Multi-modal physiological data fusion method, computer equipment and readable storage medium

Abstract

The invention provides a multi-modal physiological data fusion method, computer equipment and a readable storage medium, and relates to the technical field of medical health. According to the method, after the respective time sequence data of the multiple target physiological indexes monitored by the physiological signal telemetering equipment in the target time window are obtained, the medical knowledge base is called to construct the target knowledge graph matched with the multiple target physiological indexes; a physiological load influence degree evaluation model is called to predict physiological load influence degree weights of various target physiological indexes in the target time window based on the target knowledge graph and the time sequence data; then, according to the time sequence data and the physiological load influence degree weights corresponding to the multiple target physiological indexes, physiological load state evaluation is carried out, and therefore physiological index time sequence data deep fusion is carried out by deeply digging the medical incidence relation among the multiple physiological indexes; and the corresponding physiological data fusion result can be ensured to accurately evaluate the actual physical condition of the monitored object by fully utilizing the knowledge in the medical field.

Description

Technical Field

[0001] The present invention relates to the field of medical health technology, and in particular to a multimodal physiological data fusion method, a computer device and a readable storage medium. Background Art

[0002] With the continuous development of science and technology, people's medical concepts have gradually changed from traditional disease treatment to early prevention and health care. They often expect to remotely and timely understand the physical health status of the monitored persons (for example, patients, the elderly or children) to achieve early detection and prediction of common diseases or sudden diseases in sub-healthy and latent disease populations. For the human body, physiological index parameters such as blood pressure, heart rate, electrocardiogram, blood oxygen saturation, body temperature, exercise intensity, and respiratory rate are the most basic and important life indicators of the human body. The monitoring of these physiological index parameters can timely grasp the changes in the physical condition of the monitored persons and maximize the protection of people's physical health. Therefore, how to continuously collect various physiological index data of the monitored persons and use the fusion of multiple physiological index data to accurately evaluate the actual physical condition of the monitored persons is an important technical problem that needs to be solved in the current physiological signal telemetry process. Summary of the invention

[0003] In view of this, the purpose of the present invention is to provide a multimodal physiological data fusion method, a computer device and a readable storage medium, which can deeply explore the medical correlation between multiple physiological indicators, and deeply fuse the time series data of multiple physiological indicators based on the mined medical correlation, to ensure that the corresponding physiological data fusion results can make full use of the medical field knowledge to accurately evaluate the actual physical condition of the monitored object, and facilitate the subsequent use of data analysis technologies such as machine learning to perform rhythm analysis based on the physiological data fusion results.

[0004] In order to achieve the above purpose, the technical solution adopted by the embodiment of the present invention is as follows:

[0005] In a first aspect, the present invention provides a multimodal physiological data fusion method, the multimodal physiological data fusion method comprising:

[0006] Acquire the time series data of various target physiological indicators monitored by the physiological signal telemetry device for the target object within the target time window;

[0007] Calling a medical knowledge base to construct a knowledge graph for the multiple target physiological indicators to obtain a target knowledge graph adapted to the multiple target physiological indicators;

[0008] Calling a pre-stored physiological load influence evaluation model to perform influence weight prediction based on the target knowledge graph and the time series data of each of the multiple target physiological indicators, and obtaining the physiological load influence weight of each of the multiple target physiological indicators within the target time window;

[0009] Performing a physiological load state assessment based on the time series data corresponding to each of the multiple target physiological indicators and the physiological load influence weights to obtain a comprehensive physiological load assessment score of the target object within the target time window;

[0010] The comprehensive physiological load assessment score is output as a result of physiological data fusion of the multiple target physiological indicators at the target time window.

[0011] In an optional embodiment, the step of obtaining time series data of each of a plurality of target physiological indicators monitored by the physiological signal telemetry device for the target object within the target time window includes:

[0012] Acquire original physiological data corresponding to the multiple target physiological indicators respectively collected by the physiological signal telemetry device at different time points within the target time window;

[0013] For each target physiological index, all original physiological data of the target physiological index within the target time window are dimensionally processed to obtain normalized physiological data corresponding to different time points of the target physiological index within the target time window;

[0014] A time series is constructed for all original physiological data and all normalized physiological data of the target physiological index within the target time window to obtain the time series data of the target physiological index within the target time window.

[0015] In an optional embodiment, the time series data of each target physiological indicator includes normalized physiological data corresponding to the target physiological indicator at different time points within the target time window, and the step of performing physiological load state assessment according to the time series data corresponding to each of the multiple target physiological indicators and the physiological load influence weight to obtain the comprehensive physiological load assessment score of the target object within the target time window includes:

[0016] For each time point within the target time window, according to the normalized physiological data of each of the multiple target physiological indicators at the time point, the physiological load size caused by each of the multiple target physiological indicators to the target object at the time point is calculated, and the physiological load size and physiological load influence weight corresponding to each of the multiple target physiological indicators at the time point are weighted summed to obtain the comprehensive physiological load size of the target object at the time point;

[0017] According to the time sequence of each time point in the target time window, an ascending time weight configuration is performed for each time point in the target time window to obtain an actual time weight of each time point in the target time window;

[0018] A weighted sum operation is performed on the comprehensive physiological load magnitudes and actual time weights of all time points within the target time window to obtain the comprehensive physiological load assessment score of the target object within the target time window.

[0019] In an optional implementation, the multimodal physiological data fusion method further includes:

[0020] Performing an assessment accuracy check on the obtained comprehensive physiological load assessment score to determine whether the actual assessment accuracy of the comprehensive physiological load assessment score meets the expected accuracy standard;

[0021] When it is determined that the actual assessment accuracy of the comprehensive physiological load assessment score reaches the expected accuracy standard, the step of outputting the comprehensive physiological load assessment score as the physiological data fusion result of the multiple target physiological indicators at the target time window is performed accordingly.

[0022] In an optional embodiment, the step of performing an assessment accuracy check on the obtained comprehensive physiological load assessment score to determine whether the actual assessment accuracy of the comprehensive physiological load assessment score reaches the expected accuracy standard includes:

[0023] Obtaining an actual comprehensive physiological load true value of the target object within the target time window, and calculating a load assessment error between the comprehensive physiological load assessment score and the actual comprehensive physiological load true value;

[0024] Detecting whether the load assessment error exceeds a preset assessment error threshold;

[0025] If it is detected that the load assessment error exceeds the preset assessment error threshold, it is determined that the actual assessment accuracy of the comprehensive physiological load assessment score does not meet the expected accuracy standard; otherwise, it is determined that the actual assessment accuracy of the comprehensive physiological load assessment score meets the expected accuracy standard.

[0026] In an optional implementation, the multimodal physiological data fusion method further includes:

[0027] In the case where it is determined that the actual evaluation accuracy of the comprehensive physiological load evaluation score does not reach the expected accuracy standard, calling a graph embedding model matching the multiple target physiological indicators to perform indicator correlation detection based on the target knowledge graph, and determining multiple physiological indicators to be fused with strong correlation among the multiple target physiological indicators;

[0028] Determine the time series fusion weights of the various physiological indicators to be fused according to the actual indicator correlations among the various physiological indicators to be fused;

[0029] According to the respective time series fusion weights of the multiple physiological indicators to be fused, feature fusion is performed on the respective time series data of the multiple physiological indicators to be fused, so as to obtain the physiological signal time series fusion data of the target object within the target time window;

[0030] The physiological signal time series fusion data is directly output as the physiological data fusion result of the multiple target physiological indicators at the target time window.

[0031] In an optional embodiment, the step of calling a graph embedding model matching the multiple target physiological indicators to perform indicator correlation detection based on the target knowledge graph includes:

[0032] Calling the graph embedding model to map the medical association relationships between the multiple target physiological indicators included in the target knowledge graph into the embedding space, and obtaining actual embedding vectors of each of the multiple target physiological indicators in the embedding space;

[0033] Vector correlations are calculated according to actual embedding vectors of each of the multiple target physiological indicators in the embedding space to obtain actual indicator correlations between the multiple target physiological indicators.

[0034] In an optional embodiment, the step of determining a plurality of physiological indicators to be fused that are strongly correlated among the plurality of target physiological indicators includes:

[0035] According to the actual indicator correlations between the multiple target physiological indicators, the multiple target physiological indicators are paired according to a preset correlation threshold to obtain multiple indicator pairing combinations, wherein each indicator pairing combination is composed of two target physiological indicators whose corresponding actual indicator correlations exceed the preset correlation threshold;

[0036] Physiological indicator occurrence statistics are performed on the obtained multiple indicator pairing combinations, and all target physiological indicators involved in the multiple indicator pairing combinations are respectively used as a physiological indicator to be fused.

[0037] In a second aspect, the present invention provides a computer device, comprising a processor and a memory, wherein the memory stores a computer program that can be executed by the processor, and the processor can execute the computer program to implement the multimodal physiological data fusion method described in any one of the aforementioned embodiments.

[0038] In a third aspect, the present invention provides a readable storage medium having a computer program stored thereon, and when the computer program is executed by a computer device, the multimodal physiological data fusion method described in any one of the aforementioned embodiments is implemented.

[0039] In this case, the beneficial effects of the embodiments of the present invention may include the following:

[0040] After obtaining the time series data of each of the multiple target physiological indicators monitored by the physiological signal telemetry device for the target object within the target time window, the present invention will call the medical knowledge base to construct a knowledge graph for the multiple target physiological indicators, obtain a target knowledge graph adapted to the multiple target physiological indicators, and call the pre-stored physiological load influence evaluation model to predict the influence weight based on the target knowledge graph and the time series data of each of the multiple target physiological indicators, and obtain the physiological load influence weights of each of the multiple target physiological indicators within the target time window, and then perform a physiological load state evaluation based on the time series data corresponding to each of the multiple target physiological indicators and the physiological load influence weight, and obtain a comprehensive physiological load evaluation score of the target object within the target time window that can be used as a physiological data fusion result, thereby deeply digging into the medical association relationship between the multiple physiological indicators through the construction of the knowledge graph, and then deeply fusing the multiple physiological indicator time series data based on the mined medical association relationship, ensuring that the corresponding physiological data fusion result can make full use of the medical field knowledge to accurately evaluate the actual physical condition of the monitored object, and facilitate the subsequent use of data analysis technologies such as machine learning to perform rhythm analysis based on the physiological data fusion result.

[0041] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, preferred embodiments are given below and described in detail with reference to the accompanying drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings required for use in the embodiments are briefly introduced below. It should be understood that the following drawings only show certain embodiments of the present invention and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without creative work.

[0043] Figure 1 A schematic diagram of the composition of a computer device provided in an embodiment of the present invention;

[0044] Figure 2 One of the flowcharts of the multimodal physiological data fusion method provided in an embodiment of the present invention;

[0045] Figure 3 for Figure 2 A schematic flow chart of the sub-steps included in step 210;

[0046] Figure 4 for Figure 2 A flowchart of the sub-steps included in step 240;

[0047] Figure 5 A second flowchart of a multimodal physiological data fusion method provided by an embodiment of the present invention;

[0048] Figure 6 for Figure 5 One of the flowchart diagrams of the sub-steps included in step 270;

[0049] Figure 7 for Figure 5 The second flowchart diagram of the sub-steps included in step 270.

[0050] Icons: 10 - computer device; 11 - memory; 12 - processor; 13 - communication unit. DETAILED DESCRIPTION

[0051] In order to make the purpose, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Generally, the components of the embodiments of the present invention described and shown in the drawings here can be arranged and designed in various different configurations.

[0052] Therefore, the following detailed description of the embodiments of the present invention provided in the accompanying drawings is not intended to limit the scope of the invention claimed for protection, but merely represents selected embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0053] It should be noted that similar reference numerals and letters denote similar items in the following drawings, and therefore, once an item is defined in one drawing, further definition and explanation thereof is not required in subsequent drawings.

[0054] In the description of the present invention, it should be understood that the terms "center", "up", "down", "left", "right", "vertical", "horizontal", "inside", "outside" and the like indicate orientations or positional relationships based on the orientations or positional relationships shown in the accompanying drawings, or are the orientations or positional relationships in which the product of the invention is conventionally placed when in use, or are the orientations or positional relationships conventionally understood by those skilled in the art. They are only for the convenience of describing the present invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore should not be understood as a limitation on the present invention.

[0055] In the description of the present invention, it is also necessary to explain that, unless otherwise clearly specified and limited, the terms "set", "install", "connect", and "connect" should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection, or it can be indirectly connected through an intermediate medium, or it can be the internal communication of two elements. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0056] In addition, in the description of the present invention, it is also necessary to understand that the relational terms such as the terms "first" and "second" are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that the process, method, article or equipment including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or equipment. In the absence of further restrictions, the elements defined by the sentence "comprise one..." do not exclude the existence of other identical elements in the process, method, article or equipment including the elements. For those of ordinary skill in the art, the specific meanings of the above terms in the present invention can be understood in specific circumstances.

[0057] In conjunction with the accompanying drawings, some embodiments of the present invention are described in detail below. In the absence of conflict, the following embodiments and features in the embodiments may be combined with each other.

[0058] Please refer to Figure 1 , Figure 1: is a schematic diagram of the composition of the computer device 10 provided in an embodiment of the present invention. In an embodiment of the present invention, the computer device 10 can be connected to the physiological signal telemetry device in communication to obtain the original physiological data corresponding to each of the multiple physiological indicators collected by the physiological signal telemetry device in real time for the target object, and deeply fuse the physiological indicator time series data by digging into the medical association between the multiple physiological indicators, ensuring that the physiological data fusion results finally obtained for the multiple physiological indicators can make full use of the medical field knowledge to accurately evaluate the actual physical condition of the target object, and facilitate the subsequent use of data analysis technologies such as machine learning to perform rhythm analysis based on the physiological data fusion results. Among them, the physiological signal telemetry device is worn on the target object, and the target object is the monitored object targeted by the physiological signal telemetry device; the physiological signal telemetry device may include a plurality of physiological signal detection sensors, each physiological signal detection sensor independently collects the original physiological data of a physiological indicator, and all physiological indicators supported by the physiological signal telemetry device for monitoring may include but are not limited to: blood pressure, heart rate, electrocardiogram, blood oxygen saturation, body temperature, exercise intensity, radiation energy, etc.; the original physiological data acquired by the computer device 10 is collected by at least two physiological signal detection sensors actually used by the physiological signal telemetry device within the corresponding sliding time window; the computer device 10 may be, but not limited to, a laptop computer, a personal computer, a server, a gateway device, etc.

[0059] In the embodiment of the present invention, the computer device 10 may include a memory 11, a processor 12, and a communication unit 13. The memory 11, the processor 12, and the communication unit 13 are electrically connected to each other directly or indirectly to achieve data transmission or interaction. For example, the memory 11, the processor 12, and the communication unit 13 may be electrically connected to each other via one or more communication buses or signal lines.

[0060] In this embodiment, the memory 11 may be, but not limited to, a random access memory (RAM), a read only memory (ROM), a programmable read-only memory (PROM), an erasable programmable read-only memory (EPROM), an electrically erasable read-only memory (EEPROM), etc. The memory 11 is used to store a computer program, and the processor 12 may execute the computer program accordingly after receiving an execution instruction.

[0061] In addition, the memory 11 can also be used to store a physiological load impact assessment model and a variety of graph embedding (GE) models, wherein the number of physiological indicators and / or physiological indicator composition corresponding to each of the various graph embedding models are different from each other. Among them, the physiological load impact assessment model is used to evaluate the impact weights of each of the multiple physiological indicators actually monitored by the physiological signal telemetry device within a certain sliding time window at the physiological load level; for the physiological load impact assessment model, first, multiple medical experts are required to conduct physiological load impact weight assessment based on the physiological indicator time series data measured by different physiological indicator combinations in multiple time periods, so as to determine the schematic physiological load impact weights of various physiological indicators in different physiological indicator combinations under different indicator measurement error states, and then model training is performed through a machine learning algorithm based on the determined schematic physiological load impact weight data to obtain a physiological load impact assessment model with high impact weight prediction accuracy; each graph embedding model is used to map the medical association relationship between the multiple physiological indicators supported into a low-dimensional continuous vector space (i.e., embedding space) to visualize the actual correlation of the aforementioned multiple physiological indicators at the data level, wherein the actual embedding vectors of two physiological indicators with larger actual correlations in the embedding space are closer.

[0062] In this embodiment, the processor 12 may be an integrated circuit chip with signal processing capability. The processor 12 may be a general-purpose processor, including a central processing unit (CPU), a graphics processing unit (GPU), a network processor (NP), a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field programmable gate array (FPGA), or at least one of other programmable logic devices, discrete gate or transistor logic devices, and discrete hardware components. The general-purpose processor may be a microprocessor or the processor may be any conventional processor, etc., which may implement or execute the disclosed methods, steps, and logic block diagrams in the embodiments of the present invention.

[0063] In this embodiment, the communication unit 13 is used to establish a communication connection between the computer device 10 and other electronic devices through a network, and to send and receive data through the network, wherein the network includes a wired communication network and a wireless communication network. For example, the computer device 10 can obtain the original physiological data of the various physiological indicators monitored by the physiological signal telemetry device for the target object within a specific time window through the communication unit 13; the computer device 10 can communicate with a server with a medical knowledge base through the communication unit 13, so as to call the medical knowledge base by accessing the server to perform medical association mining for the various physiological indicators actually monitored by the physiological signal telemetry device.

[0064] In this embodiment, the computer device 10 may pre-store a specific computer program related to multimodal physiological data operations in the memory 11, and drive the processor 12 to execute the specific computer program stored in the memory 11 accordingly, so as to deeply explore the medical correlation between the various physiological indicators actually involved in the physiological signal telemetry equipment within a specific time window through the construction of a knowledge graph, and deeply fuse the time series data corresponding to the various physiological indicators based on the mined medical correlation, to ensure that the final physiological data fusion result can make full use of the medical field knowledge to accurately evaluate the actual physical condition of the monitored object, so as to facilitate the subsequent use of data analysis technologies such as machine learning to perform rhythm analysis based on the physiological data fusion results.

[0065] Understandably, Figure 1 The block diagram shown is only a schematic diagram of a composition of the computer device 10. The computer device 10 may also include Figure 1 More or fewer components as shown, or with Figure 1 Different configurations are shown. Figure 1Each component shown in the figure can be implemented by hardware, software or a combination thereof.

[0066] In the present invention, in order to ensure that the computer device 10 can deeply explore the medical correlation between the various physiological indicators actually involved in the physiological signal telemetry device within a specific time window to perform deep fusion of physiological indicator time series data, so that the final physiological data fusion result can make full use of medical field knowledge to accurately evaluate the actual physical condition of the monitored object, and facilitate the subsequent use of data analysis technologies such as machine learning to perform rhythm analysis based on the physiological data fusion results, the embodiment of the present invention provides a multimodal physiological data fusion method to achieve the above purpose. The multimodal physiological data fusion method provided by the present invention is described in detail below.

[0067] Please refer to Figure 2 , Figure 2 FIG. 1 is one of the flow charts of the multimodal physiological data fusion method provided in the embodiment of the present invention. In the embodiment of the present invention, Figure 2 The multimodal physiological data fusion method shown may include steps 210 to 250 .

[0068] Step 210: acquiring respective time series data of a plurality of target physiological indicators monitored by the physiological signal telemetry device for the target object within the target time window.

[0069] In this embodiment, the target time window is any sliding time window of the physiological signal telemetry device in the process of physiological signal monitoring; each target physiological indicator is a physiological indicator actually monitored by the physiological signal telemetry device for the target object within the target time window; the target physiological indicator composition corresponding to the physiological signal telemetry device in different sliding time windows can be completely the same, partially the same, or completely different. For example, the target physiological indicator composition corresponding to the physiological signal telemetry device in a certain sliding time window includes blood pressure, heart rate, electrocardiogram, blood oxygen saturation, body temperature and exercise intensity, and the target physiological indicator composition corresponding to the physiological signal telemetry device in a certain sliding time window includes blood pressure, heart rate, electrocardiogram, blood oxygen saturation, body temperature, exercise intensity and respiratory rate; the time series data of each target physiological indicator in the target time window includes the normalized physiological data and the original physiological data corresponding to the target physiological indicator at different time points in the target time window.

[0070] Optionally, see Figure 3 , Figure 3 yes Figure 2Schematic diagram of the flow chart of the sub-steps included in step 210. In this embodiment, step 210 may include sub-steps 211 to 213 to ensure that the time series data constructed for each target physiological indicator can minimize the impact of numerical differences, meet the data requirements of the machine learning algorithm that may be used later, and retain the traceability and efficient query possibility of the physiological data.

[0071] Sub-step 211, obtaining original physiological data corresponding to a plurality of target physiological indicators respectively collected by the physiological signal telemetry device at different time points within the target time window.

[0072] Sub-step 212, for each target physiological index, dimensionless processing is performed on all original physiological data of the target physiological index within the target time window to obtain normalized physiological data corresponding to different time points of the target physiological index within the target time window.

[0073] In this embodiment, after obtaining all the original physiological data corresponding to each of the multiple target physiological indicators within the target time window, the computer device 10 can align the timestamps of all the original physiological data of each of the multiple target physiological indicators according to a preset time granularity (for example, one time point every 5 minutes), and perform missing data interpolation processing on the original physiological data of each of the multiple target physiological indicators within the target time window, and then perform dimensionless processing on all the original physiological data of each target physiological indicator after the interpolation processing, so as to map all the original physiological data of the target physiological indicator to a numerical range of 0 to 1, so as to achieve the normalization effect of the physiological data of the target physiological indicator, and obtain the normalized physiological data corresponding to the target physiological indicator at different time points within the target time window, while minimizing the impact of the numerical differences between different target physiological indicators to meet the data requirements of the machine learning algorithms that may be used later.

[0074] Among them, there are differences in the dimensionless processing schemes corresponding to different physiological indicators. For example, for each physiological indicator of heart rate, blood pressure, body temperature and respiratory rate, the physiological indicator data can be normalized based on normalized, standardized or centralized data processing methods; for blood oxygen saturation, the normal value range (for example, 80-100%) can be set, and the corresponding original physiological data that exceeds the normal value range can be replaced with the range critical value of the normal value range (for example, the blood oxygen saturation value collected at a certain time point is 75%, then the blood oxygen saturation value must be replaced by 80%; the blood oxygen saturation value collected at a certain time point is 120%, then the blood oxygen saturation value must be replaced by 100%). For exercise intensity, the exercise intensity data of the monitored object can be mapped to a range of 0 to 1 by dividing the exercise intensity of the number of steps taken by humans in a unit time (for example, 0.5 hours) (for example, 0 to 500 steps correspond to an intensity of 0.2 representing "stationary", 501 to 2000 steps correspond to an intensity of 0.4 representing "mild", 2001 to 4000 steps correspond to an intensity of 0.6 representing "moderate", 4001 to 6000 steps correspond to an intensity of 0.8 representing "violent", and more than 6000 steps correspond to an intensity of 1 representing "extreme").

[0075] In one implementation of the present embodiment, the computer device 10 may convert all the original physiological data of the multiple target physiological indicators after interpolation processing into RDF (Resource Description Framework) format, and then perform SPARQL query on all the original physiological data of each target physiological indicator to determine the mathematical statistics (e.g., mean, standard deviation, etc.) between all the original physiological data of the target physiological indicator, and then perform data normalization reasoning on the original physiological data and mathematical statistics corresponding to the target physiological indicator by calling the SPIN (SPARQL Inferencing Notation) rule for characterizing the dimensionless processing scheme that matches the target physiological indicator, so as to obtain the normalized physiological data corresponding to the target physiological indicator at different time points within the target time window.

[0076] Sub-step 213, constructing a time series of all original physiological data and all normalized physiological data of the target physiological index within the target time window, to obtain the time series data of the target physiological index within the target time window.

[0077] Therefore, the present invention can ensure that the time series data constructed for each target physiological indicator can minimize the impact of numerical differences by executing the above sub-steps 211 to 213, meet the data requirements of machine learning algorithms that may be used subsequently, while retaining the traceability and efficient query possibility of physiological data, and can achieve efficient query of physiological data by establishing data indexes according to time.

[0078] Step 220, calling the medical knowledge base to construct a knowledge graph for multiple target physiological indicators, and obtaining a target knowledge graph adapted to the multiple target physiological indicators.

[0079] In this embodiment, the computer device 10 can call the medical knowledge base through the communication unit 13 to obtain the medical correlation knowledge of the multiple target physiological indicators when facing different diseases (for example, when a certain lung disease will show symptoms such as increased breathing rate, increased body temperature, increased heart rate, and decreased activity due to general weakness, blood oxygen saturation and breathing rate are directly related, body temperature and breathing rate are directly related, blood oxygen saturation and heart rate are directly related, blood oxygen saturation and body temperature are directly related, exercise intensity and heart rate, heart rate and body temperature are directly related, etc.), and then construct a knowledge graph for the multiple target physiological indicators based on the acquired medical correlation knowledge, so as to deeply explore the medical correlation relationship between the multiple target physiological indicators actually involved in the physiological signal telemetry equipment within the target time window, and obtain a target knowledge graph adapted to the multiple target physiological indicators.

[0080] Step 230, calling a pre-stored physiological load influence evaluation model to perform influence weight prediction based on the target knowledge graph and the time series data of each of the multiple target physiological indicators, and obtaining the physiological load influence weights of each of the multiple target physiological indicators within the target time window.

[0081] In this embodiment, the computer device 10 may preferentially extract the error characteristic parameters of each of the multiple target physiological indicators within the target time window, which are used to characterize the indicator measurement error state of the corresponding target physiological indicators, from the acquired time series data of each of the multiple target physiological indicators, and then input the constructed target knowledge graph and the extracted error characteristic parameters of each of the multiple target physiological indicators into the physiological load influence assessment model. The physiological load influence assessment model analyzes the medical associations between the multiple target physiological indicators included in the target knowledge graph and the indicator measurement error states of the multiple target physiological indicators within the target time window at the physiological load level at the physiological load level at the target time window, so as to ensure that the physiological load influence weights of each of the multiple target physiological indicators at the physiological load level at the target time window that are finally determined can effectively reflect the physiological indicator correlation and indicator data reliability.

[0082] Among them, the weight prediction logic of the physiological load influence assessment model may include, but is not limited to: "If the number of other physiological indicators associated with a certain physiological indicator is greater, the physiological load influence weight corresponding to this physiological indicator is greater", "If the direct correlation between a certain physiological indicator and other physiological indicators is greater, this physiological indicator needs to maintain a larger physiological load influence weight relative to the other physiological indicators", "If the indirect correlation between a certain physiological indicator and other physiological indicators is greater, this physiological indicator needs to maintain a larger physiological load influence weight relative to the other physiological indicators", "If the error characteristic parameter of a certain physiological indicator indicates that the indicator measurement error of the corresponding physiological indicator is greater, the physiological load influence weight corresponding to this physiological indicator is smaller", "The influence of the indicator measurement error on the size of the physiological load influence weight is greater than the influence of the physiological indicator association relationship on the size of the physiological load influence weight", "The influence of the physiological indicator association relationship on the size of the physiological load influence weight is greater than the influence of the indicator measurement error on the size of the physiological load influence weight", etc.

[0083] In one implementation of this embodiment, a mathematical statistic (e.g., mean, variance, standard deviation, etc.) between all the original physiological data of the target physiological index in the corresponding time series data can be used as the error characteristic parameter of the target physiological index. For example, if the mean of the original physiological data of the target physiological index in a certain time window is used as the error characteristic parameter, the greater the absolute value of the difference between the means of the original physiological data of the target physiological index in two adjacent time windows, the greater the indicator measurement error of the target physiological index in the latter time window; if the variance of the original physiological data or the standard deviation of the original physiological data of the target physiological index in a certain time window is used as the error characteristic parameter, the greater the variance of the original physiological data or the standard deviation of the original physiological data of the target physiological index in the corresponding time window, the greater the indicator measurement error of the target physiological index in the corresponding time window.

[0084] Step 240 , performing a physiological load state assessment based on the time series data corresponding to each of the multiple target physiological indicators and the physiological load influence weights, and obtaining a comprehensive physiological load assessment score of the target object within the target time window.

[0085] In this embodiment, the computer device 10 can calculate the comprehensive physiological load caused to the target object by the multiple target physiological indicators at the same time point in the target time window, and then perform weighted operation on the comprehensive physiological load sizes corresponding to all time points in the target time window to obtain the comprehensive physiological load assessment score of the target object in the target time window, so as to deeply integrate the physiological indicator data based on the medical association relationship between the multiple target physiological indicators mined, and ensure that the final calculated comprehensive physiological load assessment score can make full use of medical knowledge to accurately assess the actual physical condition of the target object.

[0086] Optionally, see Figure 4 , Figure 4 yes Figure 2 In this embodiment, step 240 may include sub-steps 241 to 243 to ensure that the corresponding calculated comprehensive physiological load assessment score can more comprehensively assess the actual physical condition of the target object within the target time window.

[0087] Sub-step 241, for each time point within the target time window, calculate the physiological load caused by each of the multiple target physiological indicators to the target object at that time point based on the normalized physiological data of each of the multiple target physiological indicators at that time point, and perform a weighted sum operation on the physiological load size and physiological load influence weight corresponding to each of the multiple target physiological indicators at that time point to obtain the comprehensive physiological load size of the target object at that time point.

[0088] In this embodiment, the physiological load calculation strategies for different target physiological indicators at a single time point are different from each other; for example, for each physiological indicator of heart rate, blood pressure, body temperature and respiratory rate, the corresponding physiological load calculation strategy can be expressed as "(normalized physiological data of the corresponding physiological indicator at a certain time point - the standard normalized value of the corresponding physiological indicator when it is normal)^2"; for blood oxygen saturation or exercise intensity, the corresponding physiological load calculation strategy can be expressed as "using the normalized physiological data of the corresponding physiological indicator at a certain time point directly as the physiological load size at the corresponding time point".

[0089] Sub-step 242, according to the time sequence of each time point in the target time window, an ascending time weight configuration is performed for each time point in the target time window to obtain an actual time weight of each time point in the target time window.

[0090] In this embodiment, the later the actual time weight of the time point in the target time window is, the greater the actual time weight is, and the more effective the comprehensive physiological load corresponding to the corresponding time point is in the physical condition assessment process. The sum of the actual time weights of all time points in the target time window is 1.

[0091] Sub-step 243, performing a weighted sum operation on the comprehensive physiological load magnitude and the actual time weight of each of all time points within the target time window to obtain the comprehensive physiological load assessment score of the target object within the target time window.

[0092] Therefore, the present invention can more comprehensively evaluate the actual physical condition of the target object within the target time window from the dimensions of indicator data timeliness and indicator data correlation by executing the above sub-steps 241 to 243, thereby ensuring that the corresponding calculated comprehensive physiological load assessment score can accurately describe the actual physical condition of the target object.

[0093] Step 250: Output the comprehensive physiological load assessment score as a physiological data fusion result of multiple target physiological indicators at a target time window.

[0094] Therefore, the present invention can directly execute the above steps 210 to 250 to deeply explore the medical correlation between the various physiological indicators actually involved in the physiological signal telemetry equipment within a specific time window to perform deep fusion of physiological indicator time series data, so that the final physiological data fusion result can make full use of medical knowledge to accurately evaluate the actual physical condition of the monitored object, and facilitate the subsequent use of data analysis technologies such as machine learning to perform rhythm analysis based on the physiological data fusion result.

[0095] Optionally, see Figure 5 , Figure 5 FIG. 2 is a flow chart of a multimodal physiological data fusion method provided in an embodiment of the present invention. Figure 2 Compared with the multimodal physiological data fusion method shown in Figure 5 The multimodal physiological data fusion method shown further includes a step 260 between step 240 and step 250 to verify and evaluate the data reliability of the calculated comprehensive physiological load assessment score.

[0096] Step 260 , performing an assessment accuracy check on the obtained comprehensive physiological load assessment score to determine whether the actual assessment accuracy of the comprehensive physiological load assessment score reaches the expected accuracy standard.

[0097] In this embodiment, after the computer device 10 executes the above step 240, it will determine whether the actual assessment accuracy of the calculated comprehensive physiological load assessment score reaches the expected accuracy standard by executing step 260, and complete the verification and evaluation operation of the data reliability of the comprehensive physiological load assessment score. Then, when it is determined that the actual assessment accuracy of the comprehensive physiological load assessment score reaches the expected accuracy standard, the above step 250 is executed accordingly to ensure the data reliability of the corresponding physiological data fusion result.

[0098] Optionally, in an implementation of this embodiment, the step of performing an evaluation accuracy check on the obtained comprehensive physiological load evaluation score to determine whether the actual evaluation accuracy of the comprehensive physiological load evaluation score reaches the expected accuracy standard may include:

[0099] Obtaining an actual comprehensive physiological load true value of the target object within the target time window, and calculating a load assessment error between the comprehensive physiological load assessment score and the actual comprehensive physiological load true value;

[0100] Detecting whether the load assessment error exceeds a preset assessment error threshold;

[0101] If it is detected that the load assessment error exceeds the preset assessment error threshold, it is determined that the actual assessment accuracy of the comprehensive physiological load assessment score does not meet the expected accuracy standard; otherwise, it is determined that the actual assessment accuracy of the comprehensive physiological load assessment score meets the expected accuracy standard.

[0102] The actual true value of the comprehensive physiological load may be acquired by a professional monitoring device for the comprehensive physiological load of the target object, or may be acquired by a comprehensive evaluation of the physiological changes of the target object within a target time window by at least one expert in the medical field.

[0103] Therefore, the present invention can be Figure 2 Step 260 is performed based on the multimodal physiological data fusion method shown in the figure to verify and evaluate the data reliability of the calculated comprehensive physiological load assessment score to ensure that the physiological data fusion result finally outputted has sufficiently strong data reliability.

[0104] Optionally, in an embodiment of the present invention, Figure 2 Compared with the multimodal physiological data fusion method shown in Figure 5The multimodal physiological data fusion method shown may also include steps 270 to 2110, so that when it is determined that the data reliability of the calculated comprehensive physiological load assessment score is unqualified, based on the medical correlation relationship between the various physiological indicators actually involved in the physiological signal telemetry equipment within a specific time window, the aforementioned multiple physiological indicators are adaptively screened for strongly correlated indicators and deeply fused with the characteristics of the indicator data, to ensure that the final output physiological data fusion result has sufficiently strong data reliability, and can also accurately evaluate the actual physical condition of the monitored subject, facilitating subsequent rhythm analysis operations.

[0105] Step 270, calling a graph embedding model that matches the multiple target physiological indicators to perform indicator correlation detection based on the target knowledge graph, and determining multiple physiological indicators to be fused that are strongly correlated among the multiple target physiological indicators.

[0106] In this embodiment, after the computer device 10 executes the above step 260 and determines that the actual assessment accuracy of the comprehensive physiological load assessment score does not meet the expected accuracy standard, the computer device 10 will search for a graph embedding model that matches the multiple physiological indicators from the multiple graph embedding models stored in itself, and then input the target knowledge graph that matches the multiple physiological indicators into the searched graph embedding model, so as to visualize the actual indicator correlation of the multiple target physiological indicators at the data level through the searched graph embedding model, so as to adaptively determine which target physiological indicators belong to physiological indicators with strong correlation based on the actual indicator correlation between the multiple target physiological indicators, and respectively use the various physiological indicators with strong correlation as a kind of physiological indicator to be fused, so as to achieve the effect of adaptive screening of strongly correlated indicators of the multiple physiological indicators actually involved in the physiological signal telemetry device within a specific time window.

[0107] Optionally, see Figure 6 , Figure 6 yes Figure 5 One of the flowcharts of the sub-steps included in step 270 of FIG. In this embodiment, the step of "calling a graph embedding model matching multiple target physiological indicators to perform indicator correlation detection based on the target knowledge graph" in step 270 may include sub-steps 271 and 272 to visualize the actual indicator correlation of the multiple target physiological indicators at the data level.

[0108] Sub-step 271, calling the graph embedding model to map the medical association relationships between the multiple target physiological indicators included in the target knowledge graph into the embedding space, and obtaining the actual embedding vectors of the multiple target physiological indicators in the embedding space.

[0109] Sub-step 272, calculating the vector correlation degree according to the actual embedding vectors of each of the multiple target physiological indicators in the embedding space, and obtaining the actual indicator correlation degree between the multiple target physiological indicators.

[0110] In this embodiment, the computer device 10 can determine the actual indicator correlation between the multiple target physiological indicators by calculating the cosine similarity and / or Euclidean distance between the actual embedding vectors of the multiple target physiological indicators. If the cosine similarity between the actual embedding vectors of any two target physiological indicators is greater, the actual indicator correlation between the two target physiological indicators is greater; if the Euclidean distance between the actual embedding vectors of any two target physiological indicators is smaller, the actual indicator correlation between the two target physiological indicators is greater.

[0111] Therefore, the present invention can effectively visualize the actual indicator correlation of the multiple target physiological indicators at the data level by executing the above sub-steps 271 and 272.

[0112] Optionally, see Figure 7 , Figure 7 yes Figure 5 2 is a flow chart of the sub-steps included in step 270. In this embodiment, the step of "determining multiple physiological indicators to be fused with strong correlation among multiple target physiological indicators" in step 270 may include sub-steps 273 and 274 to perform adaptive screening of strongly correlated indicators for the multiple target physiological indicators.

[0113] Sub-step 273 , according to the actual correlation between the multiple target physiological indicators, the multiple target physiological indicators are paired according to the preset correlation threshold to obtain multiple indicator pairing combinations.

[0114] Each indicator pairing combination consists of two target physiological indicators whose corresponding actual indicator correlation exceeds a preset correlation threshold.

[0115] Sub-step 274, performing physiological indicator occurrence statistics on the obtained multiple indicator pairing combinations, and taking all target physiological indicators involved in the multiple indicator pairing combinations as a physiological indicator to be fused.

[0116] The computer device 10 may use each target physiological indicator appearing in the plurality of indicator pairing combinations as a physiological indicator to be fused that has a strong correlation with other physiological indicators among the plurality of target physiological indicators.

[0117] Therefore, the present invention can perform adaptive screening of strongly correlated indicators for the multiple target physiological indicators by executing the above sub-steps 273 and 274.

[0118] Step 280: Determine the time series fusion weights of the various physiological indicators to be fused according to the actual indicator correlations between the various physiological indicators to be fused.

[0119] In this embodiment, after the computer device 10 selects a plurality of physiological indicators to be fused from the plurality of target physiological indicators, the timing fusion weights of the plurality of physiological indicators to be fused are calculated based on the pre-configured timing weight calculation strategy. The configuration principle of the timing weight calculation strategy may be, but is not limited to: "The greater the actual indicator correlation between each physiological indicator to be fused and other physiological indicators to be fused, the more directly the physiological indicator to be fused is related to the other physiological indicators to be fused, and the physiological indicator to be fused needs to maintain a greater timing fusion weight for the other physiological indicators to be fused", "The greater the number of indicator pairings (i.e., the number of indicator pairing combinations in which the actual indicator correlation of a certain physiological indicator to be fused exceeds the preset correlation threshold) among all physiological indicators to be fused, the greater the timing fusion weight that needs to be allocated to the physiological indicator to be fused", "The greater the average correlation between all actual indicator correlations corresponding to a certain physiological indicator to be fused, the greater the timing fusion weight that needs to be allocated to the physiological indicator to be fused", etc.

[0120] In one implementation of the present embodiment, in order to simplify the timing fusion weight allocation process, the computer device 10 can perform a correlation mean calculation on the actual indicator correlation between each physiological indicator to be fused and other physiological indicators to be fused among the multiple physiological indicators to be fused, and obtain the actual correlation mean of each physiological indicator to be fused, and then perform a proportion operation on the actual correlation mean of all physiological indicators to be fused, and obtain the proportion of the correlation mean of each physiological indicator to be fused relative to all physiological indicators to be fused, and directly use the correlation mean proportion of the physiological indicator to be fused as the timing fusion weight of the physiological indicator to be fused.

[0121] Step 290 , based on the respective time series fusion weights of the multiple physiological indicators to be fused, feature fusion is performed on the respective time series data of the multiple physiological indicators to be fused, so as to obtain the physiological signal time series fusion data of the target object within the target time window.

[0122] In this embodiment, after determining the respective time series fusion weights of the multiple physiological indicators to be fused, the computer device 10 will perform normalized physiological data extraction processing on the time series data of the multiple physiological indicators to be fused within the target time window, and then perform feature splicing and fusion based on the attention mechanism according to the respective time series fusion weights of the multiple physiological indicators to be fused and the normalized physiological data, to obtain the physiological signal time series fusion data of the target object within the target time window, thereby performing deep feature fusion of the indicator time series data for the multiple strongly correlated physiological indicators selected adaptively, to ensure that the physiological signal time series fusion data finally calculated can have sufficiently strong data reliability.

[0123] Step 2110 , directly outputting the physiological signal time series fusion data as the physiological data fusion result of multiple target physiological indicators at the target time window.

[0124] Therefore, the present invention can continue to execute the above steps 270 to 2110 on the basis of executing the above step 260. When it is determined that the data reliability of the calculated comprehensive physiological load assessment score is unqualified, based on the medical correlation between the various physiological indicators actually involved in the physiological signal telemetry equipment within a specific time window, the above-mentioned multiple physiological indicators are adaptively screened for strongly correlated indicators and deeply fused with the characteristics of the indicator data, to ensure that the final output physiological data fusion result has sufficiently strong data reliability, and can also accurately evaluate the actual physical condition of the monitored object, which is convenient for subsequent rhythm analysis operations.

[0125] In the embodiments provided by the present invention, it should be understood that the disclosed devices and methods can also be implemented in other ways. The device embodiments described above are merely schematic. For example, the flowcharts and block diagrams in the accompanying drawings show the possible architecture, functions and operations of the devices, methods and computer program products according to the embodiments of the present invention. In this regard, each box in the flowchart or block diagram can represent a module, a program segment or a part of a code, and a module, a program segment or a part of a code contains one or more executable instructions for implementing the specified logical functions. It should also be noted that in some alternative implementations, the functions marked in the box can also occur in a different order from the order marked in the accompanying drawings. For example, two consecutive boxes can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each box in the block diagram and / or the flowchart, and the combination of boxes in the block diagram and / or the flowchart, can be implemented with a dedicated hardware-based system that performs a specified function or action, or can be implemented with a combination of dedicated hardware and computer instructions.

[0126] In addition, the functional modules in each embodiment of the present invention can be integrated together to form an independent part, or each module can exist separately, or two or more modules can be integrated to form an independent part. If the various functions provided by the present invention are implemented in the form of software functional modules and sold or used as independent products, they can be stored in a storage medium. Based on such an understanding, the technical solution of the present invention is essentially or the part that contributes to the prior art or the part of the technical solution can be embodied in the form of a software product, and the computer software product is stored in a readable storage medium, including several instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method recorded in each embodiment of the present invention. The aforementioned readable storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk and other media that can store program codes.

[0127] The above are only various embodiments of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art can easily think of changes or substitutions within the technical scope disclosed by the present invention, which should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention should be based on the protection scope of the claims.

Claims

1. A multimodal physiological data fusion method, characterized in that: The multimodal physiological data fusion method comprises: Acquire the time series data of various target physiological indicators monitored by the physiological signal telemetry device for the target object within the target time window; Calling a medical knowledge base to construct a knowledge graph for the multiple target physiological indicators to obtain a target knowledge graph adapted to the multiple target physiological indicators; Calling a pre-stored physiological load influence evaluation model to perform influence weight prediction based on the target knowledge graph and the time series data of each of the multiple target physiological indicators, and obtaining the physiological load influence weight of each of the multiple target physiological indicators within the target time window; Performing a physiological load state assessment based on the time series data corresponding to each of the multiple target physiological indicators and the physiological load influence weights to obtain a comprehensive physiological load assessment score of the target object within the target time window; The comprehensive physiological load assessment score is output as a result of physiological data fusion of the multiple target physiological indicators at the target time window.

2. The multimodal physiological data fusion method according to claim 1, characterized in that: The step of obtaining the time series data of each of the multiple target physiological indicators monitored by the physiological signal telemetry device for the target object within the target time window includes: Acquire original physiological data corresponding to the multiple target physiological indicators respectively collected by the physiological signal telemetry device at different time points within the target time window; For each target physiological index, all original physiological data of the target physiological index within the target time window are dimensionally processed to obtain normalized physiological data corresponding to different time points of the target physiological index within the target time window; A time series is constructed for all original physiological data and all normalized physiological data of the target physiological index within the target time window to obtain the time series data of the target physiological index within the target time window.

3. The multimodal physiological data fusion method according to claim 1, characterized in that: The time series data of each target physiological indicator includes normalized physiological data corresponding to the target physiological indicator at different time points within the target time window, and the step of performing physiological load state assessment according to the time series data corresponding to each of the multiple target physiological indicators and the physiological load influence weight to obtain the comprehensive physiological load assessment score of the target object within the target time window includes: For each time point within the target time window, according to the normalized physiological data of each of the multiple target physiological indicators at the time point, the physiological load size caused by each of the multiple target physiological indicators to the target object at the time point is calculated, and the physiological load size and physiological load influence weight corresponding to each of the multiple target physiological indicators at the time point are weighted summed to obtain the comprehensive physiological load size of the target object at the time point; According to the time sequence of each time point in the target time window, an ascending time weight configuration is performed for each time point in the target time window to obtain an actual time weight of each time point in the target time window; A weighted sum operation is performed on the comprehensive physiological load magnitudes and actual time weights of all time points within the target time window to obtain the comprehensive physiological load assessment score of the target object within the target time window.

4. The multimodal physiological data fusion method according to any one of claims 1 to 3, characterized in that: The multimodal physiological data fusion method further includes: Performing an assessment accuracy check on the obtained comprehensive physiological load assessment score to determine whether the actual assessment accuracy of the comprehensive physiological load assessment score meets the expected accuracy standard; When it is determined that the actual assessment accuracy of the comprehensive physiological load assessment score reaches the expected accuracy standard, the step of outputting the comprehensive physiological load assessment score as the physiological data fusion result of the multiple target physiological indicators at the target time window is performed accordingly.

5. The multimodal physiological data fusion method according to claim 4, characterized in that: The step of performing an assessment accuracy check on the obtained comprehensive physiological load assessment score to determine whether the actual assessment accuracy of the comprehensive physiological load assessment score reaches the expected accuracy standard includes: Obtaining an actual comprehensive physiological load true value of the target object within the target time window, and calculating a load assessment error between the comprehensive physiological load assessment score and the actual comprehensive physiological load true value; Detecting whether the load assessment error exceeds a preset assessment error threshold; If it is detected that the load assessment error exceeds the preset assessment error threshold, it is determined that the actual assessment accuracy of the comprehensive physiological load assessment score does not meet the expected accuracy standard; otherwise, it is determined that the actual assessment accuracy of the comprehensive physiological load assessment score meets the expected accuracy standard.

6. The multimodal physiological data fusion method according to claim 4, characterized in that: The multimodal physiological data fusion method further includes: In the case where it is determined that the actual evaluation accuracy of the comprehensive physiological load evaluation score does not reach the expected accuracy standard, calling a graph embedding model matching the multiple target physiological indicators to perform indicator correlation detection based on the target knowledge graph, and determining multiple physiological indicators to be fused with strong correlation among the multiple target physiological indicators; Determine the time series fusion weights of the various physiological indicators to be fused according to the actual indicator correlations among the various physiological indicators to be fused; According to the respective time series fusion weights of the multiple physiological indicators to be fused, feature fusion is performed on the respective time series data of the multiple physiological indicators to be fused, so as to obtain the physiological signal time series fusion data of the target object within the target time window; The physiological signal time series fusion data is directly output as the physiological data fusion result of the multiple target physiological indicators at the target time window.

7. The multimodal physiological data fusion method according to claim 6, characterized in that: The step of calling a graph embedding model matching the multiple target physiological indicators to perform indicator correlation detection based on the target knowledge graph includes: Calling the graph embedding model to map the medical association relationships between the multiple target physiological indicators included in the target knowledge graph into the embedding space, and obtaining actual embedding vectors of each of the multiple target physiological indicators in the embedding space; Vector correlations are calculated according to actual embedding vectors of each of the multiple target physiological indicators in the embedding space to obtain actual indicator correlations between the multiple target physiological indicators.

8. The multimodal physiological data fusion method according to claim 7, characterized in that: The step of determining a plurality of physiological indicators to be fused that are strongly correlated among the plurality of target physiological indicators comprises: According to the actual indicator correlations between the multiple target physiological indicators, the multiple target physiological indicators are paired according to a preset correlation threshold to obtain multiple indicator pairing combinations, wherein each indicator pairing combination is composed of two target physiological indicators whose corresponding actual indicator correlations exceed the preset correlation threshold; Physiological indicator occurrence statistics are performed on the obtained multiple indicator pairing combinations, and all target physiological indicators involved in the multiple indicator pairing combinations are respectively used as a physiological indicator to be fused.

9. A computer device, characterized in that: It comprises a processor and a memory, wherein the memory stores a computer program executable by the processor, and the processor can execute the computer program to implement the multimodal physiological data fusion method according to any one of claims 1 to 8.

10. A readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a computer device, the multimodal physiological data fusion method described in any one of claims 1 to 8 is implemented.

Images