Child respiratory tract infection diagnosis method and system based on multi-modal data fusion

Abstract

The invention discloses a children respiratory tract infection diagnosis method and system based on multi-modal data fusion, and relates to the technical field of respiratory tract infection diagnosis. Distributing the priority of each modal data according to the change trend of the physiological data monitored in real time; generating a comprehensive diagnosis characteristic index of the child health state; through time sequence analysis, the fused feature indexes are monitored in real time, when an abnormal mode is detected, early warning information is automatically generated, and diagnosis result suggestions are sent through a medical staff or parent interface; the comprehensive diagnosis characteristic indexes are optimized according to comparison between a diagnosis result and an actual situation; and storing each diagnosis data and early warning record in a children exclusive health database for long-term trend analysis and future diagnosis reference. According to the invention, the real-time performance and accuracy of children respiratory tract infection diagnosis are improved, the disease monitoring and health management capabilities are enhanced, and a more scientific and comprehensive health protection means is provided for children.

Description

Technical Field

[0001] The present invention relates to the technical field of respiratory tract infection diagnosis, and in particular to a method and system for diagnosing respiratory tract infection in children based on multimodal data fusion. Background Art

[0002] In recent years, with the increasing prominence of children's health issues, respiratory tract infection has become an important factor affecting children's growth and development and quality of life; traditional respiratory tract infection diagnosis methods have limitations in accuracy and efficiency, especially in early detection and timely intervention; with the continuous advancement of medical technology, biosensor technology, data processing technology and artificial intelligence technology are gradually applied to the field of children's health monitoring; among them, multimodal data fusion technology is gradually becoming a research hotspot for the diagnosis of children's respiratory tract infections because it can integrate multiple physiological signals and improve the comprehensive diagnostic capabilities of data.

[0003] Existing diagnostic technologies based on multimodal data fusion still face many challenges in practice. First, since children's physiological data are greatly affected by environmental interference factors, existing data acquisition and preprocessing methods are difficult to completely remove noise, affecting the reliability of diagnostic data. Second, existing multimodal data fusion technologies usually adopt a fixed-weight fusion method, which is difficult to adapt to the dynamic changes of different individuals and real-time health status, resulting in insufficient individualization and adaptability of diagnostic results. In addition, in terms of abnormal pattern recognition, existing technologies usually rely on fixed templates and cannot be adjusted in real time according to the dynamic changes of individual health, which may lead to low sensitivity and specificity of abnormal detection. Therefore, how to optimize the data preprocessing algorithm, improve the adaptability of multimodal data fusion, and improve the accuracy of diagnosis based on a comprehensive analysis method of time series and abnormal patterns has become an urgent problem to be solved at this stage. Summary of the invention

[0004] In view of the problems existing in the above-mentioned background technology, a diagnosis method and system for children's respiratory tract infection based on multimodal data fusion are proposed.

[0005] Therefore, the problem to be solved by the present invention is how to optimize the data preprocessing algorithm, improve the adaptability of multimodal data fusion, and improve the accuracy of diagnosis based on a comprehensive analysis method of time series and abnormal patterns.

[0006] In order to solve the above technical problems, the present invention provides the following technical solutions: In a first aspect, the present invention provides a method for diagnosing respiratory tract infection in children based on multimodal data fusion, which comprises: Step S1: collecting the child's real-time physiological data through a portable monitoring device and performing standardized processing to form a multi-modal raw data set; Step S2: Analyze the change trend of the standardized real-time physiological data based on the time series, and calculate the real-time change rate of the real-time physiological data; calculate the correlation weight of each modality data to the target diagnostic feature, and form a dynamic correlation weight matrix; Extract key features from multimodal data based on the dynamic correlation weight matrix, sort them into a priority list according to the weight values, and retain high-priority features as diagnostic input; Using a weight-based multimodal feature fusion algorithm, high-priority features are weighted and summed to generate a comprehensive diagnostic feature index of the child's health status; Step S3: Through time series analysis, the fused characteristic indicators are monitored in real time to determine whether the child has an abnormal physiological state; Step S4: When an abnormal pattern is detected, a warning message is automatically generated and a diagnosis result suggestion is sent through medical staff; the comprehensive diagnostic characteristic indicators are optimized based on the comparison between the diagnosis result and the actual situation; each diagnosis data and warning record is stored in a child-specific health database.

[0007] As a preferred solution of the method for diagnosing respiratory tract infection in children based on multimodal data fusion of the present invention, the specific implementation process of step S1 includes: After authorization by the user, a portable monitoring device is installed on the user's body surface, and the portable monitoring device is used to detect the user's real-time physiological data, which includes body temperature data, blood oxygen data and respiratory rate data.

[0008] The user's real-time physiological data is standardized and an adaptive noise reduction algorithm is used to filter out environmental interference to ensure the clarity and reliability of the input data.

[0009] As a preferred solution of the method for diagnosing respiratory tract infection in children based on multimodal data fusion of the present invention, the specific implementation process of step S2 includes: Based on the time series, the change trend of the standardized real-time physiological data is analyzed, the real-time change rate of the real-time physiological data is calculated, and the change trend index is generated, as follows: ; in, Indicates j The real-time change rate of real-time physiological data, Indicates j Real-time physiological data t The long-term trend term of time, Indicates j Real-time physiological data t -1 long-term trend term; Calculate the change trend indicator, the calculation formula is as follows: ; in, Indicates j Real-time physiological data change trend indicators, and Respectively represent j The real-time change rate and fluctuation weight of real-time physiological data, Indicates j Real-time physiological data t The short-term fluctuation of time, Represents the standard deviation of short-term volatility; Based on the changing trend indicator and historical health data, calculate the relevance weights of each modality data to the target diagnostic features, and form a dynamic relevance weight matrix, as follows: ; in, Indicates j The relevance weight of the real-time physiological data to the target diagnostic features, Indicates j Real-time physiological data trend indicators and j Real-time physiological data and historical health data The covariance of and Respectively represent j Real-time physiological data trend indicators and j Real-time physiological data and historical health data The standard deviation of .

[0010] Generate dynamic relevance weight matrix ,in, represents the relevance weight of body temperature data to the target diagnostic feature, represents the correlation weight of blood oxygen data to the target diagnostic feature, Represents the relevance weight of respiratory rate data to the target diagnostic feature.

[0011] As a preferred solution of the method for diagnosing respiratory tract infection in children based on multimodal data fusion of the present invention, the specific implementation process of step S3 includes: The comprehensive diagnostic characteristic indicators are segmented into time series, and the indicator changes in continuous time intervals are constructed into dynamic characteristic sequences for subsequent analysis.

[0012] The dynamic feature sequence is matched with the template pattern in the pre-built abnormal pattern library, the matching similarity index is calculated, and the high similarity patterns are screened for key analysis, as follows: ; in, Represents the matching similarity index, Indicates j The weight of the change trend indicator of real-time physiological data, Indicates the exception pattern library k The first mode j A real-time physiological data change trend indicator, max() represents the maximum value function; Set the filtering rules as follows: ; in, Represents the filtered set of high similarity patterns, Represents a template pattern in the exception pattern library, represents a library of pre-built exception patterns, Indicates the preset similarity threshold.

[0013] As a preferred solution of the method for diagnosing respiratory tract infection in children based on multimodal data fusion of the present invention, the specific implementation process of step S3 also includes: According to the matched high similarity pattern, combined with the change amplitude and deviation range of the actual feature sequence, the abnormality index is calculated, and the warning level is generated according to the threshold classification, as follows: ; ; in, represents the abnormality index (the subscript i corresponds to the number of the current analysis object (such as an individual or a moment)), Indicates j The weight of the change trend indicator of real-time physiological data, Represents the high similarity pattern set The mean of all modes in , Indicates the exception pattern library k The first j Real-time physiological data change trend indicators, Represents a collection of highly similar patterns The total number of medium and high similarity patterns.

[0014] As a preferred solution of the method for diagnosing respiratory tract infection in children based on multimodal data fusion of the present invention, the specific implementation process of step S4 includes: To set the warning level rules: ; in, Indicates the warning level. ∈[1,3], L 1 , L 2 and L 3 Represent low risk, medium risk and high risk respectively. and Represent the abnormality threshold limits of low risk and medium risk respectively.

[0015] According to the abnormality index and warning level, multi-dimensional warning information including abnormality type, degree and recommended measures is generated and sent to the medical staff or parent interface, and the information is recorded in the diagnosis database.

[0016] The multi-dimensional warning information is matched and analyzed with the actual diagnostic feedback provided by medical staff, and the consistency index is calculated to evaluate the accuracy of the diagnostic results. The diagnostic data of each time and the optimized comprehensive diagnostic characteristic indicators are stored in the children's exclusive health database, and the health trend characteristics are regularly extracted using time series analysis methods to provide reference for future diagnosis.

[0017] In a second aspect, the present invention provides a children's respiratory tract infection diagnosis system based on multimodal data fusion, which comprises: The multimodal data acquisition module is used to collect children's real-time physiological data through portable monitoring equipment to form a multimodal raw data set; standardize the collected multimodal data, and use an adaptive noise reduction algorithm to filter environmental interference to ensure the clarity and reliability of the input data; The dynamic weight adjustment and feature extraction module is used to apply the dynamic weight adjustment method according to the trend of physiological data changes monitored in real time, assign the priority of each modality data, and optimize the adaptability of data fusion; extract key diagnostic features from the processed multimodal data, and use the weight-based fusion algorithm to generate comprehensive diagnostic feature indicators of children's health status; The abnormal state monitoring and early warning module is used to monitor the fused characteristic indicators in real time through time series analysis to determine whether the child has an abnormal physiological state; when an abnormal pattern is detected, early warning information is automatically generated and the diagnosis result suggestions are sent to medical staff; The personalized diagnosis strategy optimization module is used to optimize the comprehensive diagnostic characteristic indicators based on the comparison of the diagnosis results with the actual situation, so as to form a diagnosis strategy that is more suitable for the individual; and store each diagnosis data and warning record in a child-specific health database.

[0018] In a third aspect, the present invention provides a computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: when the computer program instructions are executed by the processor, the steps of the method for diagnosing respiratory tract infections in children based on multimodal data fusion as described in the first aspect of the present invention are implemented.

[0019] In a fourth aspect, the present invention provides a computer-readable storage medium having a computer program stored thereon, wherein: when the computer program instructions are executed by a processor, the steps of the method for diagnosing respiratory tract infections in children based on multimodal data fusion as described in the first aspect of the present invention are implemented.

[0020] The beneficial effects of the present invention are as follows: the present invention solves the problems of low diagnostic accuracy and insufficient real-time performance of a single data modality in the diagnosis of respiratory tract infections in children through the technical means of multimodal data fusion; utilizes data collection, dynamic weight adjustment, feature extraction, time series analysis and model optimization to gradually achieve high-precision, personalized real-time diagnosis; the setting of each step helps to improve the reliability, real-time performance and accuracy of the diagnosis, and provide comprehensive support for children's health management. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0022] Figure 1 Schematic diagram of the steps of the diagnosis method of children's respiratory tract infection based on multimodal data fusion.

[0023] Figure 2 This is a schematic diagram of the structure of the children's respiratory infection diagnosis system based on multimodal data fusion. DETAILED DESCRIPTION

[0024] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are described in detail below in conjunction with the accompanying drawings.

[0025] In the following description, many specific details are set forth to facilitate a full understanding of the present invention, but the present invention may also be implemented in other ways different from those described herein, and those skilled in the art may make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0026] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The term "in one embodiment" that appears in different places in this specification does not necessarily refer to the same embodiment, nor does it refer to a separate or selective embodiment that is mutually exclusive with other embodiments.

[0027] Example 1: Reference Figure 1~Figure 2 , which is the first embodiment of the present invention, and provides a method for diagnosing respiratory tract infection in children based on multimodal data fusion, comprising: Step S1: The real-time physiological data (body temperature, blood oxygen, respiratory rate) of children are collected through portable monitoring equipment to form a multimodal raw data set; the collected multimodal data are standardized, and an adaptive noise reduction algorithm is used to filter environmental interference to ensure the clarity and reliability of the input data.

[0028] Specifically, after authorization by the user, a portable monitoring device is installed on the user's body surface, and the portable monitoring device is used to detect the user's real-time physiological data, and the physiological data includes body temperature data, blood oxygen data and respiratory rate data.

[0029] Furthermore, the user's real-time physiological data is standardized, and an adaptive noise reduction algorithm is used to filter environmental interference to ensure the clarity and reliability of the input data.

[0030] Step S2: According to the changing trend of physiological data monitored in real time, a dynamic weight adjustment method is applied to assign the priority of each modal data to optimize the adaptability of data fusion; key diagnostic features are extracted from the processed multimodal data, and a weight-based fusion algorithm is used to generate a comprehensive diagnostic feature indicator of the child's health status.

[0031] Specifically, the change trend of the standardized real-time physiological data is analyzed based on the time series, the real-time change rate of the real-time physiological data is calculated, and the change trend index is generated, as follows: ; in, Indicates j The real-time change rate of real-time physiological data, Indicates j Real-time physiological data t The long-term trend term of time, Indicates j Real-time physiological data t -1 long-term trend term; Calculate the change trend indicator, the calculation formula is as follows: ; in, Indicates j Real-time physiological data change trend indicators, and Respectively represent j The real-time change rate and fluctuation weight of real-time physiological data, Indicates j Real-time physiological data t The short-term fluctuation of time, Represents the standard deviation of short-term volatility.

[0032] Based on the changing trend indicator and historical health data, calculate the relevance weights of each modality data to the target diagnostic features, and form a dynamic relevance weight matrix, as follows: ; in, Indicates j The relevance weight of the real-time physiological data to the target diagnostic features, Indicates j Real-time physiological data trend indicators and j Real-time physiological data and historical health data The covariance of and Respectively represent j Real-time physiological data trend indicators and j Real-time physiological data and historical health data The standard deviation of .

[0033] Generate dynamic relevance weight matrix ,in, represents the relevance weight of body temperature data to the target diagnostic feature, represents the correlation weight of blood oxygen data to the target diagnostic feature, Represents the relevance weight of respiratory rate data to the target diagnostic feature.

[0034] Furthermore, according to the dynamic correlation weight matrix Key features are extracted from multimodal data and sorted into a priority list according to weight values, and high-priority features are retained as diagnostic input.

[0035] A weight-based multimodal feature fusion algorithm is used to perform weighted summation on high-priority features to generate a comprehensive diagnostic feature index of children's health status.

[0036] Step S3: Through time series analysis, the fused characteristic indicators are monitored in real time to determine whether the child has an abnormal physiological state; when an abnormal pattern is detected, an early warning message is automatically generated, and the diagnosis result suggestions are sent through the medical staff or parent interface, and high-risk situations are marked.

[0037] Specifically, the comprehensive diagnostic characteristic indicators are segmented into time series, and indicator changes within continuous time intervals are constructed as dynamic characteristic sequences.

[0038] The dynamic feature sequence is matched with the template pattern in the pre-built abnormal pattern library, the matching similarity index is calculated, and the high similarity pattern is screened, as follows: ; in, Represents the matching similarity index, Indicates j The weight of the change trend indicator of real-time physiological data, Indicates the exception pattern library k The first j A real-time physiological data change trend indicator, max() represents the maximum value function; Set the filtering rules as follows: ; in, Represents the filtered set of high similarity patterns, Represents a template pattern in the exception pattern library, represents a library of pre-built exception patterns, Indicates the preset similarity threshold.

[0039] Furthermore, according to the matched high similarity pattern, combined with the change amplitude and deviation range of the actual feature sequence, the abnormality index is calculated, and the warning level is generated according to the threshold classification, as follows: ; ; in, Represents the abnormality index, Indicates j The weight of the change trend indicator of real-time physiological data, Represents the high similarity pattern set The mean of all modes in , Indicates the exception pattern library k The first j Real-time physiological data change trend indicators, Represents a collection of highly similar patterns The total number of medium and high similarity patterns.

[0040] Step S4: Based on the comparison between the diagnosis result and the actual situation, the comprehensive diagnostic characteristic index is optimized to form a more suitable diagnosis strategy for the individual and improve the accuracy of subsequent diagnosis; each diagnosis data and warning record is stored in a child-specific health database for long-term trend analysis and future diagnosis reference.

[0041] Specifically, set the warning level rules: ; in, Indicates the warning level. ∈[1,3], L 1 , L 2 and L 3 Represent low risk, medium risk and high risk respectively. and Represent the abnormality threshold limits of low risk and medium risk respectively.

[0042] According to the abnormality index and warning level, multi-dimensional warning information including abnormality type, degree and recommended measures is generated and sent to the medical staff or parent interface, and the information is recorded in the diagnosis database.

[0043] The multi-dimensional warning information is matched and analyzed with the actual diagnostic feedback provided by medical staff, and the consistency index is calculated to evaluate the accuracy of the diagnostic results.

[0044] According to the consistency index and historical diagnosis data, the comprehensive diagnostic characteristic index is optimized, and the individual difference factor is introduced to dynamically update the weight allocation strategy.

[0045] The diagnostic data of each time and the optimized comprehensive diagnostic characteristic indicators are stored in a children's exclusive health database, and the health trend characteristics are regularly extracted using time series analysis methods to provide a reference for future diagnosis.

[0046] Furthermore, the present embodiment also provides a pediatric respiratory infection diagnosis system based on multimodal data fusion, including: a multimodal data acquisition module, which is used to collect children's real-time physiological data (body temperature, blood oxygen, respiratory rate) through a portable monitoring device to form a multimodal original data set; standardize the collected multimodal data, and use an adaptive noise reduction algorithm to filter environmental interference to ensure the clarity and reliability of the input data; a dynamic weight adjustment and feature extraction module, which is used to apply a dynamic weight adjustment method according to the change trend of physiological data monitored in real time, and assign the priority of each modality data to optimize the adaptability of data fusion; extract key diagnostic features from the processed multimodal data, and use A weight-based fusion algorithm is used to generate comprehensive diagnostic characteristic indicators of children's health status; the abnormal state monitoring and early warning module is used to monitor the fused characteristic indicators in real time through time series analysis to determine whether the child has an abnormal physiological state; when an abnormal pattern is detected, early warning information is automatically generated, and diagnostic result suggestions are sent through the medical staff or parent interface, and high-risk situations are marked; a personalized diagnostic strategy optimization module is used to optimize the comprehensive diagnostic characteristic indicators based on the comparison of the diagnostic results with the actual situation, form a diagnostic strategy that is more suitable for the individual, and improve the accuracy of subsequent diagnosis; each diagnostic data and early warning record is stored in a child-specific health database for long-term trend analysis and future diagnostic reference.

[0047] This embodiment also provides a computer device, which is suitable for the case of a method for diagnosing respiratory tract infections in children based on multimodal data fusion, and includes a memory and a processor; the memory is used to store computer executable instructions, and the processor is used to execute computer executable instructions to implement the method for diagnosing respiratory tract infections in children based on multimodal data fusion as proposed in the above embodiment.

[0048] The computer device may be a terminal, and the computer device includes a processor, a memory, a communication interface, a display screen and an input device connected through a system bus. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The communication interface of the computer device is used to communicate with an external terminal in a wired or wireless manner, and the wireless manner can be achieved through WIFI, an operator network, NFC (near field communication) or other technologies. The display screen of the computer device may be a liquid crystal display screen or an electronic ink display screen, and the input device of the computer device may be a touch layer covered on the display screen, or a key, trackball or touchpad provided on the housing of the computer device, or an external keyboard, touchpad or mouse, etc.

[0049] This embodiment also provides a storage medium on which a computer program is stored. When the program is executed by a processor, the method for diagnosing respiratory tract infections in children based on multimodal data fusion as proposed in the above embodiment is implemented.

[0050] In summary, the present invention solves the problems of low diagnostic accuracy and lack of real-time performance of a single data modality in the diagnosis of children's respiratory infections through the technical means of multimodal data fusion; it gradually realizes high-precision, personalized real-time diagnosis by utilizing data collection, dynamic weight adjustment, feature extraction, time series analysis and model optimization; the setting of each step helps to improve the reliability, real-time performance and accuracy of the diagnosis, and provide comprehensive support for children's health management.

[0051] Example 2: In this Example 2, a method for diagnosing respiratory tract infections in children based on multimodal data fusion is provided. In order to verify the beneficial effects of the present invention, scientific demonstration is carried out through simulation experiments.

[0052] The experiment selected 100 children with suspected respiratory tract infection (hereinafter referred to as "experimental subjects") admitted to a community medical center as the research subjects, ranging from 3 to 8 years old. The equipment used in the experiment includes: portable physiological monitors, experimental data acquisition systems, and diagnostic analysis algorithm platforms.

[0053] Each subject wore a portable monitoring device to collect real-time physiological data for 24 hours, including body temperature (unit: °C), blood oxygen saturation (unit:%) and respiratory rate (unit: times / minute). After the data collection was completed, the data was denoised using an adaptive noise reduction algorithm and standardized to ensure data quality and uniformity.

[0054] Based on time series analysis, the change rate and fluctuation trend of the collected data are calculated to generate change trend indicators (such as body temperature increase rate, respiratory rate fluctuation range, etc.).

[0055] Through the dynamic weight adjustment method, the characteristic indicators of each modal data are calculated, and their correlation with the target diagnostic features is calculated, and low-correlation features are eliminated to finally form a key feature set.

[0056] The integrated diagnostic indicators are modeled in time series and compared with the preset abnormal pattern library. If a pattern with a similarity higher than the threshold is found, the system automatically issues a warning message and gives diagnostic suggestions.

[0057] The weight matrix and anomaly detection algorithm are optimized based on the comparative analysis of diagnostic recommendations and clinical results. The diagnostic records of each experimental subject are stored in the health database and subsequently used for trend analysis and personalized diagnosis optimization.

[0058] Table 1 Diagnosis data of respiratory tract infection Subject No. Body temperature (℃) Blood oxygen saturation (%) Respiratory rate (times / min) Changing trend indicators Comprehensive diagnostic features Diagnosis A01 37.5 96.5 28 0.75 0.82 normal A02 39.2 92 35 1.2 1.1 High risk A03 36.8 98 24 0.5 0.6 normal A04 38 94 32 0.9 1 Medium risk A05 37 97 27 0.6 0.7 normal A06 39 93 34 1.15 1.08 High risk It can be seen from Table 1 that the present invention significantly improves the recognition accuracy of high-risk and medium-risk cases by dynamically adjusting weights to optimize the adaptability of data fusion. For example, the comprehensive diagnostic characteristic values ​​of high-risk cases numbered A02 and A06 are significantly higher than the normal range (below 0.75), and the final diagnosis is consistent with clinical practice.

[0059] Compared with the traditional diagnosis method, the present invention significantly reduces the environmental interference in data collection through the adaptive noise reduction algorithm. For example, the normal cases numbered A01 and A05 still obtained accurate diagnosis results when the respiratory rate data fluctuated slightly, avoiding false positives.

[0060] Traditional diagnostic methods mostly rely on a single physiological indicator, and the accuracy of the diagnostic results is low, especially when the data fluctuates greatly, it is easy to make a misjudgment. The present invention is based on multimodal data fusion, dynamically adjusts the weights and combines time series analysis to effectively improve the diagnostic performance in complex environments.

[0061] The present invention provides a basis for subsequent long-term health trend analysis and personalized diagnosis optimization by constructing an individualized health database.

[0062] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. A method for diagnosing respiratory tract infections in children based on multimodal data fusion, characterized by: include: Step S1: collecting the child's real-time physiological data through a portable monitoring device and performing standardized processing to form a multi-modal raw data set; Step S2: Analyze the change trend of the standardized real-time physiological data based on the time series, and calculate the real-time change rate of the real-time physiological data; calculate the correlation weight of each modality data to the target diagnostic feature, and form a dynamic correlation weight matrix; Extract key features from multimodal data based on the dynamic correlation weight matrix, sort them into a priority list according to the weight values, and retain high-priority features as diagnostic input; Using a weight-based multimodal feature fusion algorithm, high-priority features are weighted and summed to generate a comprehensive diagnostic feature index of the child's health status; Step S3: Through time series analysis, the fused characteristic indicators are monitored in real time to determine whether the child has an abnormal physiological state; Step S4: When an abnormal pattern is detected, a warning message is automatically generated and a diagnosis result suggestion is sent through medical staff; the comprehensive diagnostic characteristic indicators are optimized based on the comparison between the diagnosis result and the actual situation; each diagnosis data and warning record is stored in a child-specific health database.

2. The method for diagnosing respiratory tract infection in children based on multimodal data fusion according to claim 1, characterized in that: The specific implementation process of step S1 includes: After authorization by the user, a portable monitoring device is installed on the user's body surface, and the portable monitoring device is used to detect the user's real-time physiological data, and the physiological data includes body temperature data, blood oxygen data and respiratory rate data; The user's real-time physiological data is standardized and an adaptive noise reduction algorithm is used to filter out environmental interference to ensure the clarity and reliability of the input data.

3. The method for diagnosing respiratory tract infection in children based on multimodal data fusion according to claim 2, characterized in that: The specific implementation process of step S2 includes: Based on the time series, the change trend of the standardized real-time physiological data is analyzed, the real-time change rate of the real-time physiological data is calculated, and the change trend index is generated, as follows: ； in, Indicates j The real-time change rate of real-time physiological data, Indicates j Real-time physiological data t The long-term trend term of time, Indicates j Real-time physiological data t -1 long-term trend term; Calculate the change trend indicator, the calculation formula is as follows: ； in, Indicates j Real-time physiological data change trend indicators, and Respectively represent j The real-time change rate and fluctuation weight of real-time physiological data, Indicates j Real-time physiological data t The short-term fluctuation of time, Represents the standard deviation of short-term volatility; Based on the changing trend indicator and historical health data, calculate the relevance weights of each modality data to the target diagnostic features, and form a dynamic relevance weight matrix, as follows: ； in, Indicates j The relevance weight of the real-time physiological data to the target diagnostic features, Indicates j Real-time physiological data trend indicators and j Real-time physiological data and historical health data The covariance of and Respectively represent j Real-time physiological data trend indicators and j Real-time physiological data and historical health data The standard deviation of Generate dynamic relevance weight matrix ,in, represents the relevance weight of body temperature data to the target diagnostic feature, represents the correlation weight of blood oxygen data to the target diagnostic feature, Represents the relevance weight of respiratory rate data to the target diagnostic feature.

4. The method for diagnosing respiratory tract infection in children based on multimodal data fusion according to claim 3, characterized in that: The specific implementation process of step S3 includes: Segmenting the comprehensive diagnostic characteristic indicators into time series, and constructing indicator changes within continuous time intervals into dynamic characteristic sequences; The dynamic feature sequence is matched with the template pattern in the pre-built abnormal pattern library, the matching similarity index is calculated, and the high similarity pattern is screened, as follows: ； in, Represents the matching similarity index, Indicates j The weight of the change trend indicator of real-time physiological data, Indicates the exception pattern library k The first mode j A real-time physiological data change trend indicator, max() represents the maximum value function; Set the filtering rules as follows: ； in, Represents the filtered set of high similarity patterns, Represents a template pattern in the exception pattern library, represents a library of pre-built exception patterns, Indicates the preset similarity threshold.

5. The method for diagnosing respiratory tract infection in children based on multimodal data fusion according to claim 4, characterized in that: The specific implementation process of step S3 also includes: According to the matched high similarity pattern, combined with the change amplitude and deviation range of the actual feature sequence, the abnormality index is calculated, and the warning level is generated according to the threshold classification, as follows: ； ； in, Represents the abnormality index, Indicates j The weight of the change trend indicator of real-time physiological data, Represents the high similarity pattern set The mean of all modes in , Indicates the exception pattern library k The first mode j Real-time physiological data change trend indicators, Represents a collection of highly similar patterns The total number of medium and high similarity patterns.

6. The method for diagnosing respiratory tract infection in children based on multimodal data fusion according to claim 5, characterized in that: The specific implementation process of step S4 includes: To set the warning level rules: ； in, Indicates the warning level. ∈[1,3], L 1. L 2 and L 3 Represent low risk, medium risk and high risk respectively. and Represent the abnormality threshold limits of low risk and medium risk respectively; Generate multi-dimensional warning information including abnormal type, degree and recommended measures according to the abnormality index and warning level; The multi-dimensional warning information is matched and analyzed with the actual diagnostic feedback provided by medical staff, and the consistency index is calculated to evaluate the accuracy of the diagnostic results; each diagnostic data and optimized comprehensive diagnostic feature indicators are stored in a child-specific health database, and health trend characteristics are regularly extracted using time series analysis methods.

7. A pediatric respiratory tract infection diagnosis system based on multimodal data fusion, based on the pediatric respiratory tract infection diagnosis method based on multimodal data fusion according to any one of claims 1 to 6, characterized in that: include: A multimodal data acquisition module, used to collect real-time physiological data of children through portable monitoring equipment to form a multimodal raw data set; Standardize the collected multimodal data and use an adaptive noise reduction algorithm to filter out environmental interference to ensure the clarity and reliability of the input data; The dynamic weight adjustment and feature extraction module is used to apply the dynamic weight adjustment method according to the trend of physiological data changes monitored in real time, assign the priority of each modality data, and optimize the adaptability of data fusion; extract key diagnostic features from the processed multimodal data, and use the weight-based fusion algorithm to generate comprehensive diagnostic feature indicators of children's health status; The abnormal state monitoring and early warning module is used to monitor the fused characteristic indicators in real time through time series analysis to determine whether the child has an abnormal physiological state; when an abnormal pattern is detected, early warning information is automatically generated and the diagnosis result suggestions are sent to medical staff; The personalized diagnosis strategy optimization module is used to optimize the comprehensive diagnostic characteristic indicators based on the comparison of the diagnosis results with the actual situation, so as to form a diagnosis strategy that is more suitable for the individual; and store each diagnosis data and warning record in a child-specific health database.

8. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the method for diagnosing respiratory tract infection in children based on multimodal data fusion as described in any one of claims 1 to 6 are implemented.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method for diagnosing respiratory tract infection in children based on multimodal data fusion as described in any one of claims 1 to 6 are implemented.

Images