Real-time processing method and system for rapid identification and filing of emergency patients

Abstract

The embodiment of the invention provides a real-time processing method and system for rapid identification and filing of emergency patients. Wherein the intelligent medical fabric is worn on the skin surface of a burn patient, reflected light intensity distribution and microvascular pulsation waveforms are captured, and the light transmittance and the sampling frequency are automatically adjusted; the reflected light intensity distribution and the microvascular pulsation waveform are converted into combined parameters and matched, and a grading alarm signal is triggered when a skin tolerance dynamic interval is exceeded; packaging the graded alarm signal and the real-time physical sign data of the patient into a cross-hospital transmission data packet, and sending a data sharing request to a cooperative hospital; performing feature space alignment on the burnt skin allergy feature spectrum to generate a cross-mechanism allergy reaction prediction model; and automatically reconstructing an allergy monitoring file of the burnt patient during detection and transfer. According to the technical scheme provided by the embodiment of the invention, intelligent early warning and dynamic filing of the allergic reaction of the burn patient are realized, and the recognition accuracy and filing efficiency of the emergency patient in an emergency scene are remarkably improved.

Description

Technical Field

[0001] The embodiments of the present application relate to the field of medical information technology, and in particular to a real-time processing method and system for rapid identification and filing of emergency patients. Background Art

[0002] During the skin injury repair process of burn patients, scar hyperplasia, microvascular dysfunction and allergic reactions are common clinical challenges. It is necessary to dynamically monitor key physiological parameters such as epidermal oxygen permeability and tissue fluid migration rate in real time, and predict the allergic risk by combining multi-modal data. At the same time, when patients are transferred to other hospitals or receive cross-institutional medical treatment, it is necessary to realize the dynamic reconstruction of the allergic monitoring file and the secure sharing of data between heterogeneous medical systems to ensure the continuity of medical treatment and privacy compliance.

[0003] The current mainstream solutions are based on the collaborative analysis technology of wearable sensors and the cloud. For example, basic physiological signals such as the epidermal temperature and blood oxygen saturation of patients are collected through flexible electronic skin, and the allergic risk prediction is completed on the cloud using a deep learning model, and data interconnection between hospitals is realized through a standardized interface. Such systems usually adopt a static encryption protocol to ensure the security of data transmission and rely on a pre-trained model for allergic feature matching.

[0004] This solution has significant defects in the cross-institutional data collaboration scenario: firstly, the static encryption strategy is difficult to dynamically adjust the protection intensity according to the sensitivity of patient data, which is likely to lead to privacy leakage or over-desensitization and affect the diagnosis and treatment accuracy; secondly, the pre-trained model does not perform spatial alignment for the heterogeneous data features of different institutions, resulting in insufficient generalization ability of the allergic prediction model when applied across hospitals, and low accuracy in identifying emergency patients and filing efficiency. Summary of the Invention

[0005] The embodiments of the present application provide a real-time processing method and system for rapid identification and filing of emergency patients to solve the problems of the accuracy of identifying emergency patients and the filing efficiency in the prior art.

[0006] In a first aspect, the embodiments of the present application provide a real-time processing method for rapid identification and filing of emergency patients, including:

[0007] Wear an intelligent medical fabric on the skin surface of a burn patient, dynamically capture the reflected light intensity distribution and microvascular pulsation waveform on the skin surface through a flexible optoelectronic sensing array embedded in the intelligent medical fabric, and automatically adjust the light transmittance and sampling frequency of the flexible optoelectronic sensing array according to the degree of scar hyperplasia and epidermal temperature;

[0008] The reflected light intensity distribution and the microvascular pulsation waveform are converted into a combined parameter including epidermal oxygen permeability, tissue fluid migration rate, and local immune cell density after multi-channel noise suppression, and are matched with a pre-established burn skin allergy characteristic spectrum. When the matching result exceeds the preset skin tolerance dynamic range, the edge computing node is triggered to generate a graded alarm signal for the allergic reaction;

[0009] The edge computing node encapsulates the graded alarm signal and the real-time physical sign data of the patient into a cross-hospital transmission data packet, and sends a data sharing request to the collaborating hospital based on a privacy protection policy, and the privacy protection policy automatically adjusts the encryption intensity according to the patient identity sensitivity and data desensitization level;

[0010] After obtaining the authorization of the collaborating hospital, the burn skin allergy characteristic spectrum is aligned with the heterogeneous medical data of the collaborating hospital in the feature space to generate a cross-institutional allergic reaction prediction model, and the updated model parameters are synchronized to the edge computing node;

[0011] When it is detected that the burn patient is transferred to another hospital, the allergy monitoring file of the burn patient is automatically reconstructed according to the device compatibility protocol of the target hospital.

[0012] Optionally, the step of converting the reflected light intensity distribution and the microvascular pulsation waveform into a combined parameter including epidermal oxygen permeability, tissue fluid migration rate, and local immune cell density after multi-channel noise suppression, and matching with a pre-established burn skin allergy characteristic spectrum, and when the matching result exceeds the preset skin tolerance dynamic range, triggering the edge computing node to generate a graded alarm signal for the allergic reaction includes:

[0013] Perform dynamic baseline drift elimination on the multi-channel raw signals output by the flexible optoelectronic sensing array, establish an adaptive filter based on the light intensity fluctuation correlation between adjacent channels, remove the light intensity mutation artifacts generated by the limb movement of the burn patient, and generate a timestamped purified waveform sequence in combination with the reflected light intensity distribution and the microvascular pulsation waveform;

[0014] Perform multi-band decomposition on the purified waveform sequence through multi-channel noise suppression, extract the low-frequency band energy ratio for inverting epidermal oxygen permeability, extract the medium-frequency band phase gradient integral value to characterize the tissue fluid migration rate, and extract the high-frequency band waveform peak-valley variability and calculate the local immune cell density in combination with the multi-scale morphological decomposition result;

[0015] Fuse the epidermal oxygen permeability, the tissue fluid migration rate, and the local immune cell density into a combined parameter according to a preset ratio, construct a dynamic projection space in the pre-established burn skin allergy characteristic spectrum, and calculate the Mahalanobis distance between the combined parameter and the dynamic projection space through a sliding time window to generate an allergic reaction similarity vector;

[0016] When any dimension of the allergic reaction similarity vector exceeds the corresponding boundary of the preset skin tolerance dynamic range, the pre-stored alarm logic tree in the edge computing node is called according to the combination mode of the exceeded dimensions, and a graded alarm signal for the allergic reaction is generated by combining the autonomic nerve activation index of the burn patient.

[0017] Optionally, the method of fusing the epidermal oxygen permeability, the tissue fluid migration rate, and the local immune cell density into a combined parameter according to a preset ratio, constructing a dynamic projection space in a pre-established burn skin allergy feature spectrum, and calculating the Mahalanobis distance between the combined parameter and the dynamic projection space through a sliding time window to generate an allergic reaction similarity vector includes:

[0018] Inputting the time series of the epidermal oxygen permeability, the gradient change amount of the tissue fluid migration rate, and the spatial distribution pattern of the local immune cell density into a multi-dimensional tensor field, and generating a spatio-temporal coupled initial tensor structure according to the number of dimensions of the allergic atom space in the pre-established burn skin allergy feature spectrum, with each slice of the initial tensor structure aligned with the data block in the multi-dimensional tensor field;

[0019] Constructing a dynamic projection space in the burn skin allergy feature spectrum, and constructing an adaptive topological network with the allergic atom space as nodes in the dynamic projection space based on the multi-modal data stream of historical allergic reaction events in the burn skin allergy feature spectrum. The connection strength between nodes in the adaptive topological network is dynamically adjusted by the pathological correlation degree and time decay coefficient of the allergic atom space;

[0020] Adopting an incremental covariance matrix update mechanism, and calculating the Mahalanobis distance between the initial tensor structure and each node in the adaptive topological network frame by frame through a sliding time window. The update amount of the incremental covariance matrix is jointly controlled by the stability index of the microvascular pulsation waveform and the current sampling frequency of the flexible optoelectronic sensing array;

[0021] Inputting the reciprocal value of the Mahalanobis distance into a non-linear transformation function to generate an allergic reaction entropy value, and dynamically assigning weights to each dimension of the allergic reaction entropy value through a dynamic weight allocator embedded in the edge computing node. The parameters of the dynamic weight allocator are jointly adjusted by the metabolic rate and inflammatory factor concentration of the burn patient, and an allergic reaction similarity vector is output.

[0022] Optionally, the method of adopting an incremental covariance matrix update mechanism, calculating the Mahalanobis distance between the initial tensor structure and each node in the adaptive topological network frame by frame through a sliding time window, and the update amount of the incremental covariance matrix is jointly controlled by the stability index of the microvascular pulsation waveform and the current sampling frequency of the flexible optoelectronic sensing array includes:

[0023] Convert the stability index of the microvascular pulsation waveform into a waveform oscillation dispersion parameter, and generate a sampling frequency time element according to the current sampling frequency of the flexible optoelectronic sensing array. Multiply the waveform oscillation dispersion parameter by the sampling frequency time element to generate a covariance increment step adjustment coefficient;

[0024] Obtain the ratio of the trace norm of the historical covariance matrix to the spectral radius of the current data block through a sliding time window. Combine the covariance increment step adjustment coefficient, and use the update mechanism of the incremental covariance matrix to generate a sliding covariance basis. Divide the initial tensor structure along the time axis into overlapping sub-blocks and perform block-by-block orthogonal projection with the sliding covariance basis, and output the updated incremental covariance matrix. The update amount of the incremental covariance matrix is jointly controlled by the stability index of the microvascular pulsation waveform and the current sampling frequency of the flexible optoelectronic sensing array;

[0025] Bind a time decay sliding window in the adaptive topology network, and generate a Mahalanobis distance using the principal component direction of the updated incremental covariance matrix and the pathological correlation degree of the allergic atom space. Each component of the Mahalanobis distance is normalized in time scale by the sampling frequency time element;

[0026] Integrate the Mahalanobis distance in time series through the dynamic distance buffer of the edge computing node, adaptively adjust the interval length of the dynamic distance buffer according to the fluctuation period of the waveform oscillation dispersion parameter, and synchronize the Mahalanobis distance with the sliding time window.

[0027] Optionally, the step of binding a time decay sliding window in the adaptive topology network, generating a Mahalanobis distance using the principal component direction of the updated incremental covariance matrix and the pathological correlation degree of the allergic atom space, and normalizing each component of the Mahalanobis distance in time scale by the sampling frequency time element includes:

[0028] Bind a time decay sliding window in the adaptive topology network. Based on the pathological stage marker data in the burn skin allergy characteristic spectrum, associate the decay rate of the time decay sliding window with the acute inflammation index of the allergic atom space to generate a dynamic pathological decay weight, and the dynamic pathological decay weight is used to constrain the sliding step of the time decay sliding window;

[0029] Construct a hypersphere projection through the principal component direction of the updated incremental covariance matrix and the pathological correlation degree vector of the allergic atom space, and calculate the product of the tangential component modulus length and the normal component decay factor of the principal component direction in the hypersphere projection to generate a pathological correction Mahalanobis element;

[0030] Perform time-domain stretching transformation on the pathological correction Mahalanobis basis element using the sampling frequency time basis element, and dynamically adjust the stretching transformation smoothing factor through the stability index of the microvascular pulsation waveform to generate a standardized Mahalanobis distance component synchronized with the biological tissue metabolism cycle;

[0031] In the multi-channel fusion cache pool of the edge computing node, perform cross-channel correlation disambiguation on the standardized Mahalanobis distance component, adaptively adjust the channel bandwidth of the multi-channel fusion cache pool according to the change gradient of the dynamic pathological attenuation weight, and output a Mahalanobis distance matching the pathological evolution trend within the sliding time window to drive the non-linear transformation process of the allergic reaction entropy value.

[0032] Optionally, constructing a hypersphere projection through the principal component direction of the updated incremental covariance matrix and the pathological correlation degree vector of the allergic atom space, and calculating the product of the tangential component modulus and the normal component attenuation factor of the principal component direction in the hypersphere projection to generate a pathological correction Mahalanobis basis element, including:

[0033] Construct a dynamic rotation matrix based on the pathological correlation degree vector of the allergic atom space, project the principal component direction of the incremental covariance matrix into the hypersphere space corresponding to the dynamic rotation matrix, and generate a hypersphere projection trajectory carrying the pathological feature vector;

[0034] Decompose a tangential projection operator and a normal attenuation kernel function on the hypersphere projection trajectory. The modulus of the tangential projection operator is dynamically weighted by the cosine value of the angle between the principal component direction and the pathological correlation degree vector, and the normal attenuation kernel function is jointly calibrated by the historical change gradient of the epidermal oxygen permeability and the instantaneous fluctuation of the tissue fluid migration rate;

[0035] Input the modulus of the tangential projection operator into the normal attenuation kernel function, and perform multi-order smoothing processing through the pathological basis element optimization pool of the edge computing node. The smoothing order of the pathological basis element optimization pool is dynamically controlled by the stability index of the microvascular pulsation waveform and the update frequency of the dynamic rotation matrix;

[0036] Obtain the smoothed modulus-attenuation product based on the multi-order smoothing processing result, compensate the smoothed modulus-attenuation product with the metabolic gradient parameter of the burn patient for the dynamic viscosity coefficient of the tissue fluid, generate a set of Mahalanobis basis elements containing multi-dimensional pathological correction factors, and after eliminating the sensor drift noise through the basis element perturbation suppression factor of the edge computing node, output a pathological correction Mahalanobis basis element matching the pathological stage of the allergic atom space.

[0037] Optionally, when it is detected that the burn patient is transferred to another hospital, automatically reconstruct the allergy monitoring file of the burn patient according to the device compatibility protocol of the target hospital, including:

[0038] Capture the transfer trigger instruction of the burn patient in real time through the medical system interface, analyze the sensor data format constraints and encryption transmission rules in the device compatibility protocol of the target hospital, and generate a multimodal protocol parsing tree. Each branch node of the multimodal protocol parsing tree corresponds to an archive reconstruction mode;

[0039] According to the topological structure of the multimodal protocol parsing tree, rearrange the epidermal oxygen permeability time series, tissue fluid migration rate gradient value, and allergy reaction probability vector in the allergy monitoring file of the burn patient according to the device sampling dimension of the target hospital, and generate an archive topological structure that matches the device compatibility protocol of the target hospital. Each data unit in the archive topological structure is bound with a dynamic check code;

[0040] Based on the current burn stage mark and real-time allergy reaction entropy value of the burn patient, sort the dynamic check codes in the archive topological structure according to the metabolic urgency parameter, and divide them into high-priority data blocks and low-priority data streams. The high-priority data blocks include the stability index of the microvascular pulsation waveform and the pathological correction Mahalanobis primitive, and the low-priority data stream includes the historical sampling frequency time primitive of the intelligent medical fabric;

[0041] Through the dynamic verification token generator of the edge computing node, encapsulate the high-priority data block and the low-priority data stream into a multi-level verification data packet according to the encryption transmission rule of the target hospital. The dynamic verification token generator uses the real-time metabolic rate of the burn patient and the encryption strength of the device compatibility protocol to generate a two-way verification key. The two-way verification key performs timestamp anchoring interaction with the receiving terminal of the target hospital through the multi-band communication module of the intelligent medical fabric to complete the cross-protocol fusion archiving of allergy historical data and real-time monitoring data.

[0042] In a second aspect, an embodiment of the present application provides a real-time processing system for rapid identification and archiving of emergency patients, including:

[0043] A capture module, configured to wear an intelligent medical fabric on the skin surface of a burn patient, dynamically capture the reflected light intensity distribution and microvascular pulsation waveform on the skin surface through the flexible optoelectronic sensing array embedded in the intelligent medical fabric, and automatically adjust the light transmittance and sampling frequency of the flexible optoelectronic sensing array according to the degree of scar hyperplasia and epidermal temperature;

[0044] A conversion module for converting the reflected light intensity distribution and the microvascular pulsation waveform into a combined parameter including epidermal oxygen permeability, tissue fluid migration rate, and local immune cell density after multi-channel noise suppression, and matching it with a pre-established burn skin allergy characteristic spectrum. When the matching result exceeds a preset skin tolerance dynamic range, it triggers an edge computing node to generate a graded alarm signal for an allergic reaction;

[0045] An encapsulation module for encapsulating the graded alarm signal and the patient's real-time vital sign data into a cross-hospital transmission data packet through the edge computing node, and sending a data sharing request to a collaborating hospital based on a privacy protection policy, where the privacy protection policy automatically adjusts the encryption intensity according to the patient's identity sensitivity and data desensitization level;

[0046] A generation module for, after obtaining the authorization of the collaborating hospital, aligning the characteristic space of the burn skin allergy characteristic spectrum with the heterogeneous medical data of the collaborating hospital to generate a cross-institutional allergic reaction prediction model, and synchronizing the updated model parameters to the edge computing node;

[0047] A reconstruction module for automatically reconstructing the allergy monitoring file of the burn patient according to the device compatibility protocol of the target hospital when it is detected that the burn patient is transferred to another hospital.

[0048] In a third aspect, an embodiment of the present application provides a computing device, including a processing component and a storage component; the storage component stores one or more computer instructions; the one or more computer instructions are used to be called and executed by the processing component to implement a real-time processing method for rapid identification and file creation of emergency patients as described in the first aspect above.

[0049] In a fourth aspect, an embodiment of the present application provides a computer storage medium storing a computer program, where the computer program, when executed by a computer, implements a real-time processing method for rapid identification and file creation of emergency patients as described in the first aspect.

[0050] In an embodiment of the present application, a smart medical fabric is worn on the skin surface of a burn patient, and the reflected light intensity distribution and microvascular pulsation waveform on the skin surface are dynamically captured by a flexible photoelectric sensor array embedded in the smart medical fabric, and the transmittance and sampling frequency of the flexible photoelectric sensor array are automatically adjusted according to the degree of scar hyperplasia and the epidermal temperature; the reflected light intensity distribution and the microvascular pulsation waveform are converted into a combined parameter including epidermal oxygen permeability, tissue fluid migration rate and local immune cell density after multi-channel noise suppression, and matched with a pre-established burn skin allergy characteristic spectrum; when the matching result exceeds the preset skin tolerance dynamic range, the edge computing node is triggered to generate an allergic reaction The hierarchical alarm signal is generated; the hierarchical alarm signal and the patient's real-time vital sign data are encapsulated into a cross-hospital transmission data packet through the edge computing node, and a data sharing request is sent to the cooperative hospital based on the privacy protection strategy. The privacy protection strategy automatically adjusts the encryption strength according to the patient's identity sensitivity and the data desensitization level; after obtaining the authorization of the cooperative hospital, the burn skin allergy characteristic spectrum is aligned with the heterogeneous medical data of the cooperative hospital in feature space to generate a cross-institutional allergic reaction prediction model, and the updated model parameters are synchronized to the edge computing node; when the burn patient is detected to be transferred to another hospital, the allergy monitoring file of the burn patient is automatically reconstructed according to the equipment compatibility protocol of the target hospital. The file data is divided into high-priority physiological parameters and low-priority auxiliary information, and a two-way verification link is established with the smart medical fabric of the target hospital through the edge computing node to achieve seamless integration and archiving of the patient's allergy history data and real-time monitoring data.

[0051] The technical solution of this application has the following beneficial effects:

[0052] Real-time dynamic monitoring of skin condition is achieved, and the accuracy and stability of data collection are improved through adaptive adjustment of sensor parameters, especially to adapt to skin deformation and temperature fluctuation interference caused by scar hyperplasia; through multimodal parameter fusion and dynamic projection space matching, early identification and graded warning of allergic reactions are achieved, the false alarm rate is reduced, and the inflammatory response and tissue repair process are quantified at the same time; the security of sensitive medical data in transmission is ensured, and privacy protection and data availability are balanced through dynamic encryption strategies to meet the security and compliance requirements of different collaborating hospitals; data silos are broken, and the generalization ability of the model is optimized using multi-source heterogeneous data to improve the cross-institutional applicability and real-time nature of allergy prediction; the seamless integration of patients' allergy history and real-time data is ensured, and the equipment protocols of different hospitals are compatible to avoid data gaps or repeated monitoring due to transfer.

[0053] Furthermore, dynamic baseline drift elimination is performed on the multi-channel raw signals output by the flexible optoelectronic sensing array, and an adaptive filter is established by combining the correlation of light intensity fluctuations between adjacent channels to remove motion artifacts; low-frequency (epidermal oxygen permeability), intermediate-frequency (tissue fluid migration rate), and high-frequency (local immune cell density) characteristic parameters are extracted through multi-band decomposition, and a dynamic projection space is constructed after fusion according to a preset ratio; the allergic similarity vector is calculated based on the Mahalanobis distance, and a hierarchical warning signal is generated by combining the warning logic tree and the autonomic nerve activation index. Through dynamic signal purification and multi-band feature inversion, the quantization accuracy of epidermal oxygen permeability, tissue fluid metabolism, and immune status is significantly improved; the refined classification warning of allergic reactions (such as mild, moderate, and severe levels) is realized by using the dynamic projection space and the logic tree decision mechanism, and at the same time, the false positive rate caused by limb movement interference is reduced, and the warning response time is shortened to the second level.

[0054] These aspects or other aspects of the present application will be more clearly understood in the following description of the embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0055] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the following will briefly introduce the drawings required for the description of the embodiments or the prior art. Obviously, the drawings in the following description are some embodiments of the present application. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0056] Figure 1 FIG. shows a flowchart of a real-time processing method for rapid identification and filing of emergency patients provided by the present application;

[0057] Figure 2 FIG. shows a schematic structural diagram of a real-time processing system for rapid identification and filing of emergency patients provided by the present application;

[0058] Figure 3 FIG. shows a schematic structural diagram of a computing device provided by the present application. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0059] In order to enable those skilled in the art to better understand the solution of the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application.

[0060] In some of the processes described in the specification, claims, and the above-mentioned drawings of this application, a number of operations appear in a specific order. However, it should be clearly understood that these operations may not be executed in the order in which they appear herein or may be executed in parallel. The serial numbers of the operations, such as 101, 102, etc., are only used to distinguish the different operations, and the serial numbers themselves do not represent any order of execution. In addition, these processes may include more or fewer operations, and these operations may be executed in sequence or in parallel. It should be noted that the descriptions such as "first", "second", etc. in this article are used to distinguish different messages, devices, modules, etc., and do not represent a sequence, nor do they limit that "first" and "second" are of different types.

[0061] The research and development idea of this solution aims to achieve personalized monitoring and management of burn patients through an intelligent medical fabric system. First, a flexible optoelectronic sensing array is used to dynamically capture the optical and microvascular information on the skin surface, and the data acquisition parameters are automatically optimized according to the degree of scar hyperplasia and epidermal temperature to ensure high-precision monitoring. Then, a multi-channel noise suppression technology is adopted to convert the original data into comprehensive physiological parameters, and a hierarchical alarm signal is generated by matching with the allergy characteristic spectrum to timely detect potential allergic reactions. Then, the alarm signal and the real-time physical sign data are encrypted and packaged and sent to the collaborating hospital to ensure secure data sharing. After obtaining authorization, multi-source data is integrated to generate a cross-institutional allergy reaction prediction model to improve the prediction accuracy. Finally, when the patient is transferred to another hospital, the allergy monitoring file is automatically reconstructed, the key data is segmented and preferentially transmitted, and seamless data connection between the old and new hospitals is achieved through a two-way verification link to ensure the continuity and efficiency of treatment. This idea emphasizes the full-process optimization from precise data collection to secure data sharing and then to seamless data integration, and is committed to providing a comprehensive and reliable personalized medical solution.

[0062] Next, the technical solutions in the embodiments of the present application will be clearly and completely described in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative efforts belong to the scope of protection of the present application.

[0063] Figure 1 The following is a flowchart of a real-time processing method for rapid identification and filing of emergency patients provided for the embodiments of the present application. As Figure 1 shown, the method includes:

[0064] 101. Wear the intelligent medical fabric on the skin surface of the burn patient. Dynamically capture the reflected light intensity distribution and the microvascular pulsation waveform on the skin surface through the flexible optoelectronic sensing array embedded in the intelligent medical fabric, and automatically adjust the light transmittance and sampling frequency of the flexible optoelectronic sensing array according to the degree of scar hyperplasia and the epidermal temperature.

[0065] In this step, the flexible optoelectronic sensing array is a sensor system embedded in the intelligent medical fabric, which can dynamically capture the reflected light intensity distribution and the microvascular pulsation waveform on the skin surface.

[0066] The light transmittance adjustment mechanism automatically adjusts the light transmittance and sampling frequency of the flexible optoelectronic sensing array according to the degree of scar hyperplasia and the epidermal temperature of the patient to optimize the accuracy and comfort of data collection.

[0067] The reflected light intensity distribution is a parameter for evaluating the skin condition by detecting the change of the reflected light on the skin surface through an optical sensor.

[0068] The microvascular pulsation waveform is the pulsation signal of the skin microvessels captured by using the micro pressure sensing technology, which is used to monitor the local blood circulation condition. In the embodiment of the present application, first, the intelligent medical fabric monitors the reflected light intensity distribution on the surface of the burned skin in real time through the embedded flexible optoelectronic sensing array. These sensors use high-resolution CCD (Charge Coupled Device) or CMOS (Complementary Metal Oxide Semiconductor) image sensors, combined with Fourier transform infrared spectroscopy analysis technology, to accurately capture the change of the optical characteristics on the skin surface. Then, the microvascular pulsation waveform is captured by using a micro pressure sensor, which involves the application of the piezoelectric effect and bioelectric signal processing technology. Based on the collected data, an adaptive filtering algorithm and a machine learning model (such as Support Vector Machine SVM) are used to evaluate the degree of scar hyperplasia and the epidermal temperature, so as to dynamically adjust the light transmittance and sampling frequency. The final result is to optimize the monitoring of the burned skin condition by comprehensively analyzing these parameters and ensure the accuracy and comfort of data collection.

[0069] Suppose a burn patient needs continuous monitoring of the wound recovery situation. The doctor equips him with an intelligent medical fabric. This fabric starts to work immediately after the patient wears it. First, it captures the reflected light intensity distribution on the skin surface through the flexible optoelectronic sensing array. Specifically, the sensor records that the reflected light intensity of the burned area on the patient's left arm is 500 lux (lux), which indicates that the skin in this area is relatively dry and has mild scar hyperplasia. Then, the microvascular pulsation waveform is recorded by using a micro pressure sensor, and it is found that the microvascular pulsation frequency of the patient is 75 times per minute, slightly lower than the normal range (80 - 100 times per minute). Based on these data, the light transmittance is automatically adjusted to 60%, and the sampling frequency is set to once per second to more accurately monitor the wound recovery process.

[0070] 102. Convert the reflected light intensity distribution and the microvascular pulsation waveform into a combined parameter including epidermal oxygen permeability, tissue fluid migration rate, and local immune cell density through multi-channel noise suppression, and match it with a pre-established burn skin allergy characteristic spectrum. When the matching result exceeds the preset skin tolerance dynamic range, trigger the edge computing node to generate a graded warning signal for the allergic reaction;

[0071] In this step, the multi-channel noise suppression technology is a signal processing technology used to filter data noise from different sensors and improve data quality. It reduces interference by separating independent signal components.

[0072] The combined parameter includes epidermal oxygen permeability, tissue fluid migration rate, and local immune cell density, etc. These parameters comprehensively reflect the physiological state and pathological characteristics of the skin.

[0073] The burn skin allergy characteristic spectrum is a model constructed based on a large amount of clinical data, including characteristic patterns of various burn skin allergic reactions, and is used to identify and classify allergic reactions.

[0074] The graded warning signal is a warning signal of different levels generated when it is detected that the skin state of the patient exceeds the preset range, prompting medical staff to take corresponding measures.

[0075] In the embodiment of the present application, the data is first processed by multi-channel noise suppression, and the blind source separation algorithm (such as independent component analysis ICA) is used to eliminate unnecessary interference signals. Then, the original data is converted into a more easily analyzable combined parameter through a conversion algorithm, which involves applying principal component analysis (PCA) and non-negative matrix factorization (NMF) techniques. These parameters are matched with a pre-established allergy characteristic spectrum, which is trained based on a deep neural network (DNN) and covers various burn skin allergy characteristics. If it exceeds the preset range, the edge computing node is triggered to generate a graded warning signal. The whole process relies on efficient signal processing technology and advanced pattern recognition algorithms to ensure timely discovery of potential problems.

[0076] For example, continuing the above example, after the intelligent medical fabric has been working for a period of time, it starts to process the collected data. First, apply the blind source separation algorithm (such as independent component analysis ICA) to remove noise and separate the interference signals in the original data from the main signals. After processing, the epidermal oxygen permeability is calculated to be 15 mmHg (millimeters of mercury), the tissue fluid migration rate is 0.3 ml / min (milliliters per minute), and the local immune cell density is 200 cells / mm 2(Number of cells per square millimeter). These combined parameters are then compared with a pre-established allergy signature spectrum, and abnormal reactions are found in the patient's local skin. Specifically, the epidermal oxygen permeability is lower than the normal value (20 mmHg), and the tissue fluid migration rate is higher than the normal value (0.2 ml / min), indicating a possible early allergic reaction. Based on this, a graded warning is issued to alert medical staff to the possible allergic reaction.

[0077] 103. The edge computing node encapsulates the graded alarm signal and the patient's real-time vital sign data into a cross-hospital transmission data packet, and sends a data sharing request to the collaborating hospital based on a privacy protection strategy. The privacy protection strategy automatically adjusts the encryption strength according to the patient identity sensitivity and the data desensitization level.

[0078] In this step, the edge computing node, as a local data analysis center, is responsible for real-time processing and analysis of the collected medical data, and encapsulates the necessary information into a cross-hospital transmission data packet.

[0079] The privacy protection strategy automatically adjusts the encryption strength according to the patient identity sensitivity and the data desensitization level to ensure compliance with relevant regulations and protection of personal privacy during cross-institutional data sharing.

[0080] The cross-hospital transmission data packet packages the graded alarm signal and the patient's real-time vital sign data for secure transmission to the collaborating hospital for further analysis and processing.

[0081] The collaborating hospital cooperates with the original treating hospital to jointly participate in the patient's treatment process and provide additional medical resources and technical support.

[0082] In the embodiment of the present application, the edge computing node first encapsulates the necessary medical data, including vital sign data such as heartbeat and blood pressure, as well as the allergic reaction graded alarm information. The data packet is encrypted using homomorphic encryption technology, which allows computational operations to be performed without decrypting the data, ensuring data security. The privacy protection strategy is based on the theory of Differential Privacy and automatically adjusts the encryption strength according to the patient identity sensitivity and the data desensitization level. The encrypted data packet is then sent to the selected collaborating medical institution for further analysis and processing. The whole process ensures the security and privacy of the data.

[0083] For example, continuing with the previous example, after receiving the hierarchical alarm signal, relevant data of the patient (such as heart rate, blood pressure, etc.) and the warning level are automatically encrypted and sent to the collaborating hospital. First, necessary medical data are encapsulated, including physical sign data such as heart rate (72 beats per minute), blood pressure (120 / 80 mmHg), and allergic reaction classification alarm information (level: moderate). The data packet is encrypted using homomorphic encryption technology, and the encryption strength is automatically adjusted according to the sensitivity of the patient's identity and the data desensitization level. Specifically, since the patient's identity information is relatively sensitive, a high-level encryption strategy is adopted to ensure the security of personal information during data transmission. The encrypted data packet is sent to the collaborating hospital, and the hospital can immediately decrypt and analyze the data after receiving it to ensure the continuity and efficiency of treatment.

[0084] 104. After obtaining the authorization of the collaborating hospital, align the burn skin allergy characteristic spectrum with the heterogeneous medical data of the collaborating hospital in the feature space to generate a cross-institutional allergic reaction prediction model, and synchronize the updated model parameters to the edge computing node;

[0085] In this step, feature space alignment maps two datasets from different sources into the same dimensional space for easy comparison and integration, ensuring the consistency and comparability of the data.

[0086] Heterogeneous medical data are data from different medical institutions, including different formats, standards, and contents, which need to be uniformly processed.

[0087] The cross-institutional allergic reaction prediction model is a model established by combining data from multiple medical institutions to improve the accuracy of allergic reaction prediction.

[0088] Model parameter synchronization synchronizes the updated model parameters back to the edge computing node for subsequent use, ensuring the continuity and accuracy of monitoring.

[0089] In the embodiment of the present application, after obtaining the authorization of the collaborating hospital, align the allergy characteristic spectrum of the current patient with the heterogeneous data of the collaborating hospital in the feature space to create a more comprehensive allergic reaction prediction model. This process involves using generative adversarial networks (GANs) and variational autoencoders (VAEs) for feature extraction and alignment to ensure the consistency and comparability of the data. Next, ensemble learning algorithms such as random forest (Random Forest) and gradient boosting decision tree (GBDT) are used to generate a more accurate allergic reaction prediction model by combining multi-source data. The updated model parameters will be synchronized back to the edge computing node for subsequent use.

[0090] For example, continuing with the above example, the allergy prediction model is further trained using additional data provided by the cooperative hospital. The cooperative hospital provided detailed data on similar cases in the past five years, including parameters such as epidermal oxygen permeability, tissue fluid migration rate, and local immune cell density. First, the two sets of data are aligned in feature space using generative adversarial networks (GANs) and variational autoencoders (VAEs) to ensure data consistency and comparability. Specifically, the dataset of the original hospital (containing data from 100 patients) is aligned with the dataset of the cooperative hospital (containing data from 200 patients) to generate a comprehensive dataset containing data from 300 patients. Then, the random forest and gradient boosting decision tree (GBDT) algorithms are applied for model training, and a more accurate allergic reaction prediction model is finally generated. The updated model parameters are synchronized back to the edge computing node to ensure more accurate risk assessment in future monitoring.

[0091] 105. When it is detected that the burn patient is transferred to another hospital, the allergy monitoring file of the burn patient is automatically reconstructed according to the device compatibility protocol of the target hospital.

[0092] In this step, the device compatibility protocol is a set of standards that specifies how different medical devices exchange information to ensure data interoperability and consistency.

[0093] An allergy monitoring profile is a detailed record and management system designed to track and evaluate a patient's reactions to a variety of potential allergens.

[0094] The target hospital is the new treatment location for the patient after transfer and may have different equipment configuration and technical requirements.

[0095] In an embodiment of the present application, when the patient is transferred to another hospital, the equipment configuration of the target hospital is checked, and the allergy monitoring file is adjusted according to its compatibility protocol. This process involves the use of the HL7 (Health Level Seven) standard and the FHIR (Fast Healthcare Interoperability Resources) protocol to achieve seamless communication between devices. First, the device compatibility requirements of the target hospital are parsed, and then the existing data format and transmission protocol are adjusted using the adapter pattern to ensure data interoperability. In addition, the edge computing nodes are reconfigured to adapt to the equipment environment of the new hospital to ensure the continuity and accuracy of monitoring. After the allergy monitoring file is established, the file data can also be divided into high-priority physiological parameters and low-priority auxiliary information, and a two-way verification link can be established with the smart medical fabric of the target hospital through the edge computing node to achieve seamless integration and archiving of patient allergy history data and real-time monitoring data.

[0096] For example, continuing with the previous example, when a patient needs to be transferred to another hospital for treatment, the equipment situation of the new hospital is automatically checked. First, the HL7 standard and FHIR protocol requirements of the new hospital are parsed, and it is found that the monitoring equipment used in the new hospital supports different data formats and transmission protocols. Specifically, the equipment in the new hospital uses an XML format data transmission protocol, while the original hospital uses a JSON format. The adapter pattern is used to convert the existing data format into XML format, and the edge computing nodes are reconfigured to adapt to the equipment environment of the new hospital. In addition, the allergy monitoring file is adjusted according to the equipment compatibility requirements of the new hospital, and the light transmittance and sampling frequency are reset. For example, the light transmittance is adjusted to 70%, and the sampling frequency is set to once every two seconds to ensure the accuracy and efficiency of data collection. In this way, both the original hospital and the new hospital can seamlessly dock the patient's treatment records to ensure the continuity of treatment.

[0097] To further improve the accuracy and real-time performance of allergy reaction monitoring for burn patients, the research and development idea is to design an intelligent medical fabric system that dynamically captures the reflected light intensity distribution and microvascular pulsation waveform on the skin surface, and uses multi-channel noise suppression technology to convert the original signal into comprehensive physiological parameters, which are matched with a pre-established allergy feature spectrum to generate hierarchical alarm signals.

[0098] In some embodiments, the process of converting the reflected light intensity distribution and microvascular pulsation waveform into combined parameters through multi-channel noise suppression and generating hierarchical alarm signals includes:

[0099] 1021. Perform dynamic baseline drift elimination on the multi-channel original signals output by the flexible optoelectronic sensing array, establish an adaptive filter based on the light intensity fluctuation correlation between adjacent channels, remove the light intensity mutation artifacts generated by the limb movement of the burn patient, and generate a purified waveform sequence with timestamps in combination with the reflected light intensity distribution and the microvascular pulsation waveform;

[0100] In step 1021, dynamic baseline drift elimination is a technique for removing long-term trend changes in sensor signals to maintain signal stability.

[0101] The adaptive filter is a filter established based on the light intensity fluctuation correlation between adjacent channels, which can dynamically adjust the filter coefficients to remove artifacts.

[0102] The purified waveform sequence is the processed waveform data with timestamps, which removes artifacts and baseline drift and retains the real signal.

[0103] In the embodiments of the present application, first, the multi-channel raw signals output by the flexible optoelectronic sensing array are processed by dynamic baseline drift elimination. Specifically, based on the light intensity fluctuation correlation between adjacent channels, an adaptive filter is established, and the least mean square (LMS) algorithm is used to dynamically adjust the filter coefficients to remove the light intensity mutation artifacts generated by the patient's limb movement. Then, combined with the reflected light intensity distribution and the microvascular pulsation waveform, a purified waveform sequence with timestamps is generated. These purified waveform sequences not only eliminate the artifacts but also retain the key physiological information, providing a high-quality basis for further data analysis.

[0104] 1022. Perform multi-band decomposition on the purified waveform sequence through multi-channel noise suppression, extract the low-frequency band energy ratio for inverting the epidermal oxygen permeability, extract the medium-frequency band phase gradient integral value for characterizing the tissue fluid migration rate, and extract the high-frequency band waveform peak-valley variability and calculate the local immune cell density in combination with the multi-scale morphological decomposition result;

[0105] In step 1022, multi-band decomposition is a technique for decomposing a signal into different frequency bands to extract the energy or phase information of a specific frequency band.

[0106] Epidermal oxygen permeability is a parameter reflecting the oxygen permeation ability of the skin surface and is usually used to evaluate skin health status.

[0107] The tissue fluid migration rate characterizes the flow rate of tissue fluid in the skin and reflects the local blood circulation state.

[0108] Local immune cell density is an index describing the number of immune cells in a local skin area and is used to evaluate the inflammatory response.

[0109] In the embodiments of the present application, the purified waveform sequence is subjected to multi-band decomposition through multi-channel noise suppression. The low-frequency band energy ratio is used to invert the epidermal oxygen permeability, and the wavelet transform is used to extract the low-frequency band signal characteristics; the medium-frequency band phase gradient integral value is used to characterize the tissue fluid migration rate, and the Hilbert-Huang transform (HHT) is used to calculate the phase gradient; the high-frequency band waveform peak-valley variability and the multi-scale morphological decomposition result are combined to calculate the local immune cell density, and the morphological filtering method is used to extract the high-frequency band characteristics. Finally, these parameters are integrated into a combined parameter for generating the allergy reaction similarity vector in the subsequent process.

[0110] 1023. The epidermal oxygen permeability, the tissue fluid migration rate, and the local immune cell density are fused into a combined parameter according to a preset ratio, a dynamic projection space is constructed in a pre-established burn skin allergy feature spectrum, and the Mahalanobis distance between the combined parameter and the dynamic projection space is calculated through a sliding time window to generate an allergy reaction similarity vector.

[0111] In step 1023, the combined parameter includes a comprehensive index of multiple physiological parameters such as epidermal oxygen permeability, tissue fluid migration rate, and local immune cell density.

[0112] The dynamic projection space is a multi-dimensional space in a pre-established burn skin allergy feature spectrum, which is used to evaluate the similarity between the combined parameter and the allergy feature.

[0113] The Mahalanobis distance is a statistical distance metric that measures the similarity between two data sets and is commonly used in pattern recognition and classification tasks.

[0114] In the embodiment of the present application, the epidermal oxygen permeability, the tissue fluid migration rate, and the local immune cell density are fused into a combined parameter according to a preset ratio, and a dynamic projection space is constructed in a pre-established burn skin allergy feature spectrum. The Mahalanobis distance between the combined parameter and the dynamic projection space is calculated through a sliding time window to generate an allergy reaction similarity vector. This process involves using principal component analysis (PCA) to perform dimensionality reduction processing on the combined parameter to ensure its effective representation in a high-dimensional space. In addition, a support vector machine (SVM) is also applied for classification to further improve the accuracy of similarity calculation. Finally, the similarity vector reflects the similarity between the current skin state and the known allergy features, providing a basis for generating subsequent graded alarm signals.

[0115] 1024. When any dimension of the allergy reaction similarity vector exceeds the corresponding boundary of a preset dynamic skin tolerance interval, an alarm logic tree pre-stored in the edge computing node is called according to the combined pattern of the exceeded dimension, and a graded alarm signal for the allergy reaction is generated in combination with the autonomic nerve activation index of the burn patient.

[0116] In step 1024, the allergy reaction similarity vector is a multi-dimensional vector representing the similarity between the current skin state and the known allergy features.

[0117] The edge computing node is a local data analysis center, which is responsible for real-time processing and analysis of the collected medical data and generating alarm signals.

[0118] The autonomic nerve activation index is an index reflecting the activity level of the patient's autonomic nervous system and is used to evaluate the severity of the allergy reaction.

[0119] In the embodiments of the present application, when any dimension of the allergic reaction similarity vector exceeds the corresponding boundary of the preset skin tolerance dynamic range, the pre-stored alarm logic tree in the edge computing node is called according to the combination mode of the exceeded dimensions. A graded alarm signal for the allergic reaction is generated in combination with the autonomic nerve activation index of the burn patient. Specifically, the edge computing node determines the alarm level according to the specific values of the exceeded dimensions in the similarity vector and combines the real-time physical sign data of the patient (such as heart rate, blood pressure, etc.), and applies the Fuzzy Logic algorithm to generate the corresponding alarm signal. In addition, the Reinforcement Learning algorithm is also used to optimize the decision-making path of the alarm logic tree to ensure the accuracy and timeliness of the alarm signal.

[0120] The following is a specific example:

[0121] Suppose a burn patient needs to continuously monitor the wound recovery situation, and the doctor equips him with an intelligent medical fabric. This fabric starts working immediately after the patient wears it. First, it captures the reflected light intensity distribution on the skin surface through a flexible optoelectronic sensing array, and records that the reflected light intensity of the burned area on the patient's left arm is 500 lux (lux). Then, it uses a micro pressure sensor to record the microvascular pulsation waveform and finds that the microvascular pulsation frequency of the patient is 75 beats per minute. First, the dynamic baseline drift of the multi-channel original signal is eliminated to generate a purified waveform sequence with time stamps. Then, the purified waveform sequence is decomposed into multiple frequency bands through multi-channel noise suppression, and the epidermal oxygen permeability is extracted as 15 mmHg (millimeters of mercury), the tissue fluid migration rate is 0.3 ml / min (milliliters per minute), and the local immune cell density is 200 cells / mm 2 (number of cells per square millimeter). These combined parameters are then compared with the pre-established allergy characteristic spectrum, and the Mahalanobis distance is calculated to be 0.85, exceeding the preset skin tolerance dynamic range (0.6). Finally, a graded warning signal for a moderate allergic reaction is issued based on this, and in combination with the patient's autonomic nerve activation index (heart rate 85 beats per minute, blood pressure 130 / 85 mmHg), the alarm level is generated as "moderate" to remind medical staff to pay attention to the possible allergic reaction.

[0122] This set of intelligent medical fabric system and its supporting process realize the personalized monitoring and management of burn patients by dynamically capturing the optical and microvascular information on the skin surface and using multi-channel noise suppression technology to convert the original signal into comprehensive physiological parameters. It can not only accurately detect and warn potential allergic reactions, but also automatically reconstruct the allergy monitoring file when the patient is transferred to another hospital to ensure the seamless connection of data and the continuity of treatment. By introducing a variety of advanced signal processing, machine learning, and pattern recognition algorithms, it significantly improves the nursing quality and safety of burn patients and provides important technical support for the future personalized medical field.

[0123] In order to further improve the accuracy and real-time performance of allergy reaction monitoring for burn patients, the research and development idea is to design an intelligent medical fabric system to optimize the data processing ability through a multi-dimensional tensor field and an adaptive topological network. The specific steps include: First, input the time series of epidermal oxygen permeability, the gradient change of tissue fluid migration rate, and the spatial distribution pattern of local immune cell density into the multi-dimensional tensor field; then, construct a dynamic projection space in the burn skin allergy feature spectrum and build an adaptive topological network based on historical allergy reaction events; then, use an incremental covariance matrix update mechanism to calculate the Mahalanobis distance frame by frame; finally, input the reciprocal value of the Mahalanobis distance into a non-linear transformation function to generate an allergy reaction entropy value, and output a similarity vector through a dynamic weight allocator.

[0124] In some embodiments, the process of fusing the epidermal oxygen permeability, tissue fluid migration rate, and local immune cell density into a combined parameter and constructing a dynamic projection space in the allergy feature spectrum includes:

[0125] 1031. Input the time series of the epidermal oxygen permeability, the gradient change of the tissue fluid migration rate, and the spatial distribution pattern of the local immune cell density into the multi-dimensional tensor field, and generate a spatio-temporal coupled initial tensor structure according to the number of dimensions of the allergy atom space in the pre-established burn skin allergy feature spectrum. Each slice of the initial tensor structure is aligned with the data block in the multi-dimensional tensor field;

[0126] In step 1031, the multi-dimensional tensor field is a technology for representing high-dimensional data structures, which can process time series, gradient changes, and spatial distribution patterns simultaneously.

[0127] The spatio-temporal coupled initial tensor structure is a multi-dimensional data structure generated based on epidermal oxygen permeability, tissue fluid migration rate, and local immune cell density, and each slice is aligned with the data block in the multi-dimensional tensor field.

[0128] The allergy atom space is the basic unit in the pre-established burn skin allergy feature spectrum, which is used to describe the spatial dimension of specific allergy features.

[0129] In the embodiments of the present application, first, the time series of epidermal oxygen permeability, the gradient change amount of tissue fluid migration rate, and the spatial distribution pattern of local immune cell density are input into a multi-dimensional tensor field. Using tensor decomposition techniques such as CANDECOMP / PARAFAC (CP) decomposition or Tucker decomposition, these data are converted into a high-dimensional tensor structure. According to the number of dimensions of the allergic atom space in the pre-established burn skin allergy characteristic spectrum, an initial tensor structure of spatio-temporal coupling is generated. Specifically, key features are extracted through singular value decomposition (SVD) and principal component analysis (PCA), and it is ensured that each slice is aligned with the data block in the multi-dimensional tensor field. This step not only eliminates noise but also retains the key information of the data, providing a high-quality basis for further data analysis.

[0130] 1032. Construct a dynamic projection space in the burn skin allergy characteristic spectrum. Based on the multi-modal data stream of historical allergic reaction events in the burn skin allergy characteristic spectrum, construct an adaptive topological network with the allergic atom space as nodes in the dynamic projection space. The connection strength between nodes in the adaptive topological network is dynamically adjusted by the pathological correlation degree and time decay coefficient of the allergic atom space;

[0131] In step 1032, the dynamic projection space is a multi-dimensional space for evaluating the similarity between combined parameters and allergy characteristics.

[0132] The adaptive topological network is a network structure constructed based on historical allergic reaction events, and the connection strength between nodes is dynamically adjusted by the pathological correlation degree and time decay coefficient.

[0133] The pathological correlation degree is an index reflecting the pathological correlation between different allergic atom spaces.

[0134] The time decay coefficient describes the degree to which the correlation of allergy characteristics gradually weakens over time.

[0135] In the embodiments of the present application, a dynamic projection space is constructed in the burn skin allergy characteristic spectrum. Based on the multi-modal data stream of historical allergic reaction events, an adaptive topological network with the allergic atom space as nodes is constructed in the dynamic projection space. The connection strength between nodes is dynamically adjusted by the pathological correlation degree and time decay coefficient. Specifically, a graph neural network (Graph Neural Networks, GNN) is used to construct and optimize this network. By calculating the similarity matrix between nodes and applying the PageRank algorithm to adjust the node weights. In addition, a time series model (such as LSTM or GRU) is also adopted to capture the time decay coefficient to ensure that the relationship between nodes can accurately reflect the actual pathological situation.

[0136] 1033. An incremental covariance matrix update mechanism is adopted to calculate the Mahalanobis distance between the initial tensor structure and each node in the adaptive topological network frame by frame through a sliding time window. The update amount of the incremental covariance matrix is jointly controlled by the stability index of the microvascular pulsation waveform and the current sampling frequency of the flexible optoelectronic sensing array.

[0137] In step 1033, the incremental covariance matrix update mechanism is a technique for real-time updating of the covariance matrix, ensuring that the model can adapt to new data inputs.

[0138] The sliding time window is a technique for processing time series data, calculating the changes in data frame by frame by moving the window.

[0139] The stability index of the microvascular pulsation waveform is a parameter reflecting the stability of the microvascular pulsation waveform, used to control the update amount of the covariance matrix.

[0140] In the embodiment of the present application, an incremental covariance matrix update mechanism is adopted to calculate the Mahalanobis distance between the initial tensor structure and each node in the adaptive topological network frame by frame through a sliding time window. The update amount of the incremental covariance matrix is jointly controlled by the stability index of the microvascular pulsation waveform and the current sampling frequency of the flexible optoelectronic sensing array. Specifically, the Kalman Filter is used for real-time updating of the incremental covariance matrix, and the update process is further optimized by the Recursive Least Squares (RLS). The Wavelet Transform is also applied to perform frequency domain analysis on the microvascular pulsation waveform to extract the stability index.

[0141] 1034. The reciprocal value of the Mahalanobis distance is input into a non-linear transformation function to generate an anaphylactic reaction entropy value. Each dimension of the anaphylactic reaction entropy value is dynamically weighted by the dynamic weight allocator embedded in the edge computing node. The parameters of the dynamic weight allocator are jointly adjusted by the metabolic rate and inflammatory factor concentration of the burn patient, and an anaphylactic reaction similarity vector is output.

[0142] In step 1034, the non-linear transformation function is a function for converting linear data into a non-linear form, often used for generating entropy values.

[0143] The dynamic weight allocator is a mechanism for dynamically adjusting the weights of each dimension, ensuring that the model can flexibly respond to different data inputs.

[0144] The anaphylactic reaction entropy value is an entropy value reflecting the complexity of the anaphylactic reaction, used to evaluate the possibility of an anaphylactic reaction.

[0145] In the embodiments of the present application, the reciprocal value of the Mahalanobis distance is input into a non-linear transformation function to generate an allergic reaction entropy value, and each dimension is dynamically weighted by a dynamic weight allocator embedded in the edge computing node. The parameters of the dynamic weight allocator are jointly adjusted by the metabolic rate and the concentration of inflammatory factors of the burn patient. Specifically, the fuzzy logic algorithm is used for dynamic weight allocation, and the particle swarm optimization (PSO) algorithm is combined to optimize the weight allocation strategy. The non-linear transformation function adopts the sigmoid function or the ReLU function to convert the reciprocal value of the Mahalanobis distance into an entropy value. Finally, an allergic reaction similarity vector is output, providing a comprehensive and accurate allergic reaction assessment result.

[0146] The following is a specific example:

[0147] Suppose a patient with a deep second-degree burn area of 25% wears intelligent medical fabric. The flexible optoelectronic sensing array detects that the time series of epidermal oxygen permeability in the scar area of the left forearm is 12.8, 14.2, 11.5 mmHg (sampled every 2 hours), the gradient change of the tissue fluid migration rate is 0.4 ml / min (the difference from the wound edge to the center), and the spatial distribution pattern of the local immune cell density shows that the density in the central area is 180 cells / mm 2 and the density in the edge area is 220 cells / mm 2 . The multi-dimensional tensor field integrates the above data into an initial tensor structure with spatio-temporal coupling, where each slice corresponds to a three-dimensional data block of oxygen permeability, migration rate, and immune cell distribution at different time points in the scar area, and is aligned with three allergic atom spaces in the pre-established burn skin allergy feature spectrum (corresponding to the inflammatory phase, proliferation phase, and remodeling phase respectively).

[0148] In the dynamic projection space, based on multimodal data streams of historical allergic reaction events (such as sudden drops in oxygen permeability and sudden surges in immune cell density in previous patient scar infection cases), an adaptive topological network with allergic atom spaces as nodes is constructed. The connection strength between nodes is dynamically adjusted according to the pathological correlation degree (such as the correlation weight between the inflammatory phase and the proliferative phase is 0.7) and the time decay coefficient (such as the correlation degree decreases by 5% every 24 hours). The incremental covariance matrix update mechanism calculates the Mahalanobis distance between the initial tensor structure and the network nodes frame by frame through a sliding time window (window length 10 minutes). The update amount of the covariance matrix is jointly controlled by the stability index of the microvascular pulsation waveform (judged as stable when the fluctuation amplitude ≤ 0.1 Lux) and the current sampling frequency (10 Hz). When the Mahalanobis distance reaches 1.3 within a certain time window (the preset tolerance interval is 0.6 - 1.0), a non-linear transformation function is triggered to generate the allergic reaction entropy value (0.75). The dynamic weight allocator weights the entropy value to 0.82 according to the patient's real-time metabolic rate (1800 kcal / day) and the concentration of inflammatory factors (IL-6 is 50 pg / ml, TNF-α is 30 pg / ml), and outputs the allergic reaction similarity vector (0.45, 0.82, 0.33). The second dimension (corresponding to the allergic atom space in the proliferative phase) exceeds the standard, generating a "high allergic risk" warning signal, which is synchronously pushed to the medical staff terminal and encrypted and transmitted to the collaborating hospital.

[0149] The technical solutions in steps 1031 - 1034 integrate the spatio-temporal distribution characteristics of epidermal oxygen permeability, tissue fluid migration rate, and immune cell density through a multi-dimensional tensor field. Combined with the pathological correlation analysis of the adaptive topological network in the dynamic projection space, the recognition accuracy of allergic patterns is significantly improved; the incremental covariance matrix update mechanism realizes the dynamic matching of real-time data and the historical model through the joint regulation of the microvascular waveform stability and the sampling frequency, reducing the error rate of Mahalanobis distance calculation to less than 5%; finally, through non-linear transformation and dynamic weight allocation, combined with multi-dimensional regulation of the metabolic rate and the concentration of inflammatory factors, the allergic reaction similarity vector is output, realizing the refined classification from "mild warning" to "high emergency".

[0150] In order to further improve the monitoring accuracy and real-time performance of allergic reactions in burn patients, the research and development idea is to design an intelligent medical fabric system. Through the incremental covariance matrix update mechanism, combined with the stability index of the microvascular pulsation waveform and the current sampling frequency of the flexible optoelectronic sensing array, the Mahalanobis distance between the initial tensor structure and each node in the adaptive topological network is calculated frame by frame. In some embodiments, the process of using the incremental covariance matrix update mechanism to calculate the Mahalanobis distance between the initial tensor structure and each node in the adaptive topological network frame by frame through a sliding time window includes:

[0151] 201. Convert the stability index of the microvascular pulsation waveform into a waveform oscillation dispersion parameter, and at the same time generate a sampling frequency time element according to the current sampling frequency of the flexible optoelectronic sensing array. Multiply the waveform oscillation dispersion parameter by the sampling frequency time element to generate a covariance increment step adjustment coefficient;

[0152] In step 201, the waveform oscillation dispersion parameter is a quantitative index reflecting the stability of the microvascular pulsation waveform, and is used to measure the fluctuation degree of the waveform.

[0153] The sampling frequency time element is a time unit generated according to the current sampling frequency of the flexible optoelectronic sensing array, and is used to standardize the time scale.

[0154] The covariance increment step adjustment coefficient is an adjustment coefficient generated by combining the waveform oscillation dispersion parameter and the sampling frequency time element, and is used to control the update speed of the covariance matrix.

[0155] In the embodiment of the present application, first, convert the stability index of the microvascular pulsation waveform into a waveform oscillation dispersion parameter. Specifically, use the fast Fourier transform (FFT) to analyze the frequency domain characteristics of the waveform, and extract its oscillation dispersion by calculating the energy distribution of the spectrum. This process can effectively capture the high-frequency and low-frequency components in the waveform, thereby quantifying the stability of the waveform. At the same time, generate a sampling frequency time element according to the current sampling frequency of the flexible optoelectronic sensing array. Multiply the waveform oscillation dispersion parameter by the sampling frequency time element to generate a covariance increment step adjustment coefficient, which is used to dynamically adjust the update step of the covariance matrix to ensure that the update speed is increased when the data fluctuates greatly and the update frequency is reduced when the data is relatively stable.

[0156] 202. Obtain the ratio of the trace norm of the historical covariance matrix to the spectral radius of the current data block through a sliding time window. Combine the covariance increment step adjustment coefficient, and use the update mechanism of the incremental covariance matrix to generate a sliding covariance basis. Divide the initial tensor structure along the time axis into overlapping sub-blocks and perform block-by-block orthogonal projection with the sliding covariance basis, and output the updated incremental covariance matrix. The update amount of the incremental covariance matrix is jointly controlled by the stability index of the microvascular pulsation waveform and the current sampling frequency of the flexible optoelectronic sensing array;

[0157] In step 202, the sliding time window is a technique for processing time series data, and calculates the change of data frame by frame by moving the window.

[0158] The trace norm is the sum of the absolute values of all eigenvalues of the matrix, and is used to measure the size of the matrix.

[0159] The spectral radius is the modulus of the largest eigenvalue of the matrix, and is used to measure the stability of the matrix.

[0160] The sliding covariance basis is a basis matrix generated based on the historical covariance matrix and the current data block, and is used for orthogonal projection.

[0161] In the embodiments of the present application, the ratio of the trace norm of the historical covariance matrix to the spectral radius of the current data block is obtained through a sliding time window. Specifically, singular value decomposition (SVD) is used to decompose the historical covariance matrix to calculate its trace norm; at the same time, the power iteration method is used to calculate the spectral radius of the current data block. Combining with the covariance increment step adjustment coefficient, an update mechanism of the incremental covariance matrix is adopted to generate the sliding covariance basis. The initial tensor structure is divided into overlapping sub-blocks along the time axis and orthogonally projected block by block with the sliding covariance basis to output the updated incremental covariance matrix. In this process, the recursive least squares (RLS) method is also applied to further optimize the update process of the incremental covariance matrix to ensure rapid response to new data inputs.

[0162] 203. Bind a time-decaying sliding window in the adaptive topology network, and generate a Mahalanobis distance by using the principal component direction of the updated incremental covariance matrix and the pathological correlation degree of the allergic atom space. Each component of the Mahalanobis distance is normalized in terms of time scale through the sampling frequency time primitive;

[0163] In step 203, the time-decaying sliding window is a technique for processing time series data, and the window length is dynamically adjusted through a time decay factor.

[0164] The principal component direction is the direction of the main eigenvector of the incremental covariance matrix and is used to represent the main change trend of the data.

[0165] The pathological correlation degree is an index reflecting the pathological correlation between different allergic atom spaces.

[0166] In the embodiments of the present application, a time-decaying sliding window is bound in the adaptive topology network, and a Mahalanobis distance is generated by using the principal component direction of the updated incremental covariance matrix and the pathological correlation degree of the allergic atom space. Specifically, principal component analysis (PCA) is used to extract the principal component direction of the incremental covariance matrix, and the Mahalanobis distance is generated in combination with the pathological correlation degree of the allergic atom space. Each component is normalized in terms of time scale through the sampling frequency time primitive to ensure the comparability of data under different time scales. In addition, the Bayesian optimization algorithm is also applied to dynamically adjust the length of the time-decaying sliding window to adapt to different data fluctuation situations.

[0167] 204. Integrate the Mahalanobis distance in a time series through the dynamic distance buffer of the edge computing node, adaptively adjust the interval length of the dynamic distance buffer according to the fluctuation period of the waveform oscillation dispersion parameter, and synchronize the Mahalanobis distance with the sliding time window.

[0168] In step 204, the dynamic distance buffer is a buffer area in the edge computing node, which is used to store and process the time series data of the Mahalanobis distance.

[0169] The fluctuation period is the periodic change characteristic of the waveform oscillation dispersion parameter, which is used to adaptively adjust the interval length of the dynamic distance buffer.

[0170] The time series integration is to integrate the time series data of the Mahalanobis distance to provide a more stable evaluation result.

[0171] In the embodiment of the present application, the Mahalanobis distance is integrated in a time series through the dynamic distance buffer of the edge computing node. First, the autoregressive integrated moving average model (ARIMA) is used to analyze the fluctuation period of the waveform oscillation dispersion parameter, and the interval length of the dynamic distance buffer is adaptively adjusted accordingly. Specifically, by analyzing the time series data of the Mahalanobis distance, its periodic fluctuation pattern is identified, and the length of the buffer is adjusted according to these patterns to ensure the validity and consistency of the data. Then, the Mahalanobis distance is synchronized with the sliding time window to ensure that the data at different time points can be accurately matched. Finally, the Mahalanobis distance integrated in the time series is output, providing a comprehensive and accurate evaluation result of the allergic reaction.

[0172] The following is a specific example:

[0173] Suppose the flexible optoelectronic sensor array worn by a patient with deep second-degree burns detects the microvascular pulsation waveform in the scar area of the left shoulder. Its stability index analyzes the frequency domain energy distribution through the fast Fourier transform (FFT), and the waveform oscillation dispersion parameter is extracted as 0.45 (dimensionless, the dispersion threshold <0.3 is the stable state). At the same time, the sampling frequency time base element of 0.1 second is generated according to the current sampling frequency of 10 Hz. After multiplying the two, the covariance increment step size adjustment coefficient of 4.5 is generated to dynamically control the covariance matrix update speed - when the waveform oscillation is severe (such as the dispersion >0.4), the covariance update step size is automatically expanded to 6.0, and the response speed is increased by 50%.

[0174] A sliding time window (window length: 5 minutes) extracts a trace norm of 15.2 (reflecting the historical data fluctuation intensity) from the historical covariance matrix, and the spectral radius of the current data block is 3.8 (characterizing the current data stability). The ratio of the two, 4.0, and the adjustment coefficient of 4.5 are jointly used to generate a sliding covariance basis. The initial tensor structure (including the epidermal oxygen permeability, tissue fluid migration rate, and immune cell density data in the scar area for 24 hours) is divided into overlapping sub-blocks along the time axis (each block is 2 hours long, with an overlap rate of 30%). The basis is aligned with the sub-blocks through block-by-block orthogonal projection, and an updated incremental covariance matrix is output (the standard deviation of the eigenvalues is reduced from the initial 2.1 to 1.3). The update amount is jointly regulated by the waveform dispersion parameter of 0.45 and the sampling frequency of 10 Hz.

[0175] In the adaptive topological network, a time-decaying sliding window is bound (the decay coefficient decreases by 5% per hour). Using the principal component direction of the updated covariance matrix (the variance contribution rate of the first principal component is 68%) and the pathological correlation degree in the allergic atom space (the atomic correlation weight in the inflammation period is 0.8), a Mahalanobis distance of 1.25 is generated. Each component is normalized on the time scale through the sampling frequency time element of 0.1 second (for example, the distance component is adjusted from 1.25×0.1 second to 0.125 second-1). The dynamic distance buffer of the edge computing node adaptively adjusts the interval length to 25 minutes according to the fluctuation period of the waveform oscillation dispersion (the detected period is 20 minutes), and synchronously processes the Mahalanobis distance with the sliding time window.

[0176] Steps 201 to 204, through the joint regulation of the waveform oscillation dispersion parameter and the sampling frequency time element, enable the update speed of the covariance matrix to dynamically adapt to the microvascular fluctuation intensity, shortening the data response delay; the sliding covariance basis and the orthogonal projection mechanism effectively suppress noise interference, improving the eigenvalue stability; the time-scale normalization and the dynamic buffer adjustment solve the scale difference problem of multi-time-dimensional data fusion, reducing the false alarm rate of Mahalanobis distance calculation.

[0177] To further improve the monitoring accuracy and real-time performance of the allergic reactions of burn patients, the research and development idea is to design an intelligent medical fabric system. Through the time-decaying sliding window in the adaptive topological network, combining the principal component direction of the updated incremental covariance matrix and the pathological correlation degree in the allergic atom space to generate the Mahalanobis distance, and performing time-domain stretching transformation and cross-channel correlation disambiguation. In some embodiments, the process of binding a time-decaying sliding window in the adaptive topological network and using the principal component direction of the updated incremental covariance matrix and the pathological correlation degree in the allergic atom space to generate the Mahalanobis distance includes:

[0178] 301. Bind a time-decaying sliding window in the adaptive topological network. Based on the pathological stage marker data in the burn skin allergy feature spectrum, associate the decay rate of the time-decaying sliding window with the acute inflammation index in the allergy atom space to generate a dynamic pathological decay weight, which is used to constrain the sliding step of the time-decaying sliding window.

[0179] In step 301, the time-decaying sliding window is a technique for processing time series data, and the window length is dynamically adjusted by a time decay factor.

[0180] The dynamic pathological decay weight is a weight generated based on the pathological stage marker data in the burn skin allergy feature spectrum and is used to constrain the sliding step of the time-decaying sliding window.

[0181] The acute inflammation index is a quantitative index reflecting the degree of acute inflammation of burn skin.

[0182] In the embodiment of the present application, first, bind a time-decaying sliding window in the adaptive topological network. Based on the pathological stage marker data in the burn skin allergy feature spectrum, associate the decay rate of the time-decaying sliding window with the acute inflammation index in the allergy atom space to generate a dynamic pathological decay weight. Specifically, use the Bayesian optimization algorithm to dynamically adjust the decay rate of the time-decaying sliding window and generate a dynamic pathological decay weight according to the acute inflammation index. This weight is used to constrain the sliding step of the time-decaying sliding window to ensure that the window can be flexibly adjusted under different pathological stages to adapt to different data fluctuation situations. To further improve the flexibility and adaptability of the model, a Kalman Filter is also applied to update the dynamic pathological decay weight in real time to ensure that it can quickly respond to new data inputs.

[0183] 302. Construct a hypersphere projection through the principal component direction of the updated incremental covariance matrix and the pathological correlation vector in the allergy atom space, and calculate the product of the modulus length of the tangential component of the principal component direction in the hypersphere projection and the decay factor of the normal component to generate a pathological correction Mahalanobis primitive.

[0184] In step 302, the hypersphere projection is a technique for projecting high-dimensional data onto a hypersphere and is used to calculate the modulus length of the tangential component and the decay factor of the normal component.

[0185] The pathological correction Mahalanobis primitive is a standardized primitive generated by combining the principal component direction and the pathological correlation and is used for subsequent distance calculation.

[0186] The modulus length of the tangential component is the modulus length of the tangential component of the principal component direction in the hypersphere projection.

[0187] The normal component attenuation factor is the attenuation factor of the normal component in the projection of the principal component direction on the hypersphere.

[0188] In the embodiments of the present application, a hypersphere projection is constructed by the principal component direction of the updated incremental covariance matrix and the pathological correlation degree vector of the allergic atom space, the product of the modulus length of the tangential component and the normal component attenuation factor in the hypersphere projection of the principal component direction is calculated, and a pathological corrected Mahalanobis primitive is generated. Specifically, first, the principal component analysis (PCA) is used to extract the principal component direction of the incremental covariance matrix, and then it is combined with the pathological correlation degree vector to construct a hypersphere projection. Next, the product of the modulus length of the tangential component and the normal component attenuation factor is calculated to generate a pathological corrected Mahalanobis primitive. To improve the calculation accuracy, the Locally Linear Embedding (LLE) algorithm is also applied to perform dimensionality reduction on high-dimensional data to ensure the accuracy of the hypersphere projection.

[0189] 303. Perform a time-domain stretching transformation on the pathological corrected Mahalanobis primitive using the sampling frequency time primitive, and dynamically adjust the stretching transformation smoothing factor through the stability index of the microvascular pulsation waveform to generate a standardized Mahalanobis distance component synchronized with the biological tissue metabolism cycle;

[0190] In step 303, the time-domain stretching transformation is a technique for adjusting the time scale, and time-domain expansion or compression is performed through the stretching transformation smoothing factor.

[0191] The standardized Mahalanobis distance component is the Mahalanobis distance component after being processed by the time-domain stretching transformation, and is synchronized with the biological tissue metabolism cycle.

[0192] The stretching transformation smoothing factor is a parameter for dynamically adjusting the time-domain stretching transformation, ensuring the comparability of data at different time scales.

[0193] In the embodiments of the present application, a time-domain stretching transformation is performed on the pathological corrected Mahalanobis primitive using the sampling frequency time primitive, and the stretching transformation smoothing factor is dynamically adjusted through the stability index of the microvascular pulsation waveform to generate a standardized Mahalanobis distance component synchronized with the biological tissue metabolism cycle. Specifically, the wavelet transform is used to perform a time-domain stretching transformation on the pathological corrected Mahalanobis primitive, and the stretching transformation smoothing factor is dynamically adjusted according to the stability index of the microvascular pulsation waveform. To ensure the comparability of data at different time scales, the empirical mode decomposition (EMD) technique is also applied to further refine the effect of the time-domain stretching transformation. This process ensures the comparability of data at different time scales and improves the stability and reliability of the solution.

[0194] 304. In the multi-channel fusion cache pool of the edge computing node, perform cross-channel correlation disambiguation on the standardized Mahalanobis distance components, adaptively adjust the channel bandwidth of the multi-channel fusion cache pool according to the change gradient of the dynamic pathological attenuation weight, and output the Mahalanobis distance that matches the pathological evolution trend within the sliding time window to drive the non-linear transformation process of the allergic reaction entropy value.

[0195] In step 304, the multi-channel fusion cache pool is a cache area in the edge computing node, which is used to store and process data of the standardized Mahalanobis distance components.

[0196] Cross-channel correlation disambiguation performs correlation analysis on multi-channel data, eliminates redundant information, and improves data quality.

[0197] The channel bandwidth is the channel width adaptively adjusted according to the change gradient of the dynamic pathological attenuation weight, which is used to optimize data transmission efficiency.

[0198] In the embodiment of the present application, in the multi-channel fusion cache pool of the edge computing node, perform cross-channel correlation disambiguation on the standardized Mahalanobis distance components, adaptively adjust the channel bandwidth of the multi-channel fusion cache pool according to the change gradient of the dynamic pathological attenuation weight, and output the Mahalanobis distance that matches the pathological evolution trend within the sliding time window to drive the non-linear transformation process of the allergic reaction entropy value. Specifically, first use the Recursive Least Squares (RLS) method to perform cross-channel correlation disambiguation on the standardized Mahalanobis distance components to identify and eliminate redundant information. To further improve data quality, the Independent Component Analysis (ICA) technology is also applied to further decompose the multi-channel data to ensure the independence and purity of the data. Then, adaptively adjust the channel bandwidth of the multi-channel fusion cache pool according to the change gradient of the dynamic pathological attenuation weight to optimize data transmission efficiency. Finally, output the Mahalanobis distance that matches the pathological evolution trend within the sliding time window.

[0199] The following is a specific example:

[0200] Suppose the smart medical fabric worn by a patient with deep second-degree burns detects that the acute inflammation index of the scar area on the right thigh is 4.2 (the threshold < 3.0 is normal). The time-decaying sliding window correlates this index based on the pathological stage marker data (inflammatory phase, proliferative phase, remodeling phase) in the burn skin allergy feature spectrum, generating a dynamic pathological decay weight of 0.68 (weight range 0 - 1). When the microvascular pulsation waveform stability index in the scar area drops to 0.25 (stable threshold > 0.3), the decay rate is dynamically adjusted from 5% per hour to 8%, and the sliding step length is shortened from 30 minutes to 20 minutes, ensuring that the window length (initially 60 minutes) shrinks rapidly with the deterioration of inflammation and accurately captures the pathological evolution trend.

[0201] The updated principal component direction of the incremental covariance matrix (the contribution rate of the first principal component variance is 72%) and the pathological correlation degree vector of the allergy atom space (the atom correlation degree in the inflammatory phase is 0.85) construct a hypersphere projection. Calculate the product of the tangential component modulus 0.92 and the normal component decay factor 0.75 to generate a pathological correction Mahalanobis primitive 0.69. The primitive is subjected to a time-domain stretching transformation through a sampling frequency time primitive of 0.1 second, and the stretching transformation smoothing factor is dynamically adjusted (from 0.8 to 0.6) in combination with the microvascular waveform stability index of 0.25, and a standardized Mahalanobis distance component 1.38 (preset tolerance threshold 1.0) synchronized with the tissue metabolic cycle (the detected metabolic cycle is 25 minutes) is output.

[0202] In the multi-channel fusion cache pool of the edge computing node, cross-channel correlation disambiguation eliminates redundant data (the redundancy is reduced by 40%). According to the change gradient of the dynamic pathological decay weight (the weight drops by 0.05 per hour), the channel bandwidth is adaptively adjusted from 10 MHz to 15 MHz, and finally a Mahalanobis distance 1.42 matching the pathological trend within the sliding time window is output, driving the non-linear transformation of the allergy reaction entropy value to 0.78 (the threshold > 0.6 triggers an alarm).

[0203] Steps 301 - 304 improve the window adjustment response speed through the dynamic association between the acute inflammation index and the time-decaying sliding window; the combination of the hypersphere projection and the standardized Mahalanobis distance component solves the problem of scale differences in high-dimensional data and reduces the false alarm rate; the bandwidth adaptive adjustment and cross-channel disambiguation of the multi-channel fusion cache pool achieve real-time tracking of the pathological evolution trend and improve the model sensitivity.

[0204] In order to further improve the accuracy and real-time performance of allergy reaction monitoring for burn patients, the research and development idea is to design an intelligent medical fabric system. A hypersphere projection is constructed through the principal component direction of the updated incremental covariance matrix and the pathological correlation vector of the allergy atom space, and the product of the tangential component modulus and the normal component attenuation factor is calculated to generate a pathological correction Mahalanobis primitive. In some embodiments, the process of constructing a hypersphere projection through the principal component direction of the updated incremental covariance matrix and the pathological correlation vector of the allergy atom space, calculating the product of the tangential component modulus and the normal component attenuation factor of the principal component direction in the hypersphere projection, and generating a pathological correction Mahalanobis primitive includes:

[0205] 401. Construct a dynamic rotation matrix based on the pathological correlation vector of the allergy atom space, project the principal component direction of the incremental covariance matrix onto the hypersphere space corresponding to the dynamic rotation matrix, and generate a hypersphere projection trajectory carrying a pathological feature vector;

[0206] In step 401, the dynamic rotation matrix is a rotation matrix constructed based on the pathological correlation vector of the allergy atom space, and is used to project the principal component direction of the incremental covariance matrix into the hypersphere space.

[0207] The hypersphere space is the space where high-dimensional data is projected onto a hypersphere, and is used to represent the geometric characteristics of the data.

[0208] The hypersphere projection trajectory carrying a pathological feature vector is a data trajectory carrying a pathological feature after being projected into the hypersphere space.

[0209] In the embodiments of the present application, first, a dynamic rotation matrix is constructed based on the pathological correlation vector of the allergy atom space. Specifically, the quaternion rotation technology is used to project the principal component direction of the incremental covariance matrix onto the hypersphere space corresponding to the dynamic rotation matrix, and a hypersphere projection trajectory carrying a pathological feature vector is generated. In order to ensure the accuracy of the projection, the Lagrange Interpolation method is also applied to preprocess the pathological feature vector to reduce noise interference. This process provides a more accurate hypersphere projection trajectory.

[0210] 402. Decompose a tangential projection operator and a normal attenuation kernel function on the hypersphere projection trajectory. The modulus of the tangential projection operator is dynamically weighted by the cosine value of the angle between the principal component direction and the pathological correlation vector, and the normal attenuation kernel function is jointly calibrated by the historical change gradient of the epidermal oxygen permeability and the instantaneous fluctuation of the tissue fluid migration rate;

[0211] In step 402, the tangential projection operator is the tangential component decomposed on the hypersphere projection trajectory, and is used to represent the main change trend of the data.

[0212] The normal attenuation kernel function is the normal component decomposed on the hypersphere projection trajectory and is used to represent the attenuation characteristics of data.

[0213] The cosine value of the included angle is the cosine value of the included angle between the principal component direction and the pathological correlation vector, and is used to dynamically weight the modulus length of the tangential projection operator.

[0214] In the embodiment of the present application, a tangential projection operator and a normal attenuation kernel function are decomposed on the hypersphere projection trajectory. Specifically, the singular value decomposition (SVD) technique is used to decompose the tangential projection operator and the normal attenuation kernel function on the hypersphere projection trajectory. The modulus length of the tangential projection operator is dynamically weighted by the cosine value of the included angle between the principal component direction and the pathological correlation vector, and the normal attenuation kernel function is jointly calibrated by the historical change gradient of the epidermal oxygen permeability and the instantaneous fluctuation of the tissue fluid migration rate. To improve the calculation accuracy, a radial basis function (RBF) is also applied to fit the normal attenuation kernel function to ensure that it can accurately reflect the attenuation characteristics of the data.

[0215] 403. Input the modulus length of the tangential projection operator into the normal attenuation kernel function, and perform multi-order smoothing processing through the pathological primitive optimization pool of the edge computing node. The smoothing order of the pathological primitive optimization pool is dynamically controlled by the stability index of the microvascular pulsation waveform and the update frequency of the dynamic rotation matrix;

[0216] In step 403, the pathological primitive optimization pool is a cache area in the edge computing node and is used to store and process data of pathological primitives.

[0217] The multi-order smoothing processing is to perform multiple smoothing processes on the pathological primitives to eliminate noise and improve data quality.

[0218] The smoothing order is the number of smoothing times dynamically controlled according to the stability index of the microvascular pulsation waveform and the update frequency of the dynamic rotation matrix.

[0219] In the embodiments of the present application, the modulus of the tangential projection operator is input into the normal attenuation kernel function, and multi-order smoothing processing is performed through the pathological primitive optimization pool of the edge computing node. Specifically, the Kalman Filter is used to perform preliminary smoothing processing on the modulus of the tangential projection operator, and then Gaussian Filtering is applied for further multi-order smoothing processing. The smoothing order of the pathological primitive optimization pool is dynamically controlled by the stability index of the microvascular pulsation waveform and the update frequency of the dynamic rotation matrix to ensure the comparability of data at different time scales. To further improve the data quality, the AutoRegressive Moving Average (ARMA) model is also applied to perform time series analysis on the smoothed data to ensure that it can accurately reflect the pathological evolution trend. Finally, the result of multi-order smoothing processing is output.

[0220] 404. Obtain the smoothed modulus-attenuation product based on the result of multi-order smoothing processing, compensate the smoothed modulus-attenuation product with the metabolic gradient parameter of the burn patient to generate a set of Mahalanobis primitives containing multi-dimensional pathological correction factors. After eliminating the sensor drift noise through the primitive perturbation suppression factor of the edge computing node, the pathological correction Mahalanobis primitives matching the pathological stage of the allergic atom space are output.

[0221] In step 404, the dynamic viscosity coefficient of tissue fluid is a coefficient reflecting the flow characteristics of tissue fluid and is used to compensate the smoothed modulus-attenuation product.

[0222] The primitive perturbation suppression factor is a factor in the edge computing node and is used to eliminate the sensor drift noise.

[0223] The pathological correction Mahalanobis primitive is a Mahalanobis primitive containing multi-dimensional pathological correction factors and is used for subsequent distance calculation.

[0224] In the embodiments of the present application, the smoothed modulus-attenuation product is obtained based on the result of multi-order smoothing processing, and the dynamic viscosity coefficient of tissue fluid is used to compensate the smoothed modulus-attenuation product with the metabolic gradient parameter of the burn patient. Specifically, the Nonlinear Regression technology is used to fit the metabolic gradient parameter to generate the dynamic viscosity coefficient of tissue fluid, and it is applied to the smoothed modulus-attenuation product to generate a set of Mahalanobis primitives containing multi-dimensional pathological correction factors. To eliminate the sensor drift noise, the primitive perturbation suppression factor is also applied to further process the data through an Adaptive Filter to ensure that it can accurately reflect the changes in the pathological stage. Finally, the pathological correction Mahalanobis primitives matching the pathological stage of the allergic atom space are output.

[0225] The following is a specific example:

[0226] Suppose that the intelligent medical fabric worn by a patient with deep second-degree burns detects that the historical change gradient of epidermal oxygen permeability in the scar area of the right lower leg is 15.2, 17.8, 14.5 mmHg (sampled every 3 hours), the instantaneous fluctuation range of the tissue fluid migration rate is from 0.5 ml / min to 0.2 ml / min, and the microvascular pulsation waveform stability index is 0.3 (stable threshold > 0.4). Based on the pathological correlation vector of the allergic atom space (inflammatory phase correlation weight 0.7, proliferation phase 0.6), the dynamic rotation matrix projects the principal component direction of the incremental covariance matrix (the variance contribution rate of the first principal component is 75%) into the hypersphere space to generate a hypersphere projection trajectory carrying the pathological feature vector. The tangential projection operator modulus length of 1.2 (dynamically weighted by the cosine value of 0.85 of the angle between the principal component direction and the pathological correlation vector) and the normal attenuation kernel function of 0.6 (jointly calibrated by the oxygen permeability gradient and the migration rate fluctuation) are decomposed from the trajectory.

[0227] According to the microvascular waveform stability index of 0.3 and the dynamic rotation matrix update frequency of 10 Hz, the pathological primitive optimization pool of the edge computing node selects third-order smoothing processing to denoise the modulus-attenuation product (initial value 0.72) and outputs the smoothed product 0.68. Combining with the patient's metabolic gradient parameter (basal metabolic rate 1800 kcal / day) to generate a tissue fluid dynamic viscosity coefficient compensation factor of 1.15, and the compensated modulus-attenuation product is corrected to 0.78. The primitive perturbation suppression factor eliminates the sensor drift noise (the noise amplitude is reduced by 40%), and finally outputs the pathological corrected Mahalanobis primitive of 0.82, which matches the pathological stage of the proliferation phase in the allergic atom space.

[0228] Steps 401-404 effectively capture the correlation of pathological features in the spatial and temporal dimensions through the combination of the dynamic rotation matrix and the hypersphere projection trajectory; the joint calibration of the tangential and normal components solves the problems of data drift and noise interference in traditional methods; the viscosity coefficient compensation based on metabolic parameters and the perturbation suppression factor significantly improve the pathological matching accuracy of the Mahalanobis primitive.

[0229] In order to further improve the seamless connection of the allergy monitoring files and the treatment continuity for burn patients during transfer, the research and development idea is to design an intelligent medical system. By capturing the transfer trigger instruction in real time, parsing the device compatibility protocol of the target hospital, and generating a multi-modal protocol parsing tree; rearranging the patient's data according to the parsing tree to generate an archive topology structure compatible with the target hospital's devices; sorting the data based on the current burn stage marker and the real-time allergy reaction entropy value; generating multi-level verification data packets through edge computing nodes to complete cross-protocol fusion for file creation. In some embodiments, the process of automatically reconstructing the allergy monitoring file of a burn patient according to the device compatibility protocol of the target hospital when it is detected that the burn patient is transferred includes:

[0230] 1051. Capture the transfer trigger instruction of the burn patient in real time through the medical system interface, parse the sensor data format constraints and encryption transmission rules in the device compatibility protocol of the target hospital, and generate a multi-modal protocol parsing tree. Each branch node of the multi-modal protocol parsing tree corresponds to an archive reconstruction mode;

[0231] In step 1051, the transfer trigger instruction is a signal indicating that the patient is about to be transferred captured by the medical system interface.

[0232] The device compatibility protocol is the data format constraints and encryption transmission rules of the target hospital's devices.

[0233] The multi-modal protocol parsing tree is a tree structure generated by parsing the device compatibility protocol of the target hospital, and each branch node corresponds to an archive reconstruction mode.

[0234] In the embodiments of the present application, the transfer trigger instruction of the burn patient is captured in real time through the medical system interface, the sensor data format constraints and encryption transmission rules in the device compatibility protocol of the target hospital are parsed, and a multi-modal protocol parsing tree is generated. Specifically, natural language processing (NLP) technology is used to parse the device compatibility protocol document provided by the target hospital and convert it into structured data. Then, the decision tree algorithm is applied to generate a multi-modal protocol parsing tree, and each branch node corresponds to an archive reconstruction mode. Finally, the generated multi-modal protocol parsing tree provides a basic framework for subsequent data reorganization.

[0235] 1052. According to the topological structure of the multi-modal protocol parsing tree, rearrange the epidermal oxygen permeability time series, tissue fluid migration rate gradient value, and allergy reaction probability vector in the allergy monitoring file of the burn patient according to the device sampling dimension of the target hospital to generate an archive topology structure matching the device compatibility protocol of the target hospital. Each data unit in the archive topology structure is bound with a dynamic verification code;

[0236] In step 1052, the file topology structure is a structure generated by rearranging patient data according to the sampling dimensions of the target hospital's devices, and each data unit is bound with a dynamic verification code.

[0237] The dynamic verification code is a verification code bound to each data unit, which is used to ensure data integrity and accuracy.

[0238] In the embodiment of the present application, according to the topology structure of the multimodal protocol parsing tree, the epidermal oxygen permeability time series, the tissue fluid migration rate gradient value, and the allergic reaction probability vector in the allergy monitoring file of the burn patient are rearranged according to the sampling dimensions of the target hospital's devices to generate a file topology structure that matches the compatibility protocol of the target hospital's devices. Each data unit is bound with a dynamic verification code. Specifically, first, a graph neural network (GNN) is used to analyze the topology structure of the multimodal protocol parsing tree to determine the optimal data reorganization scheme. Then, the fast Fourier transform (FFT) is applied to perform frequency-domain analysis on the data to ensure the integrity of the data after rearrangement. In addition, a hash function is also used to generate a dynamic verification code for each data unit to ensure the accuracy and integrity of the data. The finally generated file topology structure not only meets the device requirements of the target hospital but also ensures high-quality data transmission.

[0239] 1053. Based on the current burn stage marker and the real-time allergic reaction entropy value of the burn patient, sort the dynamic verification codes in the file topology structure according to the metabolic urgency parameter, and divide them into high-priority data blocks and low-priority data streams. The high-priority data blocks include the stability index of the microvascular pulsation waveform and the pathological correction Markov primitive, and the low-priority data stream includes the historical sampling frequency time primitive of the intelligent medical fabric;

[0240] In step 1053, the metabolic urgency parameter is a quantitative index reflecting the current metabolic state of the burn patient.

[0241] The high-priority data block is a data block that includes the stability index of the microvascular pulsation waveform and the pathological correction Markov primitive.

[0242] The low-priority data stream is a data stream that includes the historical sampling frequency time primitive of the intelligent medical fabric.

[0243] In the embodiments of the present application, based on the current burn stage label of the burn patient and the real-time allergy reaction entropy value, the dynamic check codes in the file topology structure are sorted according to the metabolic urgency parameter and divided into high-priority data blocks and low-priority data streams. Specifically, first, a Support Vector Machine (SVM) is used to analyze the current burn stage label of the burn patient and the real-time allergy reaction entropy value to calculate the metabolic urgency parameter. Then, the dynamic check codes are sorted according to the metabolic urgency parameter, and the high-priority data blocks (including the stability index of the microvascular pulsation waveform and the pathological correction Mahalanobis primitive) and the low-priority data streams (including the historical sampling frequency time primitive of the intelligent medical fabric) are separated. To ensure the efficiency of data transmission, the Priority Queue algorithm is also applied to optimize the data chunking and transmission order.

[0244] 1054. Through the dynamic verification token generator of the edge computing node, the high-priority data block and the low-priority data stream are encapsulated into a multi-level verification data packet according to the encryption transmission rules of the target hospital. The dynamic verification token generator generates a two-way verification key by using the real-time metabolic rate of the burn patient and the encryption strength of the device compatibility protocol. The two-way verification key performs timestamp anchoring interaction with the receiving terminal of the target hospital through the multi-band communication module of the intelligent medical fabric, completing the cross-protocol fusion and archiving of the allergy history data and the real-time monitoring data.

[0245] In step 1054, the dynamic verification token generator is a module in the edge computing node for generating a two-way verification key.

[0246] The two-way verification key is a key generated by using the real-time metabolic rate of the patient and the encryption strength of the device compatibility protocol.

[0247] The timestamp anchoring interaction is a time synchronization mechanism between the intelligent medical fabric and the receiving terminal of the target hospital.

[0248] In the embodiments of the present application, through the dynamic verification token generator of the edge computing node, high-priority data blocks and low-priority data streams are encapsulated into multi-level verification data packets according to the encryption transmission rules of the target hospital. Specifically, a two-way verification key is generated by using the real-time metabolic rate of burn patients and the encryption strength of the device compatibility protocol. Then, the Elliptic Curve Cryptography (ECC) algorithm is used to encrypt and encapsulate the high-priority data blocks and low-priority data streams. The two-way verification key is interacted with the receiving terminal of the target hospital through the multi-band communication module of the intelligent medical fabric for timestamp anchoring to ensure the security and integrity of data transmission. To further improve the reliability of data transmission, the Redundancy Coding technology is also applied to increase the fault tolerance of data transmission. Finally, the generated multi-level verification data packets complete the cross-protocol fusion archiving of allergy history data and real-time monitoring data, ensuring the seamless connection of data and the continuity of treatment.

[0249] The following is a specific example:

[0250] Suppose a patient with 60% total body burn area needs to be urgently transferred from a county-level hospital to a provincial burn center. The intelligent medical fabric captures the transfer trigger instruction in real time and analyzes the sensor data format constraints in the device compatibility protocol of the target hospital (such as requiring JSON format and AES-256 encryption standard). The multi-modal protocol parsing tree automatically generates three branch nodes: the first corresponds to the HDF5 format reconstruction mode of the epidermal oxygen permeability time series (sampling interval of 2 minutes); the second matches the storage rules of the time series database for the tissue fluid migration rate gradient value (fluctuating between 0.2 - 0.8 ml / min); the third adapts to the encrypted field nested structure of the allergy reaction probability vector (0.45, 0.82).

[0251] The file topology structure rearranges the epidermal oxygen permeability data (12.8, 14.5, 11.9 mmHg) within 72 hours of the patient according to the device sampling dimension of the target hospital (average value every 5 minutes) to 12.9, 13.7, 12.3 mmHg, and each data unit is bound with a SHA-256 dynamic verification code (such as 0x3a7d...f9c2). According to the current burn stage of the patient (acute inflammation stage) and the real-time allergy reaction entropy value of 0.75, the dynamic verification code is divided into high-priority data blocks (including the microvascular pulsation waveform stability index of 0.3 and the pathological correction Mahalanobis primitive of 0.82) and low-priority data streams (historical sampling frequency time primitive of 0.1 second) according to the metabolic urgency parameter.

[0252] The dynamic verification token generator of the edge computing node uses the patient's real-time metabolic rate (2000kcal / day) and the encryption strength of the target hospital (key length 256 bits) to generate a two-way verification key (such as ECC curve secp256r1), and encapsulates the high-priority data block into a multi-level verification data packet with a timestamp (2025-02-27T14:30:00Z). Through the multi-band communication module (5G+LoRa dual channel) and the target hospital receiving terminal, the timestamp anchoring is completed, and the cross-protocol fusion of allergy history data (72-hour epidermal oxygen permeability trend) and real-time monitoring data (current migration rate 0.5ml / min) is realized, and the file is completed within 5 seconds.

[0253] Steps 1051 to 1054 solve the problems of heterogeneity and security barriers in cross-hospital medical data formats through protocol adaptive conversion and secure transmission mechanisms. The multimodal protocol parsing tree implements dynamic parsing and reconstruction of equipment compatibility rules to ensure seamless conversion of data formats; dynamic checksums and priority division mechanisms ensure the transmission priority of key physiological parameters (such as microvascular waveform stability); two-way verification keys combined with timestamp anchoring technology strike a balance between encryption strength and transmission efficiency to avoid data gaps during transfer. The entire solution supports real-time synchronization and precise reconstruction of allergy monitoring data for burn patients in cross-institutional and multi-protocol scenarios, providing a reliable data base for continuous treatment.

[0254] Figure 2 A schematic diagram of the structure of a real-time processing system for rapid identification and file creation of emergency patients is provided for the embodiment of the present application, such as Figure 2 As shown, the device comprises:

[0255] A capture module 21 is used to wear the smart medical fabric on the skin surface of the burn patient, dynamically capture the reflected light intensity distribution and microvascular pulsation waveform on the skin surface through the flexible photoelectric sensor array embedded in the smart medical fabric, and automatically adjust the light transmittance and sampling frequency of the flexible photoelectric sensor array according to the degree of scar hyperplasia and epidermal temperature;

[0256] The conversion module 22 is used to convert the reflected light intensity distribution and the microvascular pulsation waveform into a combination parameter including epidermal oxygen permeability, tissue fluid migration rate and local immune cell density after multi-channel noise suppression, and match them with a pre-established burn skin allergy characteristic spectrum. When the matching result exceeds a preset skin tolerance dynamic range, the edge computing node is triggered to generate a graded alarm signal of an allergic reaction;

[0257] The encapsulation module 23 is used to encapsulate the hierarchical alarm signal and the real-time physical sign data of the patient into a cross-hospital transmission data packet through the edge computing node, and send a data sharing request to the collaborating hospital based on a privacy protection policy, where the privacy protection policy automatically adjusts the encryption intensity according to the patient identity sensitivity and the data desensitization level;

[0258] The generation module 24 is used to, after obtaining the authorization of the collaborating hospital, perform feature space alignment on the burn skin allergy characteristic spectrum and the heterogeneous medical data of the collaborating hospital to generate a cross-institutional allergy reaction prediction model, and synchronize the updated model parameters to the edge computing node;

[0259] The reconstruction module 25 is used to, when it is detected that the burned patient is transferred to another hospital, automatically reconstruct the allergy monitoring file of the burned patient according to the device compatibility protocol of the target hospital.

[0260] Figure 2 The real-time processing device for rapid identification and filing of emergency patients can execute Figure 1 The real-time processing method for rapid identification and filing of emergency patients shown in the embodiments, and its implementation principle and technical effects will not be elaborated. For the real-time processing device for rapid identification and filing of emergency patients in the above embodiments, the specific ways for each module and unit to execute operations have been described in detail in the embodiments related to the method, and will not be elaborated here.

[0261] In a possible design, Figure 2 The real-time processing device for rapid identification and filing of emergency patients shown in the embodiments can be implemented as a computing device, such as Figure 3 shown, this computing device may include a storage component 31 and a processing component 32;

[0262] The storage component 31 stores one or more computer instructions, where the one or more computer instructions are called and executed by the processing component 32.

[0263] The processing component 32 is used for the Figure 1 real-time processing method for rapid identification and filing of emergency patients shown in the above

[0264] embodiments. Among them, the processing component 32 may include one or more processors to execute computer instructions to complete all or part of the steps in the above method. Of course, the processing component may also be implemented by one or more application specific integrated circuits (ASICs), digital signal processors (DSPs), digital signal processing devices (DSPDs), programmable logic devices (PLDs), field programmable gate arrays (FPGAs), controllers, microcontrollers, microprocessors or other electronic components for executing the above method.

[0265] The storage component 31 is configured to store various types of data to support the operation of the terminal. The storage component can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic memory, flash memory, magnetic disk or optical disk.

[0266] Of course, the computing device may also necessarily include other components, such as an input / output interface, a display component, a communication component, etc.

[0267] The input / output interface provides an interface between the processing component and the peripheral interface module, and the peripheral interface module may be an output device, an input device, etc.

[0268] The communication component is configured to facilitate communication between the computing device and other devices in a wired or wireless manner, etc.

[0269] Among them, the computing device may be a physical device or an elastic computing host provided by a cloud computing platform, etc. At this time, the computing device may refer to a cloud server, and the above-mentioned processing component, storage component, etc. may be basic server resources leased or purchased from a cloud computing platform.

[0270] An embodiment of the present application also provides a computer storage medium storing a computer program, and when the computer program is executed by a computer, it can implement the above Figure 1 real-time processing method for rapid identification and filing of emergency patients shown in the above embodiments.

[0271] Those skilled in the art can clearly understand that for the convenience and brevity of description, the specific working processes of the above-described systems, devices, and units can refer to the corresponding processes in the foregoing method embodiments, and will not be described herein again.

[0272] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place, or may be distributed to multiple network units. Some or all of the modules can be selected according to actual needs to achieve the purpose of the solution of this embodiment. Those of ordinary skill in the art can understand and implement it without creative efforts.

[0273] Through the description of the above embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus a necessary general hardware platform, and of course, it can also be implemented by hardware. Based on such an understanding, the essence of the above technical solution, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to enable a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or some parts of the embodiments.

[0274] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them; although the present application has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that they can still modify the technical solutions recorded in the foregoing embodiments, or perform equivalent replacements for some of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. A real-time processing method for rapid identification and filing of emergency patients, characterized in that: include: Wearing the smart medical fabric on the skin surface of the burn patient, dynamically capturing the reflected light intensity distribution and microvascular pulsation waveform on the skin surface through the flexible photoelectric sensor array embedded in the smart medical fabric, and automatically adjusting the light transmittance and sampling frequency of the flexible photoelectric sensor array according to the degree of scar hyperplasia and epidermal temperature; The reflected light intensity distribution and the microvascular pulsation waveform are converted into a combination parameter including epidermal oxygen permeability, tissue fluid migration rate and local immune cell density after multi-channel noise suppression, and matched with a pre-established burn skin allergy characteristic spectrum. When the matching result exceeds a preset skin tolerance dynamic range, the edge computing node is triggered to generate a graded alarm signal of an allergic reaction; The edge computing node encapsulates the graded alarm signal and the patient's real-time vital sign data into a cross-hospital transmission data packet, and sends a data sharing request to the cooperative hospital based on a privacy protection strategy. The privacy protection strategy automatically adjusts the encryption strength according to the patient's identity sensitivity and the data desensitization level; After obtaining authorization from the cooperative hospital, align the burn skin allergy feature spectrum with the heterogeneous medical data of the cooperative hospital in feature space, generate a cross-institutional allergic reaction prediction model, and synchronize the updated model parameters to the edge computing node; When it is detected that the burn patient is transferred to another hospital, the allergy monitoring file of the burn patient is automatically reconstructed according to the device compatibility protocol of the target hospital.

2. The method according to claim 1, characterized in that The reflected light intensity distribution and the microvascular pulsation waveform are converted into a combination parameter including epidermal oxygen permeability, tissue fluid migration rate and local immune cell density after multi-channel noise suppression, and matched with a pre-established burn skin allergy characteristic spectrum. When the matching result exceeds a preset skin tolerance dynamic range, the edge computing node is triggered to generate a graded alarm signal of an allergic reaction, including: Dynamic baseline drift elimination is performed on the multi-channel raw signal output by the flexible photoelectric sensor array, an adaptive filter is established based on the correlation of light intensity fluctuations between adjacent channels, light intensity mutation artifacts generated by the limb movement of the burn patient are removed, and a purified waveform sequence with a time stamp is generated by combining the reflected light intensity distribution and the microvascular pulsation waveform; The purified waveform sequence is decomposed into multiple frequency bands through multi-channel noise suppression, the low-frequency energy proportion is extracted for inverting the epidermal oxygen permeability, the mid-frequency phase gradient integral value is extracted for characterizing the tissue fluid migration rate, and the high-frequency waveform peak-to-valley variability is extracted and combined with the multi-scale morphological decomposition results to calculate the local immune cell density; The epidermal oxygen permeability, the tissue fluid migration rate and the local immune cell density are fused into a combined parameter according to a preset ratio, a dynamic projection space is constructed in a pre-established burn skin allergy characteristic spectrum, and the Mahalanobis distance between the combined parameter and the dynamic projection space is calculated through a sliding time window to generate an allergic reaction similarity vector; When any dimension of the allergic reaction similarity vector exceeds the corresponding boundary of the preset skin tolerance dynamic range, the alarm logic tree pre-stored in the edge computing node is called according to the combination pattern of the exceeded dimensions, and a graded alarm signal for the allergic reaction is generated in combination with the autonomic nerve activation index of the burn patient.

3. The method according to claim 2, characterized in that The epidermal oxygen permeability, the tissue fluid migration rate and the local immune cell density are fused into a combination parameter according to a preset ratio, a dynamic projection space is constructed in a pre-established burn skin allergy characteristic spectrum, and the Mahalanobis distance between the combination parameter and the dynamic projection space is calculated through a sliding time window to generate an allergic reaction similarity vector, including: Inputting the time series of the epidermal oxygen permeability, the gradient change of the tissue fluid migration rate and the spatial distribution pattern of the local immune cell density into a multidimensional tensor field, generating a spatiotemporally coupled initial tensor structure according to the quantitative dimension of the allergic atom space in the pre-established burn skin allergy characteristic spectrum, and aligning each slice of the initial tensor structure with a data block in the multidimensional tensor field; A dynamic projection space is constructed in the burn skin allergy characteristic spectrum, and based on the multimodal data stream of historical allergic reaction events in the burn skin allergy characteristic spectrum, an adaptive topological network with the allergic atom space as nodes is constructed in the dynamic projection space, wherein the connection strength between nodes in the adaptive topological network is dynamically adjusted by the pathological correlation degree and time attenuation coefficient of the allergic atom space; An incremental covariance matrix update mechanism is adopted to calculate the Mahalanobis distance between the initial tensor structure and each node in the adaptive topology network frame by frame through a sliding time window, and the update amount of the incremental covariance matrix is ​​jointly controlled by the stability index of the microvascular pulsation waveform and the current sampling frequency of the flexible photoelectric sensor array; The inverse value of the Mahalanobis distance is input into a nonlinear transformation function to generate an allergic reaction entropy value, and dynamic weight allocation is performed on each dimension of the allergic reaction entropy value through a dynamic weight allocator embedded in the edge computing node. The parameters of the dynamic weight allocator are jointly adjusted by the metabolic rate and inflammatory factor concentration of the burn patient, and an allergic reaction similarity vector is output.

4. The method according to claim 3, characterized in that The updating mechanism of the incremental covariance matrix is ​​adopted, and the Mahalanobis distance between the initial tensor structure and each node in the adaptive topology network is calculated frame by frame through a sliding time window, and the updating amount of the incremental covariance matrix is ​​jointly controlled by the stability index of the microvascular pulsation waveform and the current sampling frequency of the flexible photoelectric sensor array, including: Converting the stability index of the microvascular pulsation waveform into a waveform oscillation discreteness parameter, and generating a sampling frequency time primitive according to the current sampling frequency of the flexible photoelectric sensor array, and multiplying the waveform oscillation discreteness parameter with the sampling frequency time primitive to generate a covariance increment step adjustment coefficient; The ratio of the trace norm of the historical covariance matrix to the spectral radius of the current data block is obtained through a sliding time window, and the sliding covariance basis is generated by adopting an updating mechanism of the incremental covariance matrix in combination with the covariance incremental step adjustment coefficient, the initial tensor structure is divided into overlapping sub-blocks along the time axis and orthogonally projected block by block with the sliding covariance basis, and an updated incremental covariance matrix is ​​output, wherein the updating amount of the incremental covariance matrix is ​​jointly controlled by the stability index of the microvascular pulsation waveform and the current sampling frequency of the flexible photoelectric sensor array; A time-attenuated sliding window is bound in the adaptive topology network, and a Mahalanobis distance is generated using the principal component direction of the updated incremental covariance matrix and the pathological correlation degree of the allergic atom space, wherein each component of the Mahalanobis distance is time-scale normalized by the sampling frequency time primitive; The Mahalanobis distance is integrated in time series through the dynamic distance buffer of the edge computing node, the interval length of the dynamic distance buffer is adaptively adjusted according to the fluctuation period of the waveform oscillation discreteness parameter, and the Mahalanobis distance is synchronously processed with the sliding time window.

5. The method according to claim 4, characterized in that The method comprises: binding a time-attenuated sliding window in the adaptive topology network, generating a Mahalanobis distance using the principal component direction of the updated incremental covariance matrix and the pathological correlation degree of the allergic atom space, and normalizing each component of the Mahalanobis distance by the sampling frequency time primitive, including: A time decay sliding window is bound in the adaptive topology network, and based on the pathological stage marking data in the burn skin allergy characteristic spectrum, the decay rate of the time decay sliding window is associated with the acute inflammation index of the allergic atom space to generate a dynamic pathological decay weight, and the dynamic pathological decay weight is used to constrain the sliding step of the time decay sliding window; A hyperspherical projection is constructed by using the principal component direction of the updated incremental covariance matrix and the pathological correlation vector of the allergic atom space, and the product of the tangential component modulus and the normal component attenuation factor of the principal component direction in the hyperspherical projection is calculated to generate a pathologically corrected Mahalanobis primitive; The pathologically corrected Mahalanobis primitive is subjected to a time domain stretch transformation using the sampling frequency time primitive, and a stretch transformation smoothing factor is dynamically adjusted through a stability index of the microvascular pulsation waveform to generate a standardized Mahalanobis distance component synchronized with a metabolic cycle of biological tissues; In the multi-channel fusion cache pool of the edge computing node, cross-channel correlation disambiguation is performed on the standardized Mahalanobis distance component, and the channel bandwidth of the multi-channel fusion cache pool is adaptively adjusted according to the changing gradient of the dynamic pathology attenuation weight, and the Mahalanobis distance matching the pathology evolution trend in the sliding time window is output to drive the nonlinear transformation process of the allergic reaction entropy value.

6. The method according to claim 5, characterized in that The method comprises: constructing a hyperspherical projection through the principal component direction of the updated incremental covariance matrix and the pathological correlation vector of the allergic atom space, calculating the product of the tangential component modulus and the normal component attenuation factor of the principal component direction in the hyperspherical projection, and generating a pathologically corrected Markov primitive, including: Constructing a dynamic rotation matrix based on the pathological correlation vector of the allergic atom space, projecting the principal component direction of the incremental covariance matrix to the hyperspherical space corresponding to the dynamic rotation matrix, and generating a hyperspherical projection trajectory carrying the pathological feature vector; Decomposing a tangential projection operator and a normal attenuation kernel function on the hypersphere projection trajectory, the modulus of the tangential projection operator is dynamically weighted by the cosine value of the angle between the principal component direction and the pathological correlation vector, and the normal attenuation kernel function is jointly calibrated by the historical change gradient of the epidermal oxygen permeability and the instantaneous fluctuation of the tissue fluid migration rate; The modulus of the tangential projection operator is input into the normal attenuation kernel function, and multi-order smoothing is performed through the pathological primitive optimization pool of the edge computing node, wherein the smoothing order of the pathological primitive optimization pool is dynamically controlled by the stability index of the microvascular pulsation waveform and the update frequency of the dynamic rotation matrix; Based on the multi-order smoothing result, a smoothed modulus-attenuation product is obtained, and the metabolic gradient parameter of the burn patient is used to compensate the dynamic viscosity coefficient of the tissue fluid for the smoothed modulus-attenuation product, so as to generate a Markov primitive set including a multi-dimensional pathological correction factor. After the sensor drift noise is eliminated by the primitive disturbance suppression factor of the edge computing node, the Markov primitive set outputs a pathologically corrected Markov primitive matching the pathological stage of the allergic atom space.

7. The method according to claim 1, characterized in that When it is detected that the burn patient is transferred to another hospital, the allergy monitoring file of the burn patient is automatically reconstructed according to the device compatibility protocol of the target hospital, including: The trigger instruction for hospital transfer of the burn patient is captured in real time through the medical system interface, the sensor data format constraint conditions and encryption transmission rules in the device compatibility protocol of the target hospital are parsed, and a multimodal protocol parsing tree is generated, wherein each branch node of the multimodal protocol parsing tree corresponds to an archive reconstruction mode; According to the topological structure of the multimodal protocol parsing tree, the epidermal oxygen permeability time series, the tissue fluid migration rate gradient value and the allergic reaction probability vector in the allergy monitoring file of the burn patient are rearranged according to the equipment sampling dimension of the target hospital to generate a file topological structure that matches the equipment compatibility protocol of the target hospital, and each data unit in the file topological structure is bound to a dynamic verification code; Based on the current burn stage mark and the real-time allergic reaction entropy value of the burn patient, the dynamic check codes in the archive topology are sorted according to the metabolic urgency parameter and divided into high-priority data blocks and low-priority data streams, wherein the high-priority data blocks contain the stability index of the microvascular pulsation waveform and the pathologically corrected Markov primitive, and the low-priority data stream contains the historical sampling frequency time primitive of the smart medical fabric; Through the dynamic verification token generator of the edge computing node, the high-priority data block and the low-priority data stream are encapsulated into a multi-level verification data packet according to the encryption transmission rules of the target hospital. The dynamic verification token generator generates a two-way verification key using the real-time metabolic rate of the burn patient and the encryption strength of the device compatibility protocol. The two-way verification key interacts with the receiving terminal of the target hospital through the multi-band communication module of the smart medical fabric for timestamp anchoring, completing the cross-protocol fusion and archiving of allergy history data and real-time monitoring data.

8. A real-time processing system for rapid identification and filing of emergency patients, characterized in that: include: A capture module is used to wear the smart medical fabric on the skin surface of the burn patient, dynamically capture the reflected light intensity distribution and microvascular pulsation waveform on the skin surface through the flexible photoelectric sensor array embedded in the smart medical fabric, and automatically adjust the light transmittance and sampling frequency of the flexible photoelectric sensor array according to the degree of scar hyperplasia and epidermal temperature; A conversion module, used to convert the reflected light intensity distribution and the microvascular pulsation waveform into a combination parameter including epidermal oxygen permeability, tissue fluid migration rate and local immune cell density after multi-channel noise suppression, and match them with a pre-established burn skin allergy characteristic spectrum. When the matching result exceeds a preset skin tolerance dynamic range, the edge computing node is triggered to generate a graded alarm signal of an allergic reaction; An encapsulation module, used to encapsulate the graded alarm signal and the patient's real-time vital sign data into a cross-hospital transmission data packet through the edge computing node, and send a data sharing request to the cooperative hospital based on a privacy protection strategy, wherein the privacy protection strategy automatically adjusts the encryption strength according to the patient's identity sensitivity and the data desensitization level; A generation module, which is used to align the burn skin allergy feature spectrum with the heterogeneous medical data of the collaborative hospital after obtaining authorization from the collaborative hospital, generate a cross-institutional allergic reaction prediction model, and synchronize the updated model parameters to the edge computing node; The reconstruction module is used to automatically reconstruct the allergy monitoring file of the burn patient according to the equipment compatibility protocol of the target hospital when it is detected that the burn patient is transferred to another hospital.

9. A computing device, characterized in that It comprises a processing component and a storage component; the storage component stores one or more computer instructions; the one or more computer instructions are used to be called and executed by the processing component to implement a real-time processing method for rapid identification and filing of emergency patients as described in any one of claims 1 to 7.

10. A computer storage medium, characterized in that: A computer program is stored, and when the computer program is executed by a computer, a real-time processing method for rapid identification and filing of emergency patients as described in any one of claims 1 to 7 is implemented.

Images