A Multi-Source Vital Sign Co-matching System and Method for Emotional Health Grading

By constructing a multi-source vital sign collaborative matching emotional health classification system, dynamically tracking the inflection point of vital sign channel response, calculating the product of time sequence difference and arousal rate features, and adaptively adjusting the analysis window and model weights, the system solves the problems of inconsistent vital sign signal response and individual differences in existing systems, and achieves high-precision emotional health classification.

CN123101142APending Publication Date: 2026-11-17CHUANGYI (SHANGHAI) INFORMATION TECH CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202611392546.5
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2026-09-09
Publication Date
2026-11-17

AI Technical Summary

Technical Problem

Existing multi-source vital sign emotion determination systems assume that the response time of vital sign signals is consistent, which leads to the delay of fast response signals and the truncation of slow response signals. They cannot distinguish between sudden emotional fluctuations and persistent emotional states, and lack the ability to adapt to individual differences and long-term physiological states, resulting in a decrease in determination accuracy.

Method used

A dynamic hierarchical judgment mechanism coupled with the product of the response time difference of vital signs and the emotional arousal rate is constructed. By dynamically tracking the response inflection point of each vital sign channel, the response time difference between any two channels is calculated, and an exclusive analysis window is adaptively determined for each channel. The judgment model is optimized by combining a feedback correction mechanism.

Benefits of technology

It significantly improves the ability to distinguish between sudden and persistent emotional states, increases the accuracy of judgment, adapts to individual differences and long-term physiological state drift, and achieves dynamic and differentiated classification judgment around the clock.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN123101142A_ABST
    Figure CN123101142A_ABST
Patent Text Reader

Abstract

This invention discloses a multi-source vital sign collaborative matching emotional health grading system and method, belonging to the field of multi-source vital sign signal processing and emotional health assessment technology. It includes: a multi-source vital sign signal acquisition module, used to acquire multiple types of vital sign signals of the assessed person over a continuous time stream, obtaining real-time signal sequences of each vital sign channel; and a response inflection point tracking module, used to dynamically track the response inflection point time of each vital sign channel over a continuous time stream. The response inflection point is the turning point where the vital sign signal changes from stable to significantly rising or falling under emotional stimulation or emotional state changes. The product coupling structure ensures that the judgment must be based on the common changes of two dimensions simultaneously; an effective grading conclusion cannot be formed when either dimension exists alone. This fundamentally overcomes the systematic destruction of temporal information by a unified time window and significantly improves the ability to distinguish between sudden and persistent emotional states.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This invention relates to the field of multi-source vital sign signal processing and emotional health assessment, specifically to a multi-source vital sign collaborative matching emotional health grading system and method. Background Technology

[0002] Emotional health grading has wide applications in areas such as mental health monitoring, human-computer interaction, and occupational stress assessment. Emotion recognition methods based on physiological signs have become a research hotspot in recent years due to their advantages such as objectivity and continuous monitoring. Common multi-source physiological signs include heart rate, skin conductance, respiration, and electromyography (EMG), which reflect the activity state of the autonomic and central nervous systems from different perspectives.

[0003] Existing multi-source vital sign emotion determination systems typically treat each vital sign signal as a parallel data stream with synchronous response and change. Statistical features are extracted within a unified time window, and then the features from each channel are concatenated or simply fused before being input into a classification model to obtain the emotion state determination result. This approach is based on an implicit assumption: all vital sign channels have consistent response delays and identical response patterns to the same emotional stimulus.

[0004] However, this assumption deviates significantly from actual physiological processes. Different physiological signals exhibit fundamentally different response rates to emotional stimuli: skin conductance typically peaks 1 to 3 seconds after emotional stimulation, heart rate changes fully manifest in about 3 to 10 seconds, and respiratory rate regulation can last for tens of seconds. The fixed-time-window synchronous feature extraction method results in the delay of rapid response signals and the truncation of slow response signals, completely obliterating the most crucial temporal structure information during emotional arousal. Especially when distinguishing between "sudden emotional fluctuations" and "persistent emotional states," the synchronous processing method cannot utilize the crucial temporal information of "which signals respond first, which signals respond later, and by how much?" This leads to a high degree of overlap between the two states in the feature space, limiting the accuracy of the judgment.

[0005] Furthermore, existing multi-source vital sign emotion assessment systems typically employ fixed models trained only once, lacking the ability to adapt to individual differences and long-term physiological state drift. Different individuals naturally exhibit variations in their baseline physiological parameters and vital sign signal response patterns. Even for the same user, changes in daily routines, health status, and emotional regulation abilities can cause their vital sign response characteristics to slowly drift. While fixed models may maintain a certain level of accuracy initially, their accuracy declines over time, lacking continuous learning and dynamic correction mechanisms. Summary of the Invention

[0006] This invention provides a multi-source vital sign collaborative matching emotional health grading system and method. Addressing the problem that existing multi-source vital sign emotion assessment systems treat multiple vital sign signals such as heart rate, skin conductance, respiration, and electromyography as parallel data streams with synchronous responses and changes, directly inputting statistical features into a classification model after extraction within a unified time window, leads to the lag of rapid response signals, the truncation of slow response signals, and the complete obliteration of the fine temporal characteristics of emotional arousal, making it impossible to distinguish between sudden emotional fluctuations and persistent emotional states. This solution constructs a dynamic grading mechanism coupling the product of vital sign response temporal difference and emotional arousal rate. It dynamically tracks the response inflection point of each vital sign channel in a continuous time stream, calculates the response temporal difference between any two channels, and simultaneously extracts the current emotional arousal rate. The product of the temporal difference feature and the arousal rate feature is used as the grading basis, so that the judgment process no longer relies on a unified time window, but adaptively determines a dedicated effective analysis window for each channel based on the response temporal difference, achieving coupled judgment where the temporal difference reveals the emotional type and the arousal rate reveals the emotional intensity.

[0007] To address the shortcomings of existing multi-source emotional assessment systems, such as insufficient utilization of response time difference information from various channels, lack of dynamic adjustment mechanisms for channel coupling contribution weights, and inability of the assessment model to continuously adapt and evolve with individual user differences and long-term physiological state drift, this solution constructs a dynamic correction mechanism for response time difference-awakening rate coupling weights. The system continuously records the consistency between each assessment result and subsequent psychological scale review results, and corrects the coupling contribution weights of each channel pair and the response inflection point detection parameters online based on consistency feedback. This enables the assessment model to continuously adapt and optimize with individual user differences and long-term physiological state drift, forming a closed-loop adaptive system of time difference extraction, awakening rate coupling, graded assessment, and feedback correction. Ultimately, this achieves all-weather dynamic differentiated graded assessment of sudden and persistent emotional health states.

[0008] To achieve the above objectives, the present invention provides the following technical solution: including:

[0009] The multi-source vital sign signal acquisition module is used to acquire multiple types of vital sign signals of the subject in a continuous time stream and obtain the real-time signal sequence of each vital sign channel;

[0010] The response inflection point tracking module is used to dynamically track the response inflection point of each vital sign channel in a continuous time stream. The response inflection point is the turning point where the vital sign signal changes from stable to significantly rising or falling under emotional stimulation or changes in emotional state.

[0011] The response timing difference calculation module is used to calculate the response timing difference between any two vital signs channels based on the response inflection point time of each vital sign channel. The response timing difference reflects the difference in the order of response of different vital sign channels to the same emotional state change.

[0012] The dedicated analysis window segmentation module is used to adaptively determine the dedicated effective analysis window for each vital sign channel based on the response time difference of each channel, so that each channel is extracted with features within the time period when its response is fully manifested, rather than being forcibly aligned within a uniform time window.

[0013] The emotional arousal rate extraction module is used to extract the signal change features of each channel in the dedicated analysis window of each vital sign channel, and to weight and synthesize the slope of the signal changes of multiple channels to obtain the current emotional arousal rate.

[0014] The coupled grading determination module is used to use the product of response time difference feature and emotional arousal rate feature as the grading determination basis. It performs coupled calculation of response time difference feature and emotional arousal rate feature to output the emotional health grading determination result of the evaluated person. Among them, response time difference feature reveals the emotion type and emotional arousal rate feature reveals the emotion intensity.

[0015] As a further preferred embodiment of this technical solution: the classification judgment basis generated by the coupled classification judgment module is composed of the product of the response timing difference and the emotional arousal rate. When either the response timing difference or the emotional arousal rate approaches zero, the product value approaches zero. The coupled classification judgment module outputs a high judgment score only when both have significant values. This means that the classification judgment must be based on the common changes of both the response timing difference and the emotional arousal rate.

[0016] As a further preferred embodiment of this technical solution: when determining the dedicated analysis window for each vital sign channel, the dedicated analysis window division module dynamically adjusts the starting position and length of the window according to the current emotional arousal rate: the higher the arousal rate, the shorter the window length and the closer the starting position is to the response inflection point; the lower the arousal rate, the longer the window length and the wider the coverage area, so that the window division matches the dynamic changes in the arousal rate.

[0017] As a further preferred embodiment of this technical solution, it also includes a feedback correction module, which is used to continuously record the consistency between each grading judgment result and the subsequent psychological scale review results, and correct the coupling contribution weight of each channel pair and the response inflection point detection parameters online based on the consistency feedback, so that the grading judgment model can continuously adaptively optimize with the individual differences of the assessed person and the long-term physiological state drift.

[0018] As a further preferred embodiment of this technical solution: when the feedback correction module adaptively adjusts the coupling contribution weight of each channel pair, it dynamically increases or decreases the weight of the channel pair according to the contribution of the channel pair to the discrimination of different emotion types in historical judgments: for channel pairs with high discrimination contribution, the coupling contribution weight is increased; for channel pairs with low discrimination contribution or poor consistency with the judgment result, the coupling contribution weight is decreased.

[0019] As a further preferred embodiment of this technical solution, the following steps are included:

[0020] Multiple vital signs signals of the subject are collected over a continuous time stream to obtain real-time signal sequences for each vital sign channel;

[0021] The response inflection point of each vital sign channel is dynamically tracked. The response inflection point is the turning point where the vital sign signal changes from stable to significantly rising or falling under emotional stimulation or change in emotional state.

[0022] Based on the inflection point of each vital sign channel, calculate the response timing difference between any two vital sign channels. The response timing difference reflects the difference in the order of response of different vital sign channels to the same emotional state change.

[0023] Based on the response time difference of each vital sign channel, an exclusive effective analysis window is adaptively determined for each vital sign channel, so that each channel is extracted with features within the time period when its response is fully manifested, rather than being forcibly aligned within a uniform time window.

[0024] The signal change characteristics of each channel are extracted in the dedicated analysis window of each vital sign channel, and the slope of the signal change of multiple channels is weighted and synthesized to obtain the current emotional arousal rate.

[0025] The product of response time difference feature and emotional arousal rate feature is used as the basis for classification. The response time difference feature and emotional arousal rate feature are coupled and calculated to output the emotional health classification result of the evaluated person. Among them, the response time difference feature reveals the emotion type and the emotional arousal rate feature reveals the emotion intensity.

[0026] As a further preferred embodiment of this technical solution: the classification judgment is based on the product of the response timing difference and the emotional arousal rate. When either the response timing difference or the emotional arousal rate approaches zero, the product value approaches zero. A high judgment score is output only when both have significant values, so that the classification judgment must be based on the common changes of the two dimensions of response timing difference and emotional arousal rate.

[0027] As a further preferred embodiment of this technical solution: in the process of adaptively determining a dedicated analysis window for each vital sign channel, the starting position and length of the window are dynamically adjusted according to the current emotional arousal rate: the higher the arousal rate, the shorter the window length and the closer the starting position is to the response inflection point; the lower the arousal rate, the longer the window length and the wider the coverage area, so that the window division matches the dynamic changes in the arousal rate.

[0028] As a further preferred embodiment of this technical solution, the following steps are also included: continuously recording the consistency between each grading judgment result and the subsequent psychological scale review results, and correcting the coupling contribution weight of each channel pair and the response inflection point detection parameters online based on the consistency feedback, so that the grading judgment model can continuously adaptively optimize with the individual differences of the assessed person and the long-term physiological state drift.

[0029] As a further preferred embodiment of this technical solution: before outputting the emotional health grading result of the assessed person, the method further includes: classifying the current emotional state into two categories, sudden emotional fluctuation and persistent emotional state, based on the magnitude and distribution pattern of the response time difference; wherein, sudden emotional fluctuation corresponds to the state where the time difference between the inflection points of each channel response is less than a preset threshold, and persistent emotional state corresponds to the state where the time difference between the inflection points of each channel response is greater than a preset threshold; the grading result includes both an emotional type identifier and an emotional intensity level.

[0030] Compared with the prior art, the beneficial effects of the present invention are:

[0031] 1. Existing systems treat multi-source vital signs as synchronous responses and extract features within a unified time window. This leads to the lag of rapid response signals and the truncation of slow response signals, making it impossible to distinguish between sudden emotional fluctuations and persistent emotional states. This solution constructs a dynamic grading mechanism that couples the product of the vital sign response time difference and the emotional arousal rate. It dynamically tracks the inflection point of each vital sign channel in a continuous time stream, calculates the response time difference between any two channels, and simultaneously extracts the current emotional arousal rate. The product of the time difference feature and the arousal rate feature is used as the grading criterion. The time difference feature reveals the emotion type, while the arousal rate feature reveals the emotion intensity. The product-coupled structure ensures that the judgment must be based on the simultaneous changes of both dimensions; neither dimension alone can form an effective grading conclusion. This fundamentally overcomes the systematic destruction of temporal information by the unified time window and significantly improves the ability to distinguish between sudden and persistent emotional states.

[0032] 2. Existing systems extract features from all channels synchronously within a fixed time window. Fast and slow channels are analyzed within incorrect time intervals, resulting in the smoothing and dilution of peak information for fast signals and the truncation of the complete change process for slow signals. This proposed solution adaptively determines a dedicated effective analysis window for each channel based on the response timing difference, and dynamically adjusts the starting position and length of the window according to the current emotional arousal rate. This ensures that features are extracted from each channel within the time period during which its response is fully manifested. This dynamic windowing method guarantees that fast channels are accurately captured near their peaks, and slow channels are fully observed within their complete change range. The quality and representativeness of features from each channel are simultaneously improved, providing a reliable feature foundation for subsequent coupling judgment.

[0033] 3. Existing multi-source emotion assessment systems use fixed models trained only once, which cannot adapt to individual differences and long-term physiological state drift. The accuracy of these assessments continuously declines over time after deployment. This solution constructs a feedback correction mechanism that continuously records the consistency between each grading assessment result and subsequent psychological scale review results. Based on the consistency feedback, it online corrects the coupling contribution weights of each channel pair and the response inflection point detection parameters. Channel pairs with high discriminative contributions are automatically weighted, while those with poor consistency are automatically weighted, allowing the assessment model to continuously adapt and evolve according to individual user differences and long-term physiological state drift. Through this feedback loop, the system continuously approaches the individual's optimal assessment state during long-term operation. The accuracy of assessing sudden and persistent emotional health states continuously improves with usage time, achieving a paradigm shift from a static model to a dynamic adaptive model. Attached Figure Description

[0034] Figure 1 This is the overall main flowchart of the multi-source vital sign collaborative matching emotional health grading system and method of the present invention;

[0035] Figure 2 This is a flowchart illustrating the coupled grading determination process of a multi-source vital sign collaborative matching emotional health grading determination system and method according to the present invention.

[0036] Figure 3 This is a flowchart illustrating the dynamic division of the dedicated analysis window in the multi-source vital sign collaborative matching emotional health grading system and method of the present invention.

[0037] Figure 4 This is a flowchart illustrating the feedback correction and adaptive optimization process of a multi-source vital sign collaborative matching emotional health grading system and method according to the present invention. Detailed Implementation

[0038] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0039] Example

[0040] Please see Figures 1-4 This is a schematic diagram of some embodiments of a multi-source vital sign collaborative matching emotional health grading system and method used in this application.

[0041] Example 1

[0042] This embodiment provides a multi-source vital sign collaborative matching emotional health grading system, which is deployed on a wearable physiological signal acquisition device and a matching mobile computing terminal. The system includes a multi-source vital sign signal acquisition module, a response inflection point tracking module, a response time difference calculation module, a dedicated analysis window division module, an emotional arousal rate extraction module, and a coupled grading determination module.

[0043] The multi-source vital sign signal acquisition module is used to acquire multiple types of vital sign signals of the subject over a continuous time stream, obtaining real-time signal sequences for each vital sign channel. In this embodiment, the vital sign signals include four types: heart rate signal, skin conductance signal, respiratory signal, and electromyography signal, corresponding to the heart rate channel, skin conductance channel, respiratory channel, and electromyography channel, respectively. The acquisition frequency is 100Hz for all signals.

[0044] The response inflection point tracking module is used to dynamically track the response inflection point moments of each vital sign channel over a continuous time stream. The response inflection point is the turning point where a vital sign signal changes from stable to significantly increasing or decreasing under emotional stimulation or a change in emotional state. In this embodiment, the signals of each vital sign channel are first subjected to Savitzky-Golay smoothing filtering with a window width of 0.5 seconds, and then the first and second order difference approximate derivatives are calculated. Within each sliding window, the window length is 2 seconds, and the sliding step size is 0.2 seconds. The signal is considered significant if the absolute value of the first derivative exceeds a preset threshold. Furthermore, the location where the sign of the second derivative changes is determined as the inflection point of the response. Significant change threshold. The absolute value of the derivative under historical calm conditions for each channel is taken as three times. This response inflection point detection parameter is continuously corrected by the feedback correction module during use.

[0045] The response timing difference calculation module is used to calculate the response timing difference between any two vital signs channels based on the inflection point time of each channel's response. Let the first channel be... The response inflection point time of each channel is , No. The response inflection point time of each channel is The response timing difference between the two channels is defined as:

[0046] In the formula: For the first Individual characteristics channel and the first The response timing difference between individual characteristic channels, in seconds; For the first The inflection point of an individual's response to changes in current emotional state; For the first The inflection point of an individual's response to changes in their current emotional state.

[0047] The response timing difference reflects the difference in the order in which different vital sign channels respond to the same change in emotional state. In this embodiment, the skin conductance channel usually responds first, followed by the heart rate channel, and the respiratory channel responds the slowest.

[0048] The dedicated analysis window partitioning module adaptively determines a dedicated effective analysis window for each vital sign channel based on the response timing differences of each channel. The module's working logic is as follows: for channels with earlier response inflection points, the analysis duration is extended backward from that inflection point by a duration matching the channel's response characteristics; for channels with later response inflection points, the analysis window is similarly determined starting from its own response inflection point. Each channel's window is determined independently on its own timeline and is not forcibly aligned with others.

[0049] The emotional arousal rate extraction module extracts the signal change characteristics of each channel within its dedicated analysis window. For each channel, the average value of the signal slope within its dedicated analysis window is calculated as the slope of that channel. Then, the slopes of the multi-channel signals are weighted and synthesized to obtain the current emotional arousal rate. ;

[0050] In the formula: This represents the current emotional arousal rate. In this embodiment, the total number of vital sign channels is [number]. ; For the first The average slope of change of an individual characteristic channel within its dedicated analysis window; For the first The arousal contribution weight of the individual characteristic channel is pre-set based on the sensitivity of the channel signal changes to emotional arousal, and satisfies the following conditions: .

[0051] The coupled grading module uses the product of the response timing difference feature and the emotional arousal rate feature as the grading criterion. It couples the response timing difference feature and the emotional arousal rate feature to calculate and output the emotional health grading result of the assessed individual. In this embodiment, the response timing difference feature is represented by the timing synchronization index. For any channel pair... The timing synchronization index is defined as:

[0052] ;

[0053] For the first The and the first The temporal synchronization index between individual feature channels takes a value range of (0,1]; This is the absolute value of the timing difference between the responses of the two channels; This is a timing-sensitive constant, and in this embodiment, it is set to 3 seconds.

[0054] when hour, This indicates that the two channels are responding almost synchronously; when When it increases, A smaller value indicates a greater difference in response time between the two channels, and a lower degree of synchronization.

[0055] The coupling determination index is defined as: ;

[0056] In the formula: This is the coupling determination index, used to drive hierarchical determination; For the first The and the first The temporal synchronization index between individual symptom channels; This represents the current emotional arousal rate. For channel pair The coupling contribution weight is adaptively determined based on the channel's contribution to the discrimination of different emotion types in historical data.

[0057] In this coupling determination index, In response to time difference characteristics, The emotional arousal rate is a characteristic, and the two are coupled in a product form. When either the response timing difference or the emotional arousal rate approaches zero, the product value approaches zero. The coupling grading module only outputs a high score when both are significant. The grading module is based on... The numerical value of emotional health status is used to classify emotional health status into four levels: normal, mildly fluctuating, moderately fluctuating, and severely fluctuating.

[0058] Example 2

[0059] This embodiment further illustrates the specific logic of product coupling in the coupling hierarchical determination module.

[0060] The grading criteria generated by the coupled grading determination module are composed of the product of the response timing difference feature and the emotional arousal rate feature. In this embodiment, the response timing difference feature is the timing synchronization index. The emotional arousal rate is characterized by .

[0061] Under this product structure, when the response timing difference characteristic approaches zero, i.e. The product value approaches zero; when the emotional arousal rate When the product value approaches zero, the product value also approaches zero. Only when both the synchronization index and the wake-up rate have significant values ​​is the product value large enough for the coupling hierarchical determination module to output a high determination score.

[0062] The technical effect of this product structure is that the classification must be based on the simultaneous changes in both the response timing difference and the emotional arousal rate. If either dimension is missing or significant alone, an effective classification cannot be formed. For example, if there is only high synchronization (small response timing difference) but a low arousal rate, the product value will be low, and the system will not misjudge it as an emotional fluctuation; similarly, if there is only a high arousal rate but low synchronization (significantly misaligned response times across channels), the product value will also be low, and the system will not rashly output a high-level classification.

[0063] Example 3

[0064] This embodiment further illustrates the dynamic adjustment mechanism of the dedicated analysis window.

[0065] When determining the dedicated analysis window for each vital sign channel, the module dynamically adjusts the starting position and length of the window based on the current emotional arousal rate.

[0066] In this embodiment, the higher the arousal rate, the shorter the dedicated analysis window and the closer its starting position is to the response inflection point. This is because a high arousal rate corresponds to a rapidly changing emotional process, and a shorter window can more accurately capture the peak changes in the signal; at the same time, starting the window closer to the response inflection point avoids wasting window space during the calm phase before the inflection point. Conversely, a lower arousal rate results in a longer dedicated analysis window with a wider coverage area, because a low arousal rate corresponds to a slowly evolving emotional state, requiring a longer observation interval to extract stable change features.

[0067] The specific adjustment method is as follows: Let the inflection point of the response be... The dedicated analysis window length is The window starts at position ,but:

[0068] ;

[0069] ;

[0070] In the formula: For the first The length of the dedicated analysis window for each channel at the current wake-up rate; The basic window length of the channel is determined by the response characteristics of the channel. In this embodiment, the skin conductance channel is 5 seconds, the heart rate channel is 8 seconds, the respiratory channel is 15 seconds, and the electromyography channel is 4 seconds. This represents the current emotional arousal rate. The historical maximum wake-up rate is used for normalization; This is the window length adjustment factor, ranging from 0.3 to 0.6; This is the starting position of the window; This is the inflection point of the channel's response; The initial position offset adjustment constant is set to 0.5 to 1.0 seconds.

[0071] This adjustment method ensures that the window segmentation matches the dynamic changes in the wake-up rate, and features are extracted from each channel during the time period in which its response is fully manifested.

[0072] Example 4

[0073] This embodiment is based on Embodiment 1, and the system also includes a feedback correction module.

[0074] The feedback correction module is used to continuously record the consistency between each grading judgment result and the subsequent psychological scale review results, and corrects the coupling contribution weight of each channel pair and the response inflection point detection parameters online based on the consistency feedback.

[0075] The results of the psychological scale review are collected within 24 hours after the grading determination, and a standardized emotion state scale is used to review the assessed individuals. The feedback correction module compares the grading determination results output by the system with the psychological scale review results and calculates the determination accuracy rate. When the determination accuracy rate is lower than the preset lower limit of 75%, a correction operation is triggered; when the determination accuracy rate is higher than the preset upper limit of 90%, an optimization operation is triggered.

[0076] The feedback correction module performs a batch correction operation at the end of each evaluation day, or triggers immediate correction when the accuracy rate is below the preset lower limit for three consecutive evaluation cycles.

[0077] When the feedback correction module adaptively adjusts the coupling contribution weights of each channel pair, it dynamically increases or decreases the weights based on the channel pair's contribution to the discrimination of different emotion types in historical judgments. For channel pairs with high discrimination contributions, their coupling contribution weights are increased; for channel pairs with low discrimination contributions or poor consistency with the judgment results, their coupling contribution weights are decreased. The weight adjustment step is 5% to 10% of the current weight value. After adjustment, all weights are normalized to maintain a total weight sum of 1.

[0078] The feedback correction module corrects the response inflection point detection parameters as follows: when a systematic judgment bias occurs, the threshold for significant changes in the response inflection point is adjusted. The step size can be decreased by 0.1 times to improve the sensitivity of inflection point detection; or the step size can be increased by 0.1 times to reduce the false detection rate caused by noise. The correction step size is proportional to the absolute value of the deviation in judgment accuracy.

[0079] Through the above feedback correction, the grading judgment model continuously adapts and optimizes itself according to the individual differences of the assessed persons and the long-term physiological state drift.

[0080] Example 5

[0081] This embodiment provides a multi-source vital sign collaborative matching method for classifying emotional health, applied to the classification system described in the above embodiment. The method includes the following steps:

[0082] Step S1: Multi-source vital sign signal acquisition: Multiple vital sign signals of the subject are acquired over a continuous time stream to obtain real-time signal sequences for each vital sign channel. In this embodiment, four types of signals are acquired: heart rate, skin conductance, respiration, and electromyography, at a frequency of 100Hz.

[0083] Step S2, Dynamic Tracking of Response Inflection Points: Dynamically track the response inflection point moments of each vital sign channel. The response inflection point detection method adopts a joint determination method using the first and second derivatives of a sliding window. The specific detection method is the same as in Example 1 and will not be repeated here.

[0084] Step S3: Calculate the response timing difference: Based on the response inflection point time of each vital sign channel, calculate the response timing difference between any two vital sign channels. The calculation method is the same as in Example 1.

[0085] Step S4: Dedicated analysis window division: Based on the response timing difference of each vital sign channel, an exclusive effective analysis window is adaptively determined for each vital sign channel, and the starting position and length of the window are dynamically adjusted according to the current emotional arousal rate. The specific adjustment method is the same as in Example 3.

[0086] Step S5, Emotional Arousal Rate Extraction: Extract the signal change features of each channel in the dedicated analysis window of each vital sign channel, and weight and synthesize the slope of the multi-channel signal changes to obtain the current emotional arousal rate. The calculation method is the same as in Example 1.

[0087] Step S6, Coupled Grading Determination: The product of the response time difference feature and the emotional arousal rate feature is used as the grading determination basis. The response time difference feature and the emotional arousal rate feature are coupled and calculated to output the emotional health grading determination result of the evaluated person. The response time difference feature reveals the emotion type, and the emotional arousal rate feature reveals the emotion intensity.

[0088] In step S6, the calculation method of the coupling judgment index is the same as in Example 1. The grading judgment module divides the emotional health status into four levels: normal, mild fluctuation, moderate fluctuation, and severe fluctuation, based on the coupling judgment index.

[0089] Example 6

[0090] This embodiment further illustrates the logic for distinguishing emotion types in the method.

[0091] Before outputting the emotional health rating of the assessed individual in step S6, the method further includes the following steps:

[0092] Based on the magnitude and distribution pattern of the response time difference, the current emotional state is divided into two categories: sudden emotional fluctuation and persistent emotional state.

[0093] The specific distinction is as follows: Sudden emotional fluctuations correspond to a state where the sum of the absolute values ​​of the inflection point time differences of each channel's response is less than a preset threshold, which is 2 seconds; persistent emotional states correspond to a state where the sum of the absolute values ​​of the inflection point time differences of each channel's response is greater than a preset threshold, which is 6 seconds. When the sum of the absolute values ​​of the inflection point time differences of each channel's response is between 2 and 6 seconds, it is determined to be a transitional state, and further classified into adjacent categories based on the relative proportions of fast-response channels and slow-response channels in the time difference distribution.

[0094] The grading results include both an emotion type identifier and an emotion intensity level. The emotion type identifier is either sudden or persistent, and the emotion intensity level is categorized as normal, mildly fluctuating, moderately fluctuating, or severely fluctuating. For example, a single grading result might be output as sudden-moderate fluctuation or persistent-mildly fluctuating.

[0095] By combining emotion type identifiers with emotion intensity levels, the classification results can directly provide a basis for selecting subsequent emotional health intervention measures: sudden emotional fluctuations are suitable for immediate emotion regulation interventions, while persistent emotional states are suitable for medium- to long-term mental health tracking and management.

[0096] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus.

[0097] The present invention and its embodiments have been described above. This description is not restrictive, and the accompanying drawings are only one embodiment of the present invention; the actual structure is not limited thereto. In conclusion, if those skilled in the art are inspired by this description and design similar structures and embodiments without departing from the spirit of the invention, such designs should fall within the protection scope of the present invention.

[0098] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A multi-source vital sign collaborative matching emotional health grading system, characterized in that: include: The multi-source vital sign signal acquisition module is used to acquire multiple types of vital sign signals of the subject in a continuous time stream and obtain the real-time signal sequence of each vital sign channel; The response inflection point tracking module is used to dynamically track the response inflection point of each vital sign channel in a continuous time stream. The response inflection point is the turning point where the vital sign signal changes from stable to significantly rising or falling under emotional stimulation or changes in emotional state. The response timing difference calculation module is used to calculate the response timing difference between any two vital signs channels based on the response inflection point time of each vital sign channel. The response timing difference reflects the difference in the order of response of different vital sign channels to the same emotional state change. The dedicated analysis window segmentation module is used to adaptively determine the dedicated effective analysis window for each vital sign channel based on the response time difference of each channel, so that each channel is extracted with features within the time period when its response is fully manifested, rather than being forcibly aligned within a uniform time window. The emotional arousal rate extraction module is used to extract the signal change features of each channel in the dedicated analysis window of each vital sign channel, and to weight and synthesize the slope of the signal changes of multiple channels to obtain the current emotional arousal rate. The coupled grading determination module is used to determine the grading based on the product of the response time difference feature and the emotional arousal rate feature. It performs coupled calculation of the response time difference feature and the emotional arousal rate feature to output the emotional health grading determination result of the evaluated person. The response time difference feature reveals the emotion type, and the emotional arousal rate feature reveals the emotion intensity.

2. The multi-source vital sign collaborative matching emotional health grading system according to claim 1, characterized in that: The grading judgment basis generated by the coupled grading judgment module is composed of the product of the response timing difference and the emotional arousal rate. When either the response timing difference or the emotional arousal rate approaches zero, the product value approaches zero. The coupled grading judgment module outputs a high judgment score only when both have significant values. This means that the grading judgment must be based on the common changes of both the response timing difference and the emotional arousal rate.

3. The multi-source vital sign collaborative matching emotional health grading system according to claim 2, characterized in that: When determining the dedicated analysis window for each vital sign channel, the dedicated analysis window division module dynamically adjusts the starting position and length of the window according to the current emotional arousal rate: the higher the arousal rate, the shorter the window length and the closer the starting position is to the response inflection point; the lower the arousal rate, the longer the window length and the wider the coverage area, so that the window division matches the dynamic changes in the arousal rate.

4. The multi-source vital sign collaborative matching emotional health grading system according to claim 3, characterized in that: It also includes a feedback correction module, which continuously records the consistency between each grading judgment result and the subsequent psychological scale review results. Based on the consistency feedback, it corrects the coupling contribution weight of each channel pair and the response inflection point detection parameters online, so that the grading judgment model can continuously adaptively optimize with the individual differences of the assessed person and the long-term physiological state drift.

5. The multi-source vital sign collaborative matching emotional health grading system according to claim 4, characterized in that: When the feedback correction module adaptively adjusts the coupling contribution weight of each channel pair, it dynamically increases or decreases the weight of the channel pair according to the contribution of the channel pair to the discrimination of different emotion types in historical judgments: for channel pairs with high discrimination contribution, the coupling contribution weight is increased; for channel pairs with low discrimination contribution or poor consistency with the judgment result, the coupling contribution weight is decreased.

6. A method for classifying and determining emotional health through multi-source vital signs collaborative matching, characterized in that: Includes the following steps: Multiple vital signs signals of the subject are collected over a continuous time stream to obtain real-time signal sequences for each vital sign channel; The response inflection point of each vital sign channel is dynamically tracked. The response inflection point is the turning point where the vital sign signal changes from stable to significantly rising or falling under emotional stimulation or change in emotional state. Based on the inflection point of each vital sign channel, calculate the response timing difference between any two vital sign channels. The response timing difference reflects the difference in the order of response of different vital sign channels to the same emotional state change. Based on the response time difference of each vital sign channel, an exclusive effective analysis window is adaptively determined for each vital sign channel, so that each channel is extracted with features within the time period when its response is fully manifested, rather than being forcibly aligned within a uniform time window. The signal change characteristics of each channel are extracted in the dedicated analysis window of each vital sign channel, and the slope of the signal change of multiple channels is weighted and synthesized to obtain the current emotional arousal rate. The product of response time difference feature and emotional arousal rate feature is used as the basis for classification. The response time difference feature and emotional arousal rate feature are coupled and calculated to output the emotional health classification result of the evaluated person. Among them, the response time difference feature reveals the emotion type, and the emotional arousal rate feature reveals the emotion intensity.

7. The method for determining emotional health grading through multi-source vital signs collaborative matching according to claim 6, characterized in that: The grading criteria are based on the product of the response timing difference and the emotional arousal rate. When either the response timing difference or the emotional arousal rate approaches zero, the product value approaches zero. A high score is output only when both have significant values. Therefore, the grading criteria must be based on the common changes of both the response timing difference and the emotional arousal rate.

8. The method for determining emotional health grading through multi-source vital sign synergy matching according to claim 7, characterized in that: In the process of adaptively determining a dedicated analysis window for each vital sign channel, the starting position and length of the window are dynamically adjusted according to the current emotional arousal rate: the higher the arousal rate, the shorter the window length and the closer the starting position is to the response inflection point; the lower the arousal rate, the longer the window length and the wider the coverage area, so that the window division matches the dynamic changes in the arousal rate.

9. The method for determining emotional health grading through multi-source vital sign synergy matching according to claim 8, characterized in that: It also includes the following steps: The consistency between each grading assessment result and the subsequent psychological scale review results is continuously recorded. Based on the consistency feedback, the coupling contribution weight of each channel pair and the response inflection point detection parameters are corrected online, so that the grading assessment model can continuously adaptively optimize with the individual differences of the assessed person and the long-term physiological state drift.

10. The method for determining emotional health grading through multi-source vital sign synergy matching according to claim 9, characterized in that: Before outputting the emotional health grading results of the assessed person, the process further includes: classifying the current emotional state into two categories, sudden emotional fluctuations and persistent emotional states, based on the magnitude and distribution pattern of the response time difference; wherein, sudden emotional fluctuations correspond to the state where the time difference between the inflection points of each channel response is less than a preset threshold, and persistent emotional states correspond to the state where the time difference between the inflection points of each channel response is greater than a preset threshold; the grading results also include an emotion type identifier and an emotion intensity level.