each target area;

[0140] Integrating the gamma coefficient values ​​corresponding to the target area to generate a comprehensive gamma coefficient;

[0141] The comprehensive gamma coefficient is obtained by the following formula:

[0142]

[0143] Among them, B z represents the comprehensive grayscale coefficient; m represents the number of target areas; B i Indicates the grayscale coefficient value corresponding to the i-th target area; B c Indicates the standard deviation of the grayscale coefficient corresponding to the m target areas; K zb represents the standard deviation of the grayscale median values ​​corresponding to the m target areas; P irepresents the ratio between the area of ​​the i-th target region and the image area of ​​the skin image data;

[0144] Retrieve the initial contrast value corresponding to the skin image data;

[0145] Using the comprehensive gamma coefficient to enhance and adjust the initial contrast value of the skin image data to obtain skin image data after contrast adjustment;

[0146] The contrast value corresponding to the skin image data after contrast adjustment is obtained by the following formula:

[0147]

[0148] Among them, Q represents the contrast value corresponding to the skin image data after contrast adjustment; Q0 represents the initial contrast value; B z Indicates the comprehensive gamma.

[0149] The technical effect of the above-mentioned technical solution is that by integrating the gamma coefficient values ​​of each target area, a comprehensive gamma coefficient is generated. This coefficient not only considers the grayscale characteristics of the individual target areas but also comprehensively evaluates the grayscale distribution characteristics of all target areas by introducing parameters such as the gamma coefficient standard deviation and the standard deviation of the grayscale median value. This helps to more comprehensively reflect the overall grayscale characteristics of the skin image data. The comprehensive gamma coefficient is used to enhance and adjust the initial contrast value of the skin image data, achieving adaptive contrast adjustment. This adjustment method considers the actual grayscale distribution of the image data and avoids the over-enhancement or under-enhancement problems that can occur with traditional contrast adjustment methods. Through intelligent contrast adjustment, the boundaries between areas of different grayscale levels in the skin image data are clearer, and details are more prominent. This helps improve image readability and analysis efficiency, providing strong support for the accurate diagnosis of skin diseases. Through automated and intelligent processing, this technical solution reduces the need for manual intervention and improves the efficiency and accuracy of the diagnostic process. Doctors can focus more on image data analysis and diagnosis, without spending excessive time on image preprocessing. Because this technical solution can significantly improve the contrast and readability of skin image data, doctors can see the detailed features of skin lesions more clearly during the diagnosis process, thereby enhancing diagnostic confidence and accuracy.

[0150] In summary, this technical solution significantly improves the quality and readability of skin image data by comprehensively evaluating the grayscale characteristics of the target area and achieving adaptive contrast adjustment, providing strong support for the accurate diagnosis of skin diseases. It also optimizes the diagnostic process, improving diagnostic efficiency and accuracy.

[0151] The data classification module performs the following steps:

[0152] Classify the comprehensive data according to the data type into medical records, testing, and imaging categories;

[0153] Create corresponding folders according to the classified data types, and put medical record data, routine test data, and skin imaging data into corresponding folders respectively;

[0154] After the classification is completed, the created folders are numbered, including the file name, patient name, and ID number.

[0155] The technical effect of the above content is: by classifying the comprehensive data and creating corresponding folders according to the data type, and then storing the collected data in their respective folders according to the data type to which it belongs, medical records, testing and imaging related data can be stored separately, making search and management simpler and more efficient.

[0156] Data prediction module, including:

[0157] Feature extraction module, used to:

[0158] Extract features from the preprocessed comprehensive data and fuse the extracted features;

[0159] Divide the fused features into training set and test set with a ratio of 7:3 or 8:2;

[0160] Model building modules for:

[0161] Use machine learning to build prediction models based on the comprehensive data and fused features;

[0162] Training optimization module, used to:

[0163] Use the training set to train the constructed prediction model. During the training process, observe the performance indicators of the prediction model, including accuracy, recall rate, and F1 score, and continuously adjust the performance indicators until the prediction model reaches the optimal performance;

[0164] After completing the training of the prediction model, the performance of the prediction model is evaluated using the test set. If the prediction model performance is poor, the prediction model is optimized, including adding or deleting features and adjusting parameters.

[0165] Model deployment module, used to:

[0166] The trained prediction model is deployed in practical applications, and new comprehensive data of urticaria patients are input into the prediction model for prediction to obtain the prediction results of urticaria patients, and auxiliary diagnosis services are provided to doctors and patients based on the prediction results.

[0167] The technical effect of the above content is as follows: the feature extraction module extracts features from the preprocessed comprehensive data and fuses the extracted features. Feature extraction and fusion are used to reduce the dimension of the data, making it easier for the machine learning algorithm to converge and find the optimal solution more quickly. The model construction module uses machine learning methods such as neural networks, decision trees, random forests, etc. to construct a prediction model, which can enable the machine learning model to fit the relationship between the input features and the target variable as much as possible, thereby realizing the prediction of the disease. The training optimization module uses the training set to train the constructed prediction model. During the training process, the performance indicators of the prediction model, such as accuracy, recall rate and F1 score, are observed, and the performance indicators are continuously adjusted until the prediction model reaches the optimal performance. After the training of the prediction model is completed, the performance of the prediction model is evaluated using the test set. If the prediction model performance is found to be poor, the prediction model needs to be optimized, such as adding or deleting features, or adjusting parameters, etc. Through continuous adjustment, the performance of the prediction model is optimized. Finally, the model deployment module deploys the trained prediction model to actual applications, so that it can truly provide services for doctors and patients.

[0168] Feature extraction module, specifically:

[0169] Clustering technology or statistical methods are used to extract features from the pre-processed comprehensive data, where:

[0170] Characteristics in medical record data, including medical history, onset time, duration of illness, seasonal changes, and accompanying symptoms;

[0171] Features in routine test data, including laboratory test results for complete blood count, allergens, autoantibodies, and thyroid function;

[0172] Features in skin imaging data, including images of skin lesions such as wheals and angioedema;

[0173] After the feature extraction is completed, the extracted features are fused using the simple averaging method and the linear combination method.

[0174] The technical effect of the above content is: clustering technology and statistical methods are used to extract features from the preprocessed comprehensive data. The extracted features are very important for understanding the patient's disease status. After the feature extraction is completed, the extracted features are fused using simple averaging method and linear combination method to obtain a richer feature set, which helps to better train and optimize the prediction model, thereby improving the prediction accuracy of urticaria.

[0175] Imaging diagnostic module, including:

[0176] Model creation module for:

[0177] Build a lesion recognition model based on deep learning, and use skin image data from the comprehensive data to train and optimize the constructed lesion recognition model;

[0178] Model application module, used for:

[0179] Deploy the trained model to actual applications and identify the skin image data of the patient's last treatment period and the current treatment period to identify the lesion areas in the two periods. Specifically, perform the following steps:

[0180] Collect skin image data of the same part of the patient in two periods, which are divided into first image data and second image data, wherein the first image data is the skin image data during the last diagnosis and treatment, and the second image data is the skin image data during the current diagnosis and treatment;

[0181] Using the trained lesion recognition model to respectively recognize the first image data and the second image data;

[0182] Determining the lesion area of ​​the first image data and the lesion area of ​​the second image data through recognition by the lesion recognition model;

[0183] Region division module, used for:

[0184] The threshold segmentation method is used to divide the lesion area of ​​the last treatment period and the lesion area of ​​the current treatment period respectively. After the division is completed, the areas of the two lesion areas are compared to determine whether the patient's urticaria has improved.

[0185] The technical effects of the above content are as follows: the model creation module uses deep learning technology to build a lesion recognition model, and uses the model to train and optimize the skin image data in the comprehensive data. The model application module deploys the trained model to the actual application environment, and identifies the skin image data of the patient's last treatment period and the current treatment period. By comparing the lesion areas of the patient's two skin image data, the changes in the lesion area are determined. In the area division module, the threshold segmentation method is used to divide the two lesion areas. By comparing the areas of the two lesion areas, it is determined whether the patient's urticaria has improved, which can provide strong support for the patient's treatment.

[0186] Result interaction module, including:

[0187] Data display module, used for:

[0188] The prediction results of the data prediction module and the judgment results of the imaging diagnosis module are displayed to doctors and patients, and the displayed content can be exported and downloaded;

[0189] The doctor and the patient understand the development of the disease through the displayed content, and the doctor formulates a subsequent treatment plan based on the displayed content;

[0190] Result transmission module, used to:

[0191] The prediction and judgment results are sent to the patient's mobile terminal via SMS and email.

[0192] The technical effect of the above content is: the data display module is responsible for presenting the prediction results and judgment results generated by the data prediction module and the imaging diagnosis module to doctors and patients intuitively and clearly. In addition, the module also has data export and download functions, so that doctors and patients can view and save relevant data at any time. When doctors and patients need to understand the development of the disease, they can view the real-time updated prediction results and judgment results through the data display module. The displayed results can provide important reference for doctors to formulate subsequent treatment plans, helping doctors to grasp the changes in the patient's condition more accurately, and the result transmission module is responsible for sending the prediction results and judgment results to the patient's mobile terminal in the form of text messages and emails, which not only makes it convenient for patients to check their disease progress at any time, but also allows patients to better communicate with doctors to ensure the effective implementation of the treatment plan.

[0193] Working principle: The data acquisition module collects comprehensive data of urticaria patients through intelligent sensors, such as medical records, routine test data and skin imaging data, and pre-processes and classifies and archives them for subsequent analysis. The data prediction module builds a prediction model based on machine learning based on these features, which can provide auxiliary diagnosis services for doctors and patients. The imaging diagnosis module builds a lesion recognition model based on deep learning, which can identify the lesion area. By identifying and comparing the lesion area of ​​the patient in two periods, it can be determined whether the patient's condition has improved, which can provide strong support for the patient's treatment. By combining the predictive analysis of comprehensive data with image recognition, the doctor can analyze the patient's condition more comprehensively to achieve accurate auxiliary diagnosis of urticaria. The result interaction module will display the results of data prediction and imaging diagnosis to doctors and patients, and send the results to the patient's mobile terminal in various ways, so that doctors and patients can understand the development of the disease, and also provide important reference for doctors to formulate subsequent treatment plans.

[0194] It should be noted that, in this document, relational terms such as first and second, etc., are used only to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the terms "comprises," "comprising," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that includes a list of elements includes not only those elements but also other elements not explicitly listed, or elements inherent to such process, method, article, or apparatus.

[0195] While the embodiments of the present invention have been shown and described, it will be apparent to those skilled in the art that various changes, modifications, substitutions, and alterations can be made to the embodiments without departing from the principles and spirit of the invention.

Claims

1. A comprehensive data analysis and diagnosis support system for urticaria, characterized in that: include: Data acquisition module, used to: Collect comprehensive data from urticaria patients through smart sensors, including medical history data, routine test data, and skin imaging data; Pre-process, classify and archive the collected comprehensive data; Data prediction module, used to: Extract features from the preprocessed comprehensive data and fuse the extracted features; Based on the integrated features, a prediction model based on machine learning is constructed, and the constructed prediction model is trained and optimized; Deploy the trained prediction model into practical applications to provide auxiliary diagnosis services for doctors and patients; Imaging diagnostic module for: Build a lesion recognition model based on deep learning, and use skin image data from the comprehensive data to train and optimize the constructed lesion recognition model; The trained model is deployed in real-world applications to identify the lesion areas in two periods of skin imaging data: the patient's last treatment period and the current treatment period. Comparing the areas of the two lesion areas to determine whether the patient's urticaria has improved, wherein the two lesion areas are the lesion area during the last treatment period and the lesion area during the current treatment period; Result interaction module, used to: Present the prediction results of the data prediction module and the judgment results of the imaging diagnosis module to doctors and patients; At the same time, the prediction and judgment results are sent to the patient's mobile terminal through various methods; The data acquisition module includes: Data collection module for: Comprehensive data from urticaria patients is collected through smart sensors, including medical records, routine test data, and skin imaging data. Routine test data includes blood tests, allergen tests, autoantibody tests, and thyroid function tests. Data processing module for: Pre-process the collected comprehensive data, including data cleaning and data unification; Through data cleaning, remove incomplete, inaccurate, duplicate, damaged or non-standard data in the comprehensive data, involving removing null values ​​and outliers, filling missing values ​​and removing duplicates; Through data unification, comprehensive data from different sources and formats are converted into a unified format; Among them, image enhancement is performed on the skin image data in the comprehensive data, including contrast enhancement, smoothing and edge detection; Data classification module for: Classify and number the pre-processed comprehensive data according to data type; The contrast enhancement is performed on the skin image data in the comprehensive data, and the following steps are performed: Extracting skin image data; Performing grayscale processing on the skin image data to obtain a grayscale image corresponding to the skin image data; Extracting the grayscale value of each pixel contained in the grayscale image; Comparing the grayscale value of each pixel with a preset grayscale threshold; Extracting pixels whose grayscale values ​​are lower than the grayscale threshold as target pixels; Scanning the target pixel points, deleting target pixel points that are not connected to other target pixel points, and obtaining filtered target pixel points; Acquire multiple target areas formed by the target pixel points connected according to the connection conditions between the filtered target pixel points; Obtaining the grayscale coefficient value corresponding to each target area according to the grayscale value of the pixel points contained in each target area; The grayscale coefficient value corresponding to each target area is obtained by the following formula: Among them, B represents the grayscale coefficient value corresponding to each target area; n represents the number of pixels contained in each target area; K i Indicates the grayscale value corresponding to the i-th pixel; K z Indicates the grayscale median value corresponding to each target area; K pi K represents the grayscale average of the pixels in the target area connected to the i-th pixel; mp Indicates the grayscale average value of n pixels corresponding to each target area; K fp K represents the grayscale average of the pixels on the boundary of the non-target area adjacent to the edge of the target area; b Indicates the standard deviation of the grayscale values ​​corresponding to n pixels; performing contrast adjustment on skin image data using the gamma value; The step of adjusting the contrast of the skin image data by using the gamma coefficient value includes: Extract the grayscale coefficient value corresponding to each target area; Integrating the gamma coefficient values ​​corresponding to the target area to generate a comprehensive gamma coefficient; The comprehensive gamma coefficient is obtained by the following formula: Among them, B z represents the comprehensive grayscale coefficient; m represents the number of target areas; B i Indicates the grayscale coefficient value corresponding to the i-th target area; B c Indicates the standard deviation of the grayscale coefficient corresponding to the m target areas; K zb represents the standard deviation of the grayscale median values ​​corresponding to the m target areas; P i represents the ratio between the area of ​​the i-th target region and the image area of ​​the skin image data; Retrieving the initial contrast value corresponding to the skin image data; Using the comprehensive gamma coefficient to enhance and adjust the initial contrast value of the skin image data to obtain skin image data after contrast adjustment; The contrast value corresponding to the skin image data after contrast adjustment is obtained by the following formula: Among them, Q represents the contrast value corresponding to the skin image data after contrast adjustment; Q0 represents the initial contrast value; B z Indicates the comprehensive gamma.

2. The urticaria comprehensive data analysis and diagnosis auxiliary system according to claim 1, characterized in that: The data classification module performs the following steps: Classify the comprehensive data according to the data type into medical records, testing, and imaging categories; Create corresponding folders according to the classified data types, and put medical record data, routine test data, and skin imaging data into corresponding folders respectively; After the classification is completed, the created folders are numbered, including the file name, patient name and ID number.

3. The urticaria comprehensive data analysis and diagnosis auxiliary system according to claim 1, characterized in that: The data prediction module includes: Feature extraction module, used to: Extract features from the preprocessed comprehensive data and fuse the extracted features; Divide the fused features into training set and test set with a ratio of 7:3 or 8:2; Model building modules for: Use machine learning to build prediction models based on the comprehensive data and fused features; Training optimization module, used to: Use the training set to train the constructed prediction model. During the training process, observe the performance indicators of the prediction model, including accuracy, recall rate, and F1 score, and continuously adjust the performance indicators until the prediction model reaches the optimal performance; After completing the training of the prediction model, the performance of the prediction model is evaluated using the test set. If the prediction model performance is poor, the prediction model is optimized, including adding or deleting features and adjusting parameters. Model deployment module, used to: The trained prediction model is deployed in practical applications, and new comprehensive data of urticaria patients are input into the prediction model for prediction to obtain the prediction results of urticaria patients, and auxiliary diagnosis services are provided to doctors and patients based on the prediction results.

4. The urticaria comprehensive data analysis and diagnosis auxiliary system according to claim 3, characterized in that: The feature extraction module is specifically: Clustering technology or statistical methods are used to extract features from the pre-processed comprehensive data, where: Characteristics in medical record data, including medical history, onset time, duration of illness, seasonal changes, and accompanying symptoms; Features in routine test data, including laboratory test results for complete blood count, allergens, autoantibodies, and thyroid function; Features in skin imaging data, including images of skin lesions such as wheals and angioedema; After the feature extraction is completed, the extracted features are fused using the simple averaging method and the linear combination method.

5. The urticaria comprehensive data analysis and diagnosis auxiliary system according to claim 1, characterized in that: The imaging diagnosis module includes: Model creation module for: Build a lesion recognition model based on deep learning, and use skin image data from the comprehensive data to train and optimize the constructed lesion recognition model; Model application module, used for: Deploy the trained model to real-world applications, identify the skin image data from the patient's last treatment period and the current treatment period, and identify the lesion areas in both periods; Region division module, used for: The threshold segmentation method is used to divide the lesion area of ​​the last treatment period and the lesion area of ​​the current treatment period respectively. After the division is completed, the areas of the two lesion areas are compared to determine whether the patient's urticaria has improved.

6. The urticaria comprehensive data analysis and diagnosis auxiliary system according to claim 5, characterized in that: The model application module performs the following steps: Collect skin image data of the same part of the patient in two periods, which are divided into first image data and second image data, wherein the first image data is the skin image data during the last diagnosis and treatment, and the second image data is the skin image data during the current diagnosis and treatment; Using the trained lesion recognition model to respectively recognize the first image data and the second image data; The lesion area of ​​the first image data and the lesion area of ​​the second image data are determined through recognition by the lesion recognition model.

7. The urticaria comprehensive data analysis and diagnosis auxiliary system according to claim 1, characterized in that: The result interaction module includes: Data display module, used for: The prediction results of the data prediction module and the judgment results of the imaging diagnosis module are presented to doctors and patients, and export and download functions are provided; The doctor and the patient understand the development of the disease through the displayed content, and the doctor formulates a subsequent treatment plan based on the displayed content; Result transmission module, used to: The prediction and judgment results are sent to the patient's mobile terminal via SMS and email.

Images