Pain grading evaluation system based on multi-modal physiological signal fusion

Abstract

The invention discloses a pain grading evaluation system based on multi-modal physiological signal fusion, and relates to the related field of data processing, and the method comprises the steps: capturing muscle tissue contraction of a target area, and obtaining time sequence muscle contraction distribution; initial temperature field distribution is calculated according to thermal imaging of the target area, temperature compensation is carried out based on the basic time sequence body temperature, and time sequence temperature field distribution is output. And calling real-time physiological data for analysis, and outputting coupling fluctuation correlation characteristics. After the time sequence muscle contraction distribution and the time sequence temperature field distribution are registered, dynamic characteristics are extracted, ORI correlation characteristics and regional thermodynamic characteristics are output, coupling fluctuation correlation characteristics are combined to be input into a physiological stress grading space, and initial stress grading is matched and positioned. And performing dynamic threshold compensation on the initial stress grade according to the basic time sequence body temperature, and outputting a real-time physiological pain grade. The technical problem that a traditional pain assessment method is low in accuracy of pain assessment is solved.

Description

Technical Field

[0001] This application relates to the field of data processing, and particularly to a pain grading assessment system based on multimodal physiological signal fusion. Background Art

[0002] In the fields of medicine and biology, pain assessment usually relies on the monitoring and analysis of physiological signals. Traditional pain grading methods usually focus on subjective clinical assessments and lack real-time dynamic monitoring and quantitative analysis of physiological responses. With the continuous development of biomedical technology, research on using multimodal physiological signals for pain grading has gradually become a trend. By combining different physiological characteristics, it is possible to more accurately reflect an individual's physiological stress response and their pain state.

[0003] Therefore, in the prior art, the pain assessment method lacks an accurate real-time physiological data analysis and dynamic update mechanism, and cannot fully consider the influence of environmental factors, resulting in the technical problem of low accuracy of pain assessment. Summary of the Invention

[0004] This application provides a pain grading assessment system based on multimodal physiological signal fusion, which solves the technical problem that the pain assessment method in the prior art lacks an accurate real-time physiological data analysis and dynamic update mechanism, cannot fully consider the influence of environmental factors, and results in low accuracy of pain assessment. By combining multimodal physiological signals and using a dynamic stress vector update mechanism, this method can monitor an individual's physiological stress response in real time and visualize the changes in the stress state through a physiological trajectory sphere model, thereby achieving a more accurate and reliable pain grading.

[0005] This application provides a pain grading assessment system based on multimodal physiological signal fusion. The system includes: a contraction distribution acquisition module for capturing the muscle tissue contraction of a target area to obtain a temporal muscle contraction distribution; a temperature field distribution acquisition module for calculating an initial temperature field distribution based on the thermal imaging of the target area and performing temperature compensation according to the basic temporal body temperature transmitted back by an ear temperature sensor, and outputting a temporal temperature field distribution; a correlation feature acquisition module for docking with standard medical devices to call real-time physiological data, performing physiological fluctuation correlation analysis, and outputting coupled fluctuation correlation features; a dynamic feature acquisition module for performing dynamic feature extraction after spatially registering the temporal muscle contraction distribution and the temporal temperature field distribution, and outputting ORI correlation features and regional thermodynamic features; a stress grading module for inputting the ORI correlation features, regional thermodynamic features, and coupled fluctuation correlation features into a physiological stress grading space to match and locate an initial stress grading; and a pain grading module for performing dynamic threshold compensation on the initial stress grading based on the basic temporal body temperature and outputting a real-time physiological pain grading.

[0006] In a possible implementation, the pain grading module is further configured to: construct a real-time stress vector based on the ORI correlation feature, the regional thermodynamics feature, and the coupled fluctuation correlation feature; preset a stress monitoring window, and dynamically update the real-time stress vector through the stress monitoring window to output a plurality of incremental stress vectors; construct a physiological trajectory sphere model based on the plurality of incremental stress vectors, and perform visual rendering of the physiological trajectory sphere model using a plurality of incremental stress gradings.

[0007] In a possible implementation, the correlation feature acquisition module is further configured to: obtain respiratory rate and blood pressure parameters in real time through a standard medical device interface to obtain respiratory rate time series data and blood pressure parameter time series data; perform phase synchronization analysis on the respiratory rate time series data and the blood pressure parameter time series data to construct a respiration-blood pressure coupling sequence; perform sliding window cross-correlation analysis on the respiration-blood pressure coupling sequence, and output the coupled fluctuation correlation feature.

[0008] In a possible implementation, the correlation feature acquisition module is further configured to: construct sliding window parameters according to a standard respiratory cycle, where the sliding window parameters include a window length and a sliding step size, the window length is H standard respiratory cycles, and the sliding step size is 1 / M standard respiratory cycle; slide and segment the respiration-blood pressure coupling sequence based on the sliding window parameters to output a plurality of window coupling sequences; use a normalized cross-correlation function to calculate a plurality of cross-correlation coefficients and a plurality of coupling time delays of the plurality of window coupling sequences; sort the plurality of cross-correlation coefficients in ascending order, extract the maximum cross-correlation coefficient, and correspondingly call the target coupling time delay, where the maximum cross-correlation coefficient and the target coupling time delay constitute the coupled fluctuation correlation feature.

[0009] In a possible implementation, the contraction distribution acquisition module is further configured to: during the process of irradiating the target area with an 850 nm near-infrared light source and a 520 nm visible light source in a pulse alternating mode, synchronously capture the dual-wavelength reflected light intensity distribution through a shutter sensor to generate a reflected light intensity alternating sequence; calculate and output a light intensity difference sequence based on the reflected light intensity alternating sequence; input the light intensity difference sequence into a tissue deformation correlation model for local deformation inversion to obtain a local deformation amount sequence; perform elastic modulus inversion on the local deformation amount sequence to output an instantaneous elastic modulus sequence; calculate the sliding window change rate of the instantaneous elastic modulus sequence to output an elastic modulus time series change curve as the time series muscle contraction distribution.

[0010] In a possible implementation, the temperature field distribution acquisition module is further configured to: configure a far-infrared thermal imager according to the spatial characteristics of the target area, where the target area is at the center of the field of view of the far-infrared thermal imager; capture the original temperature of the target area based on the far-infrared thermal imager to obtain multiple frames of original temperature matrices.

[0011] In a possible implementation, the temperature field distribution acquisition module is further configured to: collect continuous resting state data of a target user through the ear temperature sensor to construct the basic temporal body temperature; locate a vascular bifurcation point matrix in the target area; after spatially aligning the multiple frames of original temperature matrices based on the vascular bifurcation point matrix as a reference feature, calculate the gradient distribution of adjacent frames of original temperature matrices and output an initial temperature field distribution; perform dynamic temperature compensation on the initial temperature field distribution according to the basic temporal body temperature and output the temporal temperature field distribution.

[0012] In a possible implementation, the dynamic feature acquisition module is further configured to: perform cross-modal spatial registration of the temporal muscle contraction distribution and the temporal temperature field distribution based on the vascular bifurcation point matrix as a reference feature and output a registered muscle contraction distribution and a registered temperature field distribution; calculate the elastic modulus change rate of the registered muscle contraction distribution, locate the peak point of the change rate as the instantaneous extreme value of muscle contraction; perform peak decay tracking starting from the peak point of the change rate and output a contraction recovery delay window, where the contraction recovery delay window and the instantaneous extreme value of muscle contraction constitute the ORI correlation feature; start from the center of the target area and traverse the registered temperature field distribution to calculate and output the regional temperature difference gradient; perform metabolic heat production prediction on the registered temperature field distribution and output the regional heat production rate, where the regional temperature difference gradient and the regional heat production rate constitute the regional thermodynamics feature.

[0013] In a possible implementation, the stress grading module is further configured to: locally de-privatize and call multiple sample multi-modal physiological features and multiple sample stress gradings; establish the physiological stress grading space according to the standard index composition of the sample multi-modal physiological features; after locating multiple sample particle points in the physiological stress grading space based on the multiple sample multi-modal physiological features, use the multiple sample stress gradings to label the multiple sample particle points to complete the data filling of the physiological stress grading space; input the ORI correlation feature, the regional thermodynamics feature, and the coupled fluctuation correlation feature into the physiological stress grading space to locate a virtual particle point; perform spatial matching of the virtual particle point with the goal of minimizing the distance, locate the nearest particle point, and use the sample stress grading of the nearest particle point as the initial stress grading. The sample multi-modal physiological features include sample ORI features, sample thermodynamics features, and sample coupling features.

[0014] One or more technical solutions provided in this application have at least the following technical effects or advantages:

[0015] The pain grading evaluation system based on multimodal physiological signal fusion provided in this application includes: a contraction distribution acquisition module, configured to capture the muscle tissue contraction of the target area to obtain a temporal muscle contraction distribution; a temperature field distribution acquisition module, configured to calculate an initial temperature field distribution based on the thermal imaging of the target area, and perform temperature compensation according to the basic temporal body temperature transmitted back by the ear temperature sensor, and output a temporal temperature field distribution; a correlation feature acquisition module, configured to connect to a standard medical device to call real-time physiological data, perform physiological fluctuation correlation analysis, and output a coupled fluctuation correlation feature; a dynamic feature acquisition module, configured to perform dynamic feature extraction after spatially registering the temporal muscle contraction distribution and the temporal temperature field distribution, and output an ORI correlation feature and a regional thermodynamic feature; a stress grading module, configured to input the ORI correlation feature, the regional thermodynamic feature, and the coupled fluctuation correlation feature into a physiological stress grading space to match and locate an initial stress grading; a pain grading module, configured to perform dynamic threshold compensation on the initial stress grading according to the basic temporal body temperature, and output a real-time physiological pain grading. This solves the technical problem in the prior art that the pain evaluation method lacks accurate real-time physiological data analysis and dynamic update mechanism, and cannot fully consider the influence of environmental factors, resulting in low accuracy of pain evaluation. By combining multimodal physiological signals and using a dynamic stress vector update mechanism, this method can monitor the physiological stress response of an individual in real time, and visualize the change of the stress state through a physiological trajectory sphere model, so as to achieve a more accurate and reliable pain grading. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings of the embodiments of the present invention will be briefly introduced below. Flowcharts are used in this application to illustrate the operations performed by the system according to the embodiments of the present application. It should be understood that the operations in the front or below do not necessarily need to be executed precisely in sequence. On the contrary, according to the need, various steps can be executed in reverse order or simultaneously. At the same time, other operations can also be added to these processes, or one or several operations can be removed from these processes.

[0017] Figure 1 It is a schematic structural diagram of the pain grading evaluation system based on multimodal physiological signal fusion provided by the embodiments of this application;

[0018] Figure 2 It is a schematic flow diagram of the contraction distribution acquisition module in the pain grading evaluation system based on multimodal physiological signal fusion of this application for determining the temporal muscle contraction distribution.

[0019] Description of the reference numerals: The contraction distribution acquisition module 11, the temperature field distribution acquisition module 12, the associated feature acquisition module 13, the dynamic feature acquisition module 14, the stress grading module 15, and the pain grading module 16. Detailed implementation manners

[0020] The above description is only an overview of the technical solution of this application. In order to be able to understand the technical means of this application more clearly, it can be implemented according to the content of the specification. And in order to make the above and other purposes, features, and advantages of this application more obvious and understandable, the following specifically gives the detailed implementation manners of this application.

[0021] In order to make the purpose, technical solution, and advantages of this application clearer, the following will further describe this application in detail with reference to the accompanying drawings. The described embodiments should not be regarded as limitations of this application. All other embodiments obtained by those of ordinary skill in the art without creative efforts fall within the scope of protection of this application.

[0022] In the following description, "some embodiments" are involved, which describe a subset of all possible embodiments. However, it can be understood that "some embodiments" can be the same subset or different subsets of all possible embodiments, and can be combined with each other without conflict. The terms "first / second" involved are only used to distinguish similar objects and do not represent a specific order for the objects. The terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusion. For example, a process, system, product, or server including a series of steps or units does not necessarily have to be limited to those steps or units clearly listed, but may include other steps or modules not clearly listed or inherent to these processes, systems, products, or devices. Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by those skilled in the technical field to which this application belongs. The terms used herein are only for the purpose of describing the embodiments of this application.

[0023] The embodiments of this application provide a pain grading evaluation system based on multimodal physiological signal fusion, as Figure 1 shown. The system includes:

[0024] The contraction distribution acquisition module 11 is used to capture the contraction of the muscle tissue in the target area to obtain the temporal muscle contraction distribution; the temperature field distribution acquisition module 12 is used to calculate the initial temperature field distribution based on the thermal imaging of the target area and perform temperature compensation according to the basic temporal body temperature transmitted back by the ear temperature sensor, and output the temporal temperature field distribution; the associated feature acquisition module 13 is used to call real-time physiological data by docking with standard medical devices, perform physiological fluctuation correlation analysis, and output the coupled fluctuation correlation features.

[0025] By arranging an 850nm near-infrared light source and a 520nm visible light source in a target area such as a patient's limb, muscle group or specific treatment area and irradiating the target area in a pulsed alternating mode, the contraction distribution acquisition module 11 is used to capture the contraction of muscle tissue in the target area to obtain a temporal muscle contraction distribution, and the temporal muscle contraction distribution is a change curve of the muscle elastic modulus over time. Subsequently, the temperature field distribution acquisition module 12 is used to obtain a thermal image by capturing the infrared radiation in the target area, and calculate the initial temperature field distribution based on the thermal imaging of the target area. And the body temperature data is collected by an ear temperature sensor, combined with the thermal imaging data for temperature compensation. Specifically, the local temperature rise caused by the inflammatory reaction is detected through the temperature difference gradient between the center and the edge of the wound, and then the user's basal body temperature is synchronously measured by a forehead temperature or ear temperature sensor for whole-body temperature compensation to eliminate the interference of the ambient temperature, so as to obtain a more accurate temporal temperature field distribution. Further, the correlation feature acquisition module 13 is used to call real-time physiological data by docking with standard medical devices. The physiological data includes the respiratory cycle and blood pressure, and real-time physiological data is constructed for physiological fluctuation correlation analysis, and coupled fluctuation correlation features are output.

[0026] Further, the correlation feature acquisition module 13 is also used to obtain the respiratory rate and blood pressure parameters in real time through the standard medical device interface to obtain the respiratory rate time series data and the blood pressure parameter time series data; perform phase synchronization analysis on the respiratory rate time series data and the blood pressure parameter time series data to construct a respiration-blood pressure coupling sequence; perform sliding window cross-correlation analysis on the respiration-blood pressure coupling sequence, and output the coupled fluctuation correlation features.

[0027] Calling real-time physiological data by docking with standard medical devices for physiological fluctuation correlation analysis and outputting coupled fluctuation correlation features, the method includes: obtaining the respiratory rate and blood pressure parameters in real time through the standard medical device interface to obtain the respiratory rate time series data and the blood pressure parameter time series data. Subsequently, perform phase synchronization analysis on the respiratory rate time series data and the blood pressure parameter time series data, such as aligning the peaks or valleys of the respiratory rate time series data and the blood pressure parameter time series data, to construct a respiration-blood pressure coupling sequence. Perform sliding window cross-correlation analysis on the respiration-blood pressure coupling sequence, and output the coupled fluctuation correlation features.

[0028] Further, the associated feature acquisition module 13 is further configured to: construct a sliding window parameter according to a standard respiratory cycle, where the sliding window parameter includes a window length and a sliding step, the window length is H standard respiratory cycles, and the sliding step is 1 / M standard respiratory cycle; slide and segment the respiration-blood pressure coupling sequence based on the sliding window parameter, and output a plurality of window coupling sequences; use a normalized cross-correlation function to calculate a plurality of cross-correlation coefficients and a plurality of coupling time delays of the plurality of window coupling sequences; sort the plurality of cross-correlation coefficients in ascending order, extract the maximum cross-correlation coefficient, and correspondingly call the target coupling time delay, where the maximum cross-correlation coefficient and the target coupling time delay constitute the coupling fluctuation association feature.

[0029] Perform sliding window cross-correlation analysis on the respiration-blood pressure coupling sequence and output the coupling fluctuation association feature. The method includes: constructing a sliding window parameter according to a standard respiratory cycle, where the standard respiratory cycle is the required duration to complete H complete respiratory cycles under normal conditions. The sliding window parameter includes a window length and a sliding step, the window length is H standard respiratory cycles, and the sliding step is 1 / M standard respiratory cycle. H and M are positive integers greater than 1. Further, slide and segment the respiration-blood pressure coupling sequence based on the sliding window parameter, segment a respiration-blood pressure coupling sequence into multiple window lengths and data corresponding to the sliding step under each window length, and output a plurality of window coupling sequences. Further, use a normalized cross-correlation function to calculate a plurality of cross-correlation coefficients and a plurality of coupling time delays of the plurality of window coupling sequences. The normalized cross-correlation function is:

[0030]

[0031] where R(t) is the time series data of the respiration signal, B(t) is the time series data of the blood pressure signal, and τ is the delay offset. The mean of the time series data of the respiration signal. is the mean of the time series data of the blood pressure signal. Finally, extract the maximum cross-correlation coefficient from the cross-correlation coefficients calculated for multiple windows. The maximum cross-correlation coefficient represents the strongest synchronization between respiration and blood pressure during the entire analysis process. Sort the plurality of cross-correlation coefficients in ascending order and extract the maximum cross-correlation coefficient. And correspondingly call the target coupling time delay. The target coupling time delay is the coupling time delay corresponding to the maximum cross-correlation coefficient, which represents the best phase synchronization point between respiration and blood pressure. The maximum cross-correlation coefficient and the target coupling time delay constitute the coupling fluctuation association feature for subsequent pain grading analysis.

[0032] Further, as Figure 2As shown, the contraction distribution acquisition module 11 is further configured to: during the process of irradiating the target area with an 850 nm near-infrared light source and a 520 nm visible light source in a pulse-alternating mode, synchronously capture the dual-wavelength reflected light intensity distribution through a shutter sensor to generate a reflected light intensity alternating sequence; calculate and output an intensity difference sequence based on the reflected light intensity alternating sequence; input the intensity difference sequence into a tissue deformation correlation model for local deformation inversion to obtain a local deformation quantity sequence; perform elastic modulus inversion on the local deformation quantity sequence to output an instantaneous elastic modulus sequence; calculate the sliding window change rate of the instantaneous elastic modulus sequence and output an elastic modulus time-series change curve as the time-series muscle contraction distribution.

[0033] Capture the muscle tissue contraction of the target area to obtain the time-series muscle contraction distribution. The method includes: using an 850 nm near-infrared light source and a 520 nm visible light source to irradiate the target area in a pulse-alternating mode, that is, the near-infrared light source and the visible light source are alternately emitted, and synchronously capture the dual-wavelength reflected light intensity distribution through a shutter sensor to ensure that the intensity of the reflected light from the target area can be accurately recorded each time the light source is switched, generating a reflected light intensity alternating sequence. Subsequently, calculate and output an intensity difference sequence based on the reflected light intensity alternating sequence. The intensity difference sequence is the difference between adjacent reflected light intensity alternating sequences, such as the difference between the first and the second reflected light intensities, the difference between the third and the fourth reflected light intensities, the difference between the fifth and the sixth reflected light intensities. The reflected light intensities that have participated in the calculation are not calculated a second time, thereby obtaining the intensity difference sequence. Based on a neural network model, construct a tissue deformation correlation model. The construction data is the intensity difference sequence and the corresponding local deformation quantity parameters in the historical muscle tissue contraction monitoring data. For example, when the intensity difference sequence is 50 units and the local deformation quantity is 0.2 mm, it represents the local deformation of the muscle surface during the contraction process. Supervise and train the neural network model with the construction data until the output accuracy of the model meets the requirements to complete the training, thereby obtaining the tissue deformation correlation model. Further, input the intensity difference sequence into the tissue deformation correlation model for local deformation inversion to obtain a local deformation quantity sequence. Use experiments or calibration methods to perform elastic modulus inversion on the local deformation quantity sequence through the local deformation quantity and the corresponding applied stress, thereby outputting an instantaneous elastic modulus sequence. Finally, calculate the sliding window change rate of the instantaneous elastic modulus sequence. The sliding window is a preset time window, and the window is the same as the interval duration of each parameter of the instantaneous elastic modulus sequence. Calculate the change rate of adjacent parameters of the instantaneous elastic modulus sequence through the sliding window, and output an elastic modulus time-series change curve according to the calculated change rate as the time-series muscle contraction distribution.

[0034] Further, the temperature field distribution acquisition module 12 is further configured to: configure a far-infrared thermal imager according to the spatial characteristics of the target area, where the target area is at the center of the field of view of the far-infrared thermal imager; capture the original temperature of the target area based on the far-infrared thermal imager to obtain multiple frames of original temperature matrices.

[0035] Configuring the far-infrared thermal imager according to the spatial characteristics of the target area makes the target area at the center of the field of view of the far-infrared thermal imager. After the configuration is completed, the original temperature of the target area is captured based on the far-infrared thermal imager to obtain multiple frames of original temperature matrices, and the temperature data corresponding to the imaging pixels at the position of the target area is displayed in the original temperature matrix.

[0036] Further, the temperature field distribution acquisition module 12 is further configured to: collect continuous resting state data of the target user through the ear temperature sensor to construct the basic timing body temperature; locate the blood vessel bifurcation point matrix in the target area; calculate the gradient distribution of adjacent frames of the original temperature matrices after spatially aligning the multiple frames of original temperature matrices based on the blood vessel bifurcation point matrix as a reference feature, and output the initial temperature field distribution; perform dynamic temperature compensation on the initial temperature field distribution according to the basic timing body temperature, and output the timing temperature field distribution.

[0037] Calculating the initial temperature field distribution based on the thermal imaging of the target area, and performing temperature compensation according to the basic timing body temperature returned by the ear temperature sensor to output the timing temperature field distribution. The method includes: collecting continuous resting state data of the target user through the ear temperature sensor, where the resting state is the body temperature measurement data obtained when the user is not affected by other interferences, and the measured temperature is closer to. Constructing the basic timing body temperature according to the continuous resting state data. Further, locating the blood vessel bifurcation point matrix in the target area, where the blood vessel bifurcation point matrix is the spatial coordinates of the blood vessel bifurcation position in the target area. Spatially aligning the multiple frames of original temperature matrices based on the blood vessel bifurcation point matrix as a reference feature, that is, aligning the spatial positions of the blood vessel bifurcation point matrix and the original temperature matrix, and the alignment process includes translation, rotation, or scaling to make the target areas corresponding to the multiple frames of data consistent. Calculating the gradient distribution of adjacent frames of the original temperature matrices and outputting the initial temperature field distribution, where the initial temperature field distribution is the temperature difference gradient between the wound center and the edge of each original temperature matrix. Finally, establishing a linear regression formula between the body temperature data and the thermal imaging temperature data, and the linear regression formula is obtained by finding the linear relationship between the basic timing body temperature and the initial temperature field distribution so that the thermal imaging temperature can be corrected with the body temperature data, and the linear regression formula can be constructed by professional technicians. Performing dynamic temperature compensation on the initial temperature field distribution according to the basic timing body temperature using the linear regression formula and outputting the timing temperature field distribution.

[0038] The dynamic feature acquisition module 14 is configured to perform dynamic feature extraction after spatially registering the temporal muscle contraction distribution and the temporal temperature field distribution, and output ORI correlation features and regional thermodynamic features. The stress grading module 15 is configured to input the ORI correlation features, the regional thermodynamic features, and the coupled fluctuation correlation features into the physiological stress grading space to match and locate the initial stress grading. The pain grading module 16 is configured to perform dynamic threshold compensation on the initial stress grading based on the basic temporal body temperature and output the real-time physiological pain grading.

[0039] The dynamic feature acquisition module 14 is used to perform spatial registration on the temporal muscle contraction distribution and the temporal temperature field distribution. After completion, dynamic feature extraction is executed to output ORI correlation features and regional thermodynamic features. Subsequently, the stress grading module 15 is used to input the ORI correlation features, the regional thermodynamic features, and the coupled fluctuation correlation features into the physiological stress grading space, locate the positions of the virtual points of the ORI correlation features, the regional thermodynamic features, and the coupled fluctuation correlation features in the physiological stress grading space, determine the closest particle point, and use the sample stress grading of the closest particle point as the initial stress grading. Finally, the pain grading module 16 is used to perform dynamic threshold compensation on the initial stress grading based on the basic temporal body temperature, obtain the preset adjustment amplitude of the stress grading corresponding to the rising or falling amplitude of the body temperature according to the change trend of the body temperature, such as a slight rise or fall of the basic body temperature, and output the real-time physiological pain grading to cope with the possible errors brought by the body temperature change, so as to make the pain assessment more accurate. For example, if the body temperature rises, it may mean that the body is experiencing a higher stress state, so the stress grading needs to be increased. For example, for every 1°C increase, the preset stress grading increase is 0.5. This preset parameter value can be adjusted according to actual data. It solves the technical problem that the pain assessment method in the prior art lacks accurate real-time physiological data analysis and dynamic update mechanism, cannot fully consider the influence of environmental factors, and results in low accuracy of pain assessment. By combining multi-modal physiological signals and using a dynamic stress vector update mechanism, this method can monitor the physiological stress response of an individual in real time and visualize the change of the stress state through a physiological trajectory sphere model, thus achieving a more accurate and reliable pain grading. It solves the technical problem that the pain assessment method in the prior art lacks accurate real-time physiological data analysis and dynamic update mechanism, cannot fully consider the influence of environmental factors, and results in low accuracy of pain assessment. By combining multi-modal physiological signals and using a dynamic stress vector update mechanism, this method can monitor the physiological stress response of an individual in real time and visualize the change of the stress state through a physiological trajectory sphere model, thus achieving a more accurate and reliable pain grading.

[0040] Further, the dynamic feature acquisition module 14 is further configured to: use the vascular bifurcation point matrix as a reference feature to perform cross-modal spatial registration of the temporal muscle contraction distribution and the temporal temperature field distribution, and output a registered muscle contraction distribution and a registered temperature field distribution; calculate the elastic modulus change rate of the registered muscle contraction distribution, locate the peak point of the change rate as the instantaneous extreme value of muscle contraction; perform peak decay tracking starting from the peak point of the change rate, and output a contraction recovery delay window, where the contraction recovery delay window and the instantaneous extreme value of muscle contraction constitute the ORI correlation feature; start from the center of the target area, traverse the registered temperature field distribution to calculate and output the regional temperature difference gradient; perform metabolic heat production prediction on the registered temperature field distribution, and output the regional heat production rate, where the regional temperature difference gradient and the regional heat production rate constitute the regional thermodynamics feature.

[0041] After spatially registering the temporal muscle contraction distribution and the temporal temperature field distribution, dynamic feature extraction is performed to output ORI correlation features and regional thermodynamic features. The method includes: using the vascular bifurcation point matrix as the reference feature to perform cross-modal spatial registration of the temporal muscle contraction distribution and the temporal temperature field distribution, so that the data of the temporal muscle contraction distribution and the temporal temperature field distribution are aligned in the spatial coordinates, and then outputting the registered muscle contraction distribution and the registered temperature field distribution. Further, calculate the rate of change of the elastic modulus of the registered muscle contraction distribution, and locate the peak point of the rate of change as the instantaneous extreme value of muscle contraction. These peak points can be used to measure the intensity of pain or stress. Suppose that during the monitoring process, the elastic modulus of the muscle changes from 10 kPa to 25 kPa within 0.5 seconds, and the rate of change of the elastic modulus is 30 kPa / second. Through calculation, the peak point of the instantaneous extreme value is obtained and expressed as the maximum muscle tension at a certain moment. Starting from the peak point of the rate of change, peak decay tracking is performed. Peak decay refers to the recovery process after muscle contraction reaches the maximum value, and the recovery time and intensity can be obtained. Output the contraction recovery delay window. The contraction recovery delay window refers to the time required from the instantaneous extreme value of muscle contraction to the muscle contraction recovering to a certain threshold. This window can reflect the recovery ability of the muscle and thus provide a basis for pain assessment. Among them, the contraction recovery delay window and the instantaneous extreme value of muscle contraction constitute the ORI correlation features. Starting from the center of the target area, traverse the registered temperature field distribution to calculate and output the regional temperature difference gradient. The calculation of the temperature difference gradient is obtained through the temperature difference between adjacent points, which reflects the speed and direction of heat transfer. Based on the registered temperature field data and the calculated temperature difference gradient, a thermodynamic model is applied for metabolic heat production prediction. The thermodynamic model is constructed based on a neural network model. By inputting the registered temperature field data corresponding temperature difference gradient and the metabolic heat production identification result obtained from historical collection into the untrained model to perform supervised training of the model, when the model converges, the thermodynamic model is obtained. By inputting the registered temperature field data and the calculated temperature difference gradient into the thermodynamic model. The heat production rate of the area is predicted by the model to obtain the regional thermodynamic features. Among them, the regional temperature difference gradient and the regional heat production rate constitute the regional thermodynamic features.

[0042] Furthermore, the stress grading module 15 is further configured to: locally de-privatize and call multiple sample multi-modal physiological features and multiple sample stress gradings; establish the physiological stress grading space according to the standard index composition of the sample multi-modal physiological features; after positioning multiple sample particle points in the physiological stress grading space based on the multiple sample multi-modal physiological features, use the multiple sample stress gradings to label the multiple sample particle points to complete the data filling of the physiological stress grading space; input the ORI correlation feature, regional thermodynamic feature, and coupled fluctuation correlation feature into the physiological stress grading space to locate virtual particle points; perform spatial matching of the virtual particle points with the goal of minimizing the distance, locate the nearest particle point, and use the sample stress grading of the nearest particle point as the initial stress grading.

[0043] By locally taking privacy protection measures to call multiple sample multi-modal physiological features and multiple sample stress gradings, the security of the user's biological data and personal information is ensured. The sample multi-modal physiological features include sample ORI features, sample thermodynamic features, and sample coupling features. The sample stress grading is the stress level of the reference sample under the condition of the corresponding sample multi-modal physiological features. Stress grading is usually divided into different levels according to the intensity of physiological reactions. For example, mild corresponds to levels 0-3, moderate corresponds to levels 4-6, and severe stress reactions correspond to levels 7-10. According to the standard index composition of the sample multi-modal physiological features, that is, establish the physiological stress grading space according to the sample ORI features, sample thermodynamic features, and sample coupling features, and transform the physiological features of each individual into a point in the stress grading space. After positioning multiple sample particle points in the physiological stress grading space based on the multiple sample multi-modal physiological features, use the multiple sample stress gradings to label the multiple sample particle points to complete the data filling of the physiological stress grading space. Further, input the obtained ORI correlation feature, regional thermodynamic feature, and coupled fluctuation correlation feature of the target user into the physiological stress grading space to locate virtual particle points. Perform spatial matching of the virtual particle points with the goal of minimizing the distance, obtain the nearest particle point of the virtual particle point in the physiological stress grading space, and use the sample stress grading of the nearest particle point as the initial stress grading.

[0044] Furthermore, the pain grading module 16 is further configured to: construct a real-time stress vector according to the ORI correlation feature, regional thermodynamic feature, and coupled fluctuation correlation feature; preset a stress monitoring window, and dynamically update the real-time stress vector through the stress monitoring window to output multiple incremental stress vectors; construct a physiological trajectory sphere model based on the multiple incremental stress vectors, and perform visual rendering of the physiological trajectory sphere model using multiple incremental stress gradings.

[0045] Construct a real-time stress vector based on the ORI correlation feature, regional thermodynamic feature, and coupled fluctuation correlation feature, and use a vector to represent the corresponding feature parameters. Subsequently, preset a stress monitoring window, which is a fixed time monitoring window set in advance, and dynamically update the real-time stress vector through the stress monitoring window to obtain the stress vectors with vector updates, and output multiple incremental stress vectors. Finally, construct a physiological trajectory sphere model based on the multiple incremental stress vectors. The physiological trajectory sphere model is used to represent the dynamic changes of the stress state on the time axis. Each incremental stress vector will be represented as a point in the three-dimensional space of the sphere model, and the connection between points shows the change trajectory of the stress state. Use multiple incremental stress gradings for visual rendering of the physiological trajectory sphere model, that is, map the incremental stress gradings to different regions of the sphere model, which can intuitively display the changes in the stress state. In the sphere model, different stress gradings are represented by different colors or sphere sizes. Low stress is represented by green, medium stress is represented by yellow, and high stress is represented by red.

[0046] In the embodiment of the present application, there is a contraction distribution acquisition module for capturing the muscle tissue contraction of the target area to obtain a time-series muscle contraction distribution; a temperature field distribution acquisition module for calculating an initial temperature field distribution according to the thermal imaging of the target area and performing temperature compensation according to the basic time-series body temperature transmitted back by the ear temperature sensor, and outputting a time-series temperature field distribution; a correlation feature acquisition module for docking standard medical devices to call real-time physiological data, performing physiological fluctuation correlation analysis, and outputting a coupled fluctuation correlation feature; a dynamic feature acquisition module for performing dynamic feature extraction after spatially registering the time-series muscle contraction distribution and the time-series temperature field distribution, and outputting an ORI correlation feature and a regional thermodynamic feature; a stress grading module for inputting the ORI correlation feature, regional thermodynamic feature, and coupled fluctuation correlation feature into a physiological stress grading space to match and locate the initial stress grading; a pain grading module for dynamically compensating the threshold of the initial stress grading according to the basic time-series body temperature and outputting a real-time physiological pain grading. This solves the technical problem in the prior art that the pain assessment method lacks accurate real-time physiological data analysis and dynamic update mechanism, and cannot fully consider the influence of environmental factors, resulting in low accuracy of pain assessment. By combining multimodal physiological signals and using a dynamic stress vector update mechanism, this method can monitor the physiological stress response of an individual in real time and visualize the changes in the stress state through a physiological trajectory sphere model, thereby achieving more accurate and reliable pain grading.

[0047] The above specific embodiments do not constitute a limitation on the protection scope of the present application. Those skilled in the art should understand that various modifications, combinations, and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principle of the present application shall be included within the protection scope of the present application. In some cases, the actions or steps recited in the present application can be executed in a different order than in the embodiments and still achieve the desired results. Additionally, the processes depicted in the drawings do not necessarily require the specific order or sequential order shown to achieve the desired results. In certain embodiments, multitasking and parallel processing are also possible or may be advantageous.

Claims

1. A pain grading evaluation system based on multimodal physiological signal fusion, characterized in that, The system includes: a contraction distribution acquisition module, configured to capture the contraction of muscle tissues in a target area to obtain a time-series muscle contraction distribution; a temperature field distribution acquisition module, configured to calculate an initial temperature field distribution based on the thermal imaging of the target area, and perform temperature compensation according to the basic time-series body temperature transmitted back by an ear temperature sensor, and output a time-series temperature field distribution; a correlation feature acquisition module, configured to call real-time physiological data through a standard medical device, perform physiological fluctuation correlation analysis, and output coupled fluctuation correlation features; a dynamic feature acquisition module, configured to perform dynamic feature extraction after spatially registering the time-series muscle contraction distribution and the time-series temperature field distribution, and output ORI correlation features and regional thermodynamics features; a stress grading module, configured to input the ORI correlation features, the regional thermodynamics features, and the coupled fluctuation correlation features into a physiological stress grading space to match and locate an initial stress grade; a pain grading module, configured to perform dynamic threshold compensation on the initial stress grade according to the basic time-series body temperature, and output a real-time physiological pain grade.

2. The pain grading evaluation system based on multi-modal physiological signal fusion according to claim 1, wherein The pain grading module is further configured to: construct a real-time stress vector according to the ORI correlation features, the regional thermodynamics features, and the coupled fluctuation correlation features; preset a stress monitoring window, and perform dynamic update of the real-time stress vector through the stress monitoring window to output a plurality of incremental stress vectors; construct a physiological trajectory sphere model according to the plurality of incremental stress vectors, and perform visual rendering of the physiological trajectory sphere model by using a plurality of incremental stress gradings.

3. The pain grading evaluation system based on multi-modal physiological signal fusion according to claim 1, characterized in that, The correlation feature acquisition module is further configured to: acquire respiratory rate and blood pressure parameters in real time through a standard medical device interface to obtain respiratory rate time-series data and blood pressure parameter time-series data; perform phase synchronization analysis on the respiratory rate time-series data and the blood pressure parameter time-series data to construct a respiration-blood pressure coupling sequence; perform sliding window cross-correlation analysis on the respiration-blood pressure coupling sequence, and output the coupled fluctuation correlation features.

4. The pain grading evaluation system based on multimodal physiological signal fusion according to claim 3, wherein The correlation feature acquisition module is further configured to: construct sliding window parameters according to a standard respiratory cycle, where the sliding window parameters include a window length and a sliding step, the window length is H standard respiratory cycles, and the sliding step is 1 / M standard respiratory cycle; slide and segment the respiration-blood pressure coupling sequence based on the sliding window parameters to output a plurality of window coupling sequences; adopt a normalized cross-correlation function to calculate a plurality of cross-correlation coefficients and a plurality of coupling time delays of the plurality of window coupling sequences; sort the plurality of cross-correlation coefficients in ascending order, extract the maximum cross-correlation coefficient, and correspondingly call a target coupling time delay, where the maximum cross-correlation coefficient and the target coupling time delay constitute the coupled fluctuation correlation features.

5. The pain grading evaluation system based on multimodal physiological signal fusion according to claim 1, characterized in that The contraction distribution acquisition module is further configured to: during the process of irradiating the target area in a pulse alternating mode by using an 850nm near-infrared light source and a 520nm visible light source, synchronously capture the dual-wavelength reflected light intensity distribution through a shutter sensor to generate a reflected light intensity alternating sequence; calculate and output an intensity difference sequence based on the reflected light intensity alternating sequence; input the intensity difference sequence into a tissue deformation correlation model to perform local deformation inversion to obtain a local deformation amount sequence; Perform elastic modulus inversion on the local deformation quantity sequence and output an instantaneous elastic modulus sequence; Calculate the change rate of the sliding window for the instantaneous elastic modulus sequence and output the time-series change curve of the elastic modulus as the time-series muscle contraction distribution.

6. The pain grading evaluation system based on multimodal physiological signal fusion according to claim 1, wherein The temperature field distribution acquisition module is further configured to: Configure a far-infrared thermal imager according to the spatial characteristics of the target area, where the target area is at the center of the field of view of the far-infrared thermal imager; Capture the original temperature of the target area based on the far-infrared thermal imager to obtain multiple frames of original temperature matrices.

7. The pain grading evaluation system based on multimodal physiological signal fusion according to claim 6, wherein The temperature field distribution acquisition module is further configured to: Collect continuous resting state data of the target user through the ear temperature sensor to construct the basic time-series body temperature; Locate the blood vessel bifurcation point matrix in the target area; After spatially aligning the multiple frames of original temperature matrices with the blood vessel bifurcation point matrix as the reference feature, calculate the gradient distribution of adjacent frames of original temperature matrices and output the initial temperature field distribution; Perform dynamic temperature compensation on the initial temperature field distribution according to the basic time-series body temperature and output the time-series temperature field distribution.

8. The pain grading evaluation system based on multimodal physiological signal fusion according to claim 7, wherein The dynamic feature acquisition module is further configured to: Perform cross-modal spatial registration on the time-series muscle contraction distribution and the time-series temperature field distribution with the blood vessel bifurcation point matrix as the reference feature and output the registered muscle contraction distribution and the registered temperature field distribution; Calculate the change rate of the elastic modulus for the registered muscle contraction distribution and locate the peak point of the change rate as the instantaneous extreme value of muscle contraction; Perform peak decay tracking starting from the peak point of the change rate and output the contraction recovery delay window, where the contraction recovery delay window and the instantaneous extreme value of muscle contraction constitute the ORI correlation feature; Starting from the center of the target area, traverse the registered temperature field distribution to calculate and output the regional temperature difference gradient; Perform metabolic heat production prediction on the registered temperature field distribution and output the regional heat production rate, where the regional temperature difference gradient and the regional heat production rate constitute the regional thermodynamics feature.

9. The pain grading evaluation system based on multi-modal physiological signal fusion according to claim 1, characterized in that, The stress grading module is further configured to: Locally de-privatize and call multiple sample multi-modal physiological features and multiple sample stress gradings; Establish the physiological stress grading space according to the standard index composition of the sample multi-modal physiological features; After locating multiple sample particle points in the physiological stress grading space based on the multiple sample multi-modal physiological features, use the multiple sample stress gradings to label the multiple sample particle points to complete the data filling of the physiological stress grading space; Input the ORI correlation feature, the regional thermodynamics feature, and the coupled fluctuation correlation feature into the physiological stress grading space to locate virtual particle points; Perform spatial matching of the virtual particle points with the goal of minimizing the distance, locate the nearest particle point, and use the sample stress grading of the nearest particle point as the initial stress grading.

10. The pain grading evaluation system based on multimodal physiological signal fusion according to claim 9, wherein The sample multi-modal physiological features include sample ORI features, sample thermodynamics features, and sample coupling features.

Images