Emotional state monitoring method and system based on skin resistance change

Abstract

The invention discloses an emotional state monitoring method and system based on skin resistance change, and relates to the technical field of physiological signal monitoring and emotional state recognition. Preprocessing the signals and extracting adaptive features; dynamically correcting an individual baseline based on skin resistance variation; time sequence modeling is introduced to capture the dynamic change of the emotional state; customized emotion health suggestions are provided in combination with individual historical data. The method has remarkable innovativeness and practicability in the aspects of multi-dimensional signal collection, data processing, individual correction, time sequence modeling and historical data utilization, the accuracy, stability and personalized recommendation effect of emotional state monitoring can be effectively improved, and the method has wide application prospects and market value.

Description

Technical Field

[0001] The present invention relates to the technical field of physiological signal monitoring and emotional state recognition, and particularly to an emotional state monitoring method and system based on skin resistance change. Background Art

[0002] With the rapid development of affective computing and physiological signal monitoring technologies, emotional state monitoring, as an important research direction in mental health management, smart wearable devices, and human-computer interaction, has gradually become one of the hotspots in interdisciplinary fields. Galvanic Skin Response (GSR), as a direct reflection of the activity of the autonomic nervous system, can provide physiological signals closely related to emotional states through changes in the secretion level of human sweat glands. In recent years, researchers have made certain progress in emotional state recognition, stress detection, and mental health monitoring based on skin resistance signals, and related technologies have been widely applied in the fields of medical health, smart wearable devices, and educational psychology. However, the real-time and accurate monitoring of emotional states still faces many challenges, especially in aspects such as individual differences, signal noise filtering, and long-term trend analysis, where there are still significant technical gaps.

[0003] Currently, most of the emotional state monitoring methods based on skin resistance changes rely on single-channel or single-time-point skin resistance signal acquisition. This method is difficult to capture the multi-dimensional characteristics of emotional states, resulting in low accuracy of emotional recognition. At the same time, existing technologies often use static filtering and feature extraction methods in signal preprocessing, which cannot adapt to the skin resistance signal change characteristics of different individuals in different emotional states, resulting in poor signal feature extraction effects. In addition, due to the significant differences in skin resistance baselines of different individuals in the resting state, existing technologies usually use fixed reference values for emotional state discrimination. This static baseline method cannot effectively eliminate the interference of individual physiological differences, reducing the generalization ability and classification accuracy of the model. The application of time series modeling in emotional state monitoring is also relatively limited. Most existing technologies lack dynamic modeling of emotional states over time and cannot capture the evolution law of emotional states, especially showing poor performance in the recognition of sudden emotional changes. In addition, current emotional state monitoring methods often ignore the historical emotional state data of individuals and fail to effectively use historical data for long-term trend analysis and the generation of customized health advice, resulting in the lack of targeted and personalized adjustment functions in the monitoring system. Therefore, existing technologies have significant deficiencies in multi-channel signal acquisition, adaptive feature extraction, individual baseline dynamic correction, time series modeling, and historical data application, and it is difficult to achieve accurate, dynamic monitoring and personalized intervention of emotional states. Summary of the Invention

[0004] In view of the above problems, the present invention is proposed.

[0005] Therefore, the technical problems solved by the present invention are: the existing emotional state monitoring methods have insufficient single-channel signal acquisition accuracy, poor adaptive feature extraction effect, insufficient individual baseline difference correction, and the problem of how to realize personalized emotional health advice based on time series modeling and historical data.

[0006] To solve the above technical problems, the present invention provides the following technical solutions: an emotional state monitoring method based on skin resistance change, including collecting multi-channel skin resistance signals; preprocessing the signals and extracting adaptive features; dynamically correcting the individual baseline based on skin resistance change; introducing time series modeling to capture the dynamic changes of emotional state; and providing customized emotional health advice in combination with individual historical data.

[0007] As a preferred embodiment of the emotional state monitoring method based on skin resistance change of the present invention, wherein: the collecting of multi-channel skin resistance signals includes collecting skin resistance signals by using a multi-channel skin sensor layout;

[0008] The sensor layout positions include the finger tips, the center of the palm, the inner side of the wrist, and the forearm; the multi-channel configuration adopts a distributed multi-channel architecture, including a main sensor and auxiliary sensors; the sensor nodes use a wireless transmission protocol to transmit the collected data to the data processing center in real time;

[0009] Adjust the sampling frequency according to the user's activity state, and the frequency is recommended to be set between 1Hz and 10Hz.

[0010] As a preferred embodiment of the emotional state monitoring method based on skin resistance change of the present invention, wherein: the preprocessing of the signals and the extraction of adaptive features include performing signal preprocessing, including filtering and noise reduction, signal smoothing, and signal normalization;

[0011] Filtering and noise reduction includes power frequency noise filtering, high-frequency noise suppression, and baseline drift correction;

[0012] Power frequency noise filtering is to use a 50Hz notch filter to remove power interference; high-frequency noise suppression is to use a filter with a cut-off frequency of 1Hz to 10Hz to remove high-frequency interference; baseline drift correction is to use an adaptive baseline correction algorithm to adjust the signal baseline in real time and eliminate the drift caused by monitoring;

[0013] Signal smoothing is to perform multi-scale decomposition of the original signal using wavelet transform, retain the feature signal, and remove high-frequency noise;

[0014] Perform normalization processing to map the signal amplitudes of different sensors to the same numerical range;

[0015] Adaptive feature extraction includes extracting time-domain features, frequency-domain features, non-linear features, and adaptive feature fusion.

[0016] As a preferred embodiment of the method for monitoring emotional state based on skin resistance change according to the present invention, wherein: the dynamic correction of individual baseline based on skin resistance change includes that the feature data contains the resistance change patterns of an individual in different emotional states. Due to the differences in physiological characteristics of different individuals, as well as the influence of external environment and sensor contact state factors, analyzing the emotional state using the original signal features will result in deviations;

[0017] By establishing a dynamic correction model for each individual, the differences between individual resting states are eliminated, and the real-time skin resistance data is dynamically adjusted. The mathematical model for dynamic correction of individual baseline based on skin resistance change is expressed as:

[0018]

[0019] where B(t) is the dynamically corrected value of the individual baseline at time t, which is used to adjust the current skin resistance value to eliminate individual differences, t 0 is the initial time point of correction, x(t) is the skin resistance value collected at time t, μ(t) is the baseline mean at time t, representing the skin resistance reference value in the current individual's resting state, α is the time decay coefficient, which controls the weight of historical data in the current correction, σ 2 (t) is the baseline variance at time t, which reflects the degree of data fluctuation, β is the variance weight decay coefficient, which controls the contribution of data in different time periods to the current correction, λ is the individual adaptive adjustment coefficient, which is used to balance the weights of the global baseline and real-time data, φ(x i ) is the feature transformation function, which maps the collected skin resistance data xi to the feature space, is the feature change rate, which reflects the contribution of instantaneous feature changes to the overall correction, ω i is the feature weight factor, which represents the importance of different feature dimensions, and N is the number of features within the current time window;

[0020] When B(t) > 0, it indicates that the current skin resistance baseline has increased compared to the resting state, reflecting that the individual is in an excited or tense state;

[0021] When B(t) < 0, it indicates that the current skin resistance baseline has decreased compared to the resting state, reflecting that the individual is in a relaxed or calm state;

[0022] When B(t) ≈ 0, it indicates that the current state is the resting state of the individual.

[0023] As a preferred embodiment of the emotional state monitoring method based on skin resistance changes according to the present invention, wherein: the introduction of time series modeling to capture the dynamic changes of emotional states includes individual baseline dynamic correction to eliminate the differences in physiological baselines among individuals, ensuring that the skin resistance characteristics of different individuals are compared and analyzed on the same standard. The emotional classification model construction step organically integrates time domain, frequency domain, statistical features, and non-linear features through mathematical modeling, comprehensively considering the time decay of historical data, the feature change rate, and the individual feature distribution, to achieve accurate classification of the current emotional state;

[0024] The emotional classification model is expressed as:

[0025]

[0026] Wherein, C is the emotional classification result, representing the optimal classification label of the emotional state at the current moment, y is a category in the set of emotional categories, Y is the set of all emotional categories, and x i (t) is the i-th skin resistance feature value collected at time t, and μ is the average value of skin resistance features within the time window; is the total variance of all features within the current time window, reflecting the degree of dispersion of the overall data, is the partial derivative of the feature mapping function, representing the instantaneous change rate of the k-th feature, and z k is the k-th feature parameter, N is the number of features within the time window, M is the dimension number of features within the time window, and K is the number of feature mappings;

[0027] If C > 0, the individual is currently in an excited or anxious state;

[0028] If C < 0, the individual is currently in a calm or relaxed state;

[0029] If C ≈ 0, the individual is currently in a neutral emotional state, and there is no fluctuation in skin resistance features.

[0030] As a preferred embodiment of the emotional state monitoring method based on skin resistance changes according to the present invention, wherein: the provision of customized emotional health advice by combining individual historical data includes combining the individual's historical emotional data, understanding the user's emotional patterns, discovering emotional health problems, and providing emotional health advice through big data analysis, trend modeling, and risk identification;

[0031] The emotional classification result, the corresponding skin resistance features, and the time label obtained from each monitoring will be recorded and stored in the database;

[0032] Individual characteristic data includes the user's gender, age, occupation, sleep pattern, and previous mental health history physical characteristic information;

[0033] Associate emotional data with individual characteristic data to construct an individual emotional data profile and store it in the form of a time series;

[0034] Use the historical emotional data of an individual for time series modeling to identify the laws of emotional changes and predict future emotional change trends;

[0035] Calculate the standard deviation and mean change rate of emotional data, analyze the frequency and amplitude of emotional fluctuations, and judge the emotional stability of an individual;

[0036] Identify abnormal emotional reactions that occur at specific times or in specific scenarios, and judge whether there are periodic changes in the emotional state.

[0037] As a preferred solution of the method for monitoring emotional state based on skin resistance change according to the present invention, wherein: providing customized emotional health advice by combining individual historical data includes setting personalized emotional fluctuation thresholds according to historical emotional data;

[0038] Risk level classification includes high risk, medium risk, and low risk;

[0039] For high-risk emotional states, it is recommended to seek psychological counseling, relaxation training, scenario avoidance, and daily emotional logs;

[0040] For medium-risk emotional states, it is recommended to have regular work and rest, relaxation training, self-regulation skills, and short-term emotional management;

[0041] For low-risk emotional states, it is recommended to maintain emotional health, have regular relaxation training, and positive psychological reinforcement.

[0042] Another object of the present invention is to provide a system for monitoring emotional state based on skin resistance change, which can dynamically correct the individual baseline based on skin resistance change, and solves the problem of insufficient correction of individual baseline differences in current methods for monitoring emotional state.

[0043] As a preferred solution of the system for monitoring emotional state based on skin resistance change according to the present invention, wherein: it includes an initialization module, a preprocessing module, an individual baseline dynamic correction module, and an emotional state dynamic tracking module; the initialization module is used to obtain the original skin resistance signal through multi-point sensor distribution; the preprocessing module is used to perform noise reduction, filtering, and feature adaptive extraction on the signal; the individual baseline dynamic correction module is used to establish a personalized skin resistance baseline model to eliminate individual differences; the emotional state dynamic tracking module is used to introduce time series modeling to capture the dynamic changes of the emotional state.

[0044] A computer device includes a memory and a processor. The memory stores a computer program, and the processor executes the computer program to implement the steps of a method for monitoring an emotional state based on changes in skin resistance.

[0045] A computer-readable storage medium stores a computer program thereon. When the computer program is executed by a processor, it implements the steps of a method for monitoring an emotional state based on changes in skin resistance.

[0046] Advantages of the present invention: The method for monitoring an emotional state based on changes in skin resistance provided by the present invention ensures the comprehensiveness and real-time nature of data by collecting multi-channel skin resistance signals; effectively removes noise and extracts high-dimensional emotional features by preprocessing the signals and extracting adaptive features; eliminates interference caused by individual differences by dynamically correcting the individual baseline based on changes in skin resistance; captures the dynamic changes in the emotional state by introducing time series modeling, realizing real-time dynamic tracking and prediction of the emotional state; finally, provides customized emotional health advice by combining individual historical data, achieving personalized and scientific emotional health management. Overall, the present invention has significant innovation and practicality in multi-dimensional signal acquisition, data processing, individual correction, time series modeling, and utilization of historical data, can effectively improve the accuracy, stability, and personalized recommendation effect of emotional state monitoring, and has broad application prospects and market value. Description of the Drawings

[0047] To more clearly illustrate the technical solutions of the embodiments of the present invention, the following will briefly introduce the drawings required for description in the embodiments. Obviously, the following described drawings are only some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0048] Figure 1 It is the overall flowchart of a method for monitoring an emotional state based on changes in skin resistance provided by the first embodiment of the present invention.

[0049] Figure 2 It is the risk level classification diagram of a method for monitoring an emotional state based on changes in skin resistance provided by the first embodiment of the present invention.

[0050] Figure 3 It is the overall flowchart of a system for monitoring an emotional state based on changes in skin resistance provided by the third embodiment of the present invention. Detailed Embodiments

[0051] To make the above objects, features, and advantages of the present invention more obvious and understandable, the following provides a detailed description of the specific embodiments of the present invention in conjunction with the accompanying drawings of the specification. Obviously, the described embodiments are part of the embodiments of the present invention, rather than all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the scope of protection of the present invention.

[0052] Example 1, referring to Figure 1 - Figure 2 , which is an embodiment of the present invention, provides a method for monitoring emotional state based on skin resistance changes, including:

[0053] S1: Collect multi-channel skin resistance signals.

[0054] Furthermore, the galvanic skin response (GSR) is a direct reflection of the activity of the human autonomic nervous system and is mainly regulated by the activity of sweat glands. When emotions fluctuate, the activation of the sympathetic nervous system causes changes in sweat gland secretion, thereby causing significant changes in skin resistance.

[0055] Traditional single-channel skin resistance signal acquisition only relies on a single sensor, and this method has obvious deficiencies in terms of signal stability, diversity, and anti-interference ability. To improve the accuracy and integrity of the acquired signals, the present invention uses a multi-channel sensor layout for skin resistance signal acquisition.

[0056] It should be noted that for the sensor layout strategy, high-sensitivity and high-precision skin resistance sensors are selected for the sensor type, which can capture weak skin conductance changes. The sensor needs to have strong anti-interference ability and low-noise characteristics.

[0057] Sensor layout position:

[0058] The tip of the finger (such as the index finger, middle finger): The sweat glands on the finger are dense and can significantly reflect emotional fluctuations.

[0059] The center of the palm: The palm is a relatively sensitive part for emotional reactions.

[0060] The inner side of the wrist: The skin on the wrist is thinner and the signal conduction is more stable.

[0061] Specific areas of the forearm: As an auxiliary acquisition area to enhance the robustness of the overall data.

[0062] Adopt a distributed multi-channel architecture, including a main sensor (core part, such as the finger) and auxiliary sensors (palm, wrist, forearm).

[0063] The sensor node uses a wireless transmission protocol (such as Bluetooth 5.0 or Wi-Fi) to transmit the collected data to the data processing center in real time.

[0064] Use clock synchronization technology to ensure the consistency of multiple sensor signals in the time dimension. Timestamp the multi-channel data to ensure the accuracy of signal processing.

[0065] Design an adaptive sampling frequency mechanism to adjust the sampling frequency according to the user's activity status (for example, reduce the frequency in the resting state and increase the frequency when the emotion changes drastically). The recommended conventional sampling frequency is set between 1Hz and 10Hz to capture the minute changes in skin resistance.

[0066] The data collected by the sensor is transmitted to the edge computing unit or the cloud server through the wireless communication module. The data is stored in the local cache in real time and encrypted to ensure data security. During the data transmission process, a data integrity verification algorithm is used to prevent data loss or transmission errors.

[0067] S2: Preprocess the signal and extract adaptive features.

[0068] Furthermore, the original skin resistance signal is often affected by various interference factors, such as:

[0069] Power frequency noise: 50Hz power frequency interference.

[0070] High-frequency noise: Device hardware, electromagnetic interference.

[0071] Baseline drift: Signal baseline shift caused by the sensor contact state or the user's minute movements.

[0072] In order to extract useful emotion feature signals, signal preprocessing is required, including steps such as denoising, filtering, and feature extraction. The present invention proposes an adaptive feature extraction algorithm to improve the accuracy and robustness of emotion feature extraction.

[0073] It should be noted that for signal preprocessing, filtering and noise reduction:

[0074] Power frequency noise filtering: Use a 50Hz notch filter to remove power interference.

[0075] High-frequency noise suppression: Use a low-pass filter (cutoff frequency 1Hz - 10Hz) to remove high-frequency interference.

[0076] Baseline drift correction: Use an adaptive baseline correction algorithm to adjust the signal baseline in real time and eliminate the drift caused by long-term monitoring.

[0077] Perform multi-scale decomposition on the original signal using wavelet transform, retain the key feature signals, and remove high-frequency noise. Use an adaptive smoothing algorithm to dynamically adjust the smoothing parameter to ensure signal fidelity.

[0078] Perform normalization to map the signal amplitudes of different sensors to the same numerical range (e.g., [0, 1]). Eliminate the amplitude differences between sensors to ensure the accuracy of feature extraction.

[0079] Time-domain feature extraction:

[0080] Instantaneous resistance value: The instantaneous change amplitude of skin resistance.

[0081] Rate of change: The rate of change of skin resistance.

[0082] Peak feature: Extract the amplitude and distribution of peak points.

[0083] Duration: The duration of skin resistance change under a specific emotional state.

[0084] Use fast Fourier transform to convert the time-domain signal into a frequency-domain signal. Extract key parameters such as the main frequency component, energy distribution, and power spectral density. Identify the spectral features under different emotional states.

[0085] Use a weighted feature fusion algorithm to perform weighted integration on time-domain, frequency-domain, and non-linear features to form a multi-dimensional feature vector. Introduce principal component analysis for feature dimensionality reduction, remove redundant features, and improve computational efficiency.

[0086] S3: Dynamically correct the individual baseline based on the change of skin resistance.

[0087] Furthermore, in the steps of multi-channel skin resistance signal acquisition, signal preprocessing, and adaptive feature extraction, we obtained a large number of original skin resistance signals and their multi-dimensional features after time-domain, frequency-domain, and non-linear processing. These feature data contain the resistance change patterns of individuals under different emotional states. However, due to the differences in physiological characteristics among different individuals, as well as the influence of external environment, sensor contact status, and other factors, relying solely on the original signal features for emotional state analysis will lead to significant deviations.

[0088] Therefore, individual baseline dynamic correction is an important basic step for emotional state classification and tracking. By establishing a dynamic correction model for each individual, eliminating the differences between individual resting states, and dynamically adjusting the baseline of real-time skin resistance data, the accuracy and robustness of subsequent emotional classification model construction and emotional state dynamic tracking can be effectively improved.

[0089] It should be noted that by establishing a dynamic correction model for each individual, the differences between individual resting states are eliminated, and dynamic baseline adjustment is performed on real-time skin resistance data. The individual baseline dynamic correction mathematical model based on skin resistance changes is expressed as:

[0090]

[0091] Among them, B(t) is the individual baseline dynamic correction value at time t, which is used to adjust the current skin resistance value to eliminate individual differences. t 0 is the initial time point of correction, x(t) is the skin resistance value collected at time t, μ(t) is the baseline mean at time t, representing the skin resistance reference value under the current individual's resting state, α is the time decay coefficient, which controls the weight of historical data in the current correction, σ 2 (t) is the baseline variance at time t, which reflects the degree of data fluctuation, β is the variance weight decay coefficient, which controls the contribution of data in different time periods to the current correction, λ is the individual adaptive adjustment coefficient, which is used to balance the weights of the global baseline and real-time data, φ(x i ) is the feature transformation function, which maps the collected skin resistance data xi to the feature space, is the feature change rate, which reflects the contribution of instantaneous feature changes to the overall correction, ω i is the feature weight factor, which represents the importance of different feature dimensions, and N is the number of features within the current time window.

[0092] Furthermore, B(t) > 0: The individual's current state is higher than the resting state and may be in an excited or anxious state.

[0093] B(t) < 0: The individual's current state is lower than the resting state and may be in a calm or relaxed state.

[0094] B(t) ≈ 0: The individual's current state is close to the resting state, and the emotional change is not significant.

[0095] The range of the dynamic baseline correction value B(t) stabilizes within [-1, 1] within a certain time window, and the actual range is affected by the adjustment of parameters α, β, and λ.

[0096] S4: Introduce time series modeling to capture the dynamic changes of emotional states.

[0097] Furthermore, emotion classification is the core step of the emotion state monitoring method based on skin resistance changes, directly affecting the accuracy and real-time performance of the subsequent emotion state dynamic tracking and feedback system. In the previous steps, multi-channel skin resistance signal acquisition provides rich original resistance data, signal preprocessing and adaptive feature extraction effectively filter, denoise and extract features from these data, and individual baseline dynamic correction eliminates the differences in individual physiological baselines, ensuring that the skin resistance characteristics of different individuals are compared and analyzed on the same standard.

[0098] On this basis, in the emotion classification model construction step, through mathematical modeling, the time domain, frequency domain, statistical features and non-linear features are organically integrated, and factors such as the time decay of historical data, feature change rate, and individual feature distribution are comprehensively considered. Finally, the accurate classification of the current emotion state is achieved, and a stable initial classification benchmark is provided for the next emotion state dynamic tracking.

[0099] It should be noted that the emotion classification model is expressed as:

[0100]

[0101] Among them, C is the emotion classification result, representing the optimal classification label of the emotion state at the current moment, y is a category in the emotion category set, Y is the set of all emotion categories, and x i (t) is the i-th skin resistance feature value collected at time t, and μ is the mean value of skin resistance features within the time window; is the total variance of all features within the current time window, reflecting the degree of dispersion of the overall data, is the partial derivative of the feature mapping function, representing the instantaneous change rate of the k-th feature, and z k is the k-th feature parameter, N is the number of features within the time window, M is the dimension number of features within the time window, and K is the number of feature mappings;

[0102] Furthermore, when C > 0: it indicates that the current individual's skin resistance signal is in a relatively high active state, which may reflect excitement or anxiety emotions.

[0103] When C < 0: it indicates that the current individual's skin resistance signal is relatively stable, which may reflect relaxation or calm emotions.

[0104] When C ≈ 0: it indicates that the current individual's skin resistance signal has no significant change, which may reflect a neutral emotion state.

[0105] The classification result C takes values in the set Y of different emotion category labels, ensuring the interpretability of the classification result in the actual emotion state.

[0106] S5: Provide customized emotion health advice by combining individual historical data.

[0107] Furthermore, in the emotional state monitoring method based on skin resistance changes, real-time monitoring can only capture an individual's emotional state at a specific moment, while the management of emotional health requires attention to long-term emotional change trends and personalized characteristics. Therefore, by combining an individual's historical emotional data, through big data analysis, trend modeling, and risk identification, it is possible to more comprehensively understand the user's emotional patterns, discover potential emotional health problems, and provide scientific and personalized emotional health advice. This can not only improve the practicality and reliability of the system but also help users establish a long-term and effective emotional management plan to prevent the occurrence of mental health problems.

[0108] The emotional classification results obtained from each monitoring (such as excitement, calmness, anxiety, anger, etc.), the corresponding skin resistance characteristics (such as mean, change rate, variance, etc.), and time tags will be recorded and stored in the database.

[0109] Individual characteristic data: including individual characteristic information such as the user's gender, age, occupation, sleep pattern, and previous mental health history.

[0110] Associate the emotional data with the individual characteristic data to construct a complete individual emotional data profile and store it in the form of a time series for subsequent analysis and modeling.

[0111] Use the individual's historical emotional data for time series modeling to identify emotional change patterns and predict future emotional change trends. Calculate the standard deviation and mean change rate of the emotional data, analyze the frequency and amplitude of emotional fluctuations, and judge the individual's emotional stability. Identify abnormal emotional reactions that occur at specific times or in specific situations (such as being in an anxious state for a long time). Judge whether there are periodic changes in the emotional state (such as being prone to anxiety at specific time periods).

[0112] Risk threshold setting: Set personalized emotional fluctuation thresholds based on historical emotional data (such as high long-term variance or high entropy representing an unstable emotional state).

[0113] Risk level classification:

[0114] High risk: Severe emotional fluctuations, being in an anxious or angry state for a long time, may have mental health problems.

[0115] Medium risk: Occasional emotional fluctuations, overall relatively stable, but there are potential risks.

[0116] Low risk: Stable emotional state, no significant risk signals are found.

[0117] Scenario risk identification: Analyze the specific scenarios where emotional abnormalities occur (such as high work pressure, lack of sleep, specific interpersonal communication events, etc.).

[0118] Suggestions for high - risk emotional states:

[0119] Psychological counseling suggestions: It is recommended that users seek the help of professional psychological counselors for in - depth mental health assessment.

[0120] Relaxation training: Provide professional relaxation techniques such as deep breathing, meditation training, and progressive muscle relaxation.

[0121] Situation avoidance: Avoid or reduce exposure to high - pressure environments or scenarios that trigger negative emotions.

[0122] Daily emotion log: It is recommended that users record their daily emotional states to assist the system in further analysis.

[0123] Suggestions for medium - risk emotional states:

[0124] Regular schedule suggestions: Ensure sufficient sleep and maintain a regular daily schedule.

[0125] Relaxation training suggestions: Conduct 10 - minute relaxation training daily, such as meditation and yoga.

[0126] Self - regulation techniques: Provide emotion regulation methods, such as simple exercises of cognitive behavioral therapy (CBT).

[0127] Short - term emotion management: Provide short - term goals and plans to gradually reduce the frequency of negative emotions.

[0128] Suggestions for low - risk emotional states:

[0129] Emotional health maintenance: It is recommended that users maintain their current good emotional states and conduct regular self - checks on emotions.

[0130] Regular relaxation training: Conduct relaxation training 1 - 2 times a week to maintain good psychological resilience.

[0131] Positive psychological reinforcement: Provide encouraging feedback to strengthen good emotional habits.

[0132] When the user's emotional state deviates from the normal range, the system immediately issues reminders and suggestions. Generate personalized emotional health reports weekly and monthly, showing the changing trend of emotional states, fluctuation frequency, and potential risks. Display emotional trends, risk levels, and health suggestions in the form of visual charts on the APP or wearable devices. The system will continuously optimize the personalized suggestion generation model according to the user's historical response situations to make it more in line with the actual needs of users.

[0133] Example 2, an embodiment of the present invention, provides a method for monitoring emotional states based on changes in skin resistance. In order to verify the beneficial effects of the present invention, scientific demonstrations are carried out through economic benefit calculations and simulation experiments.

[0134] First of all, this embodiment aims to verify the accuracy and effectiveness of the emotional state monitoring method based on skin resistance changes under different emotional states, involving multiple technical links such as signal acquisition, feature extraction, individual baseline correction, time series modeling, and historical data analysis. In this experiment, 10 subjects (5 males and 5 females) aged between 20 and 45 years old with different occupations and living habits were selected. The experiment was carried out in a controlled laboratory environment with the room temperature maintained at 24±1°C and the humidity maintained between 50% and 60% to minimize the interference of the external environment.

[0135] Each subject was equipped with a multi-channel skin resistance sensor, and the sensors were respectively arranged at the finger tips, the center of the palm, the inner side of the wrist, and the forearm to ensure comprehensive and stable skin resistance data collection. The system uses a wireless transmission protocol (Bluetooth 5.0) to transmit multi-channel signals to the data processing center. The sampling frequency is adjusted according to the activity state of the subject and is set in the range of 1Hz to 10Hz. Each subject was continuously monitored for 1 hour, during which emotional inductions in a quiet state, a mild stress state, and a high-pressure state were carried out respectively to ensure the diversity of skin resistance data. The collected original skin resistance signals were preprocessed through the following steps:

[0136] Filtering and noise reduction: Use a 50Hz notch filter to remove power interference and a **low-pass filter (cutoff frequency 10Hz)** to suppress high-frequency noise.

[0137] Baseline drift correction: Eliminate baseline drift through an adaptive baseline correction algorithm to ensure the stability of the signal.

[0138] Signal smoothing: Use wavelet transform for multi-scale decomposition, retain key feature signals, and remove redundant noise.

[0139] Signal normalization: Normalize the data of different sensors to ensure the comparability of data from different sampling channels.

[0140] Adaptive feature extraction includes time-domain features (such as mean, variance, rate of change), frequency-domain features (spectrum energy distribution), and non-linear features (approximate entropy, fractal dimension), and feature fusion is carried out to form a high-dimensional emotional feature vector.

[0141] Based on the skin resistance signals of the subjects in the resting state, a dynamic baseline model is constructed to perform personalized baseline correction for each subject, eliminating the physiological differences between individuals. The system dynamically adjusts the real-time signals of each subject to ensure the accuracy of the subsequent emotion classification process. Based on the above feature data, a time series model is introduced to dynamically model the emotional states in different time periods, capturing the emotional change trends. This model comprehensively considers the time decay of historical data, the feature change rate, and the individual feature distribution, and real-time tracks the emotional states of each subject. Combining the historical emotional state data of each subject, the system conducts emotional trend analysis, stability assessment, and risk level classification (high risk, medium risk, low risk). For different risk levels, customized health management solutions such as psychological counseling, relaxation training, and regular work and rest suggestions are provided respectively.

[0142] Table 1 Experimental data table

[0143]

[0144] As can be seen from the table, different subjects show significant skin resistance change patterns under different sampling frequencies and emotion induction states. Subjects S3 and S7 have a relatively large standard deviation of skin resistance and significantly higher baseline correction values under high-pressure emotional states. The emotion classification results are "anxiety", and they are judged as high risk and need psychological counseling. In contrast, subjects S2 and S5 have lower baseline correction values and relatively stable skin resistance under resting or relaxed states, showing a low-risk state, and are recommended to have regular work and rest and positive psychological strengthening.

[0145] After introducing dynamic baseline correction, the error caused by the physiological baseline differences between different subjects is effectively reduced, and the accuracy of emotion state classification is improved. Time series modeling enables the system to capture the real-time change patterns of emotion states, especially showing high sensitivity when emotions fluctuate suddenly. Combining historical data analysis, the long-term emotional trends of each subject are comprehensively identified, and personalized health suggestions are generated.

[0146] Compared with the traditional emotion monitoring method with single-channel and static baseline, the present invention shows significant advantages in terms of accuracy, real-time performance, and personalization through multi-channel signal acquisition, dynamic baseline correction, time series modeling, and personalized emotion health suggestions. This not only improves the robustness of the system but also provides a more scientific and accurate emotion health management solution for different users, demonstrating the innovation and practicality of the present invention in actual application scenarios.

[0147] Finally, the experimental results prove that the present invention has significant technical advantages and innovation in emotion state monitoring and health management, and has broad application prospects.

[0148] Example 3, refer toFigure 3 , which is an embodiment of the present invention, provides an emotional state monitoring system based on skin resistance change, including an initialization module, a preprocessing module, an individual baseline dynamic correction module, and an emotional state dynamic tracking module.

[0149] Among them, the initialization module is used to obtain the original skin resistance signal through multi-point sensor distribution, the preprocessing module is used to perform noise reduction, filtering and feature adaptive extraction on the signal, the individual baseline dynamic correction module is used to establish a personalized skin resistance baseline model to eliminate individual differences, and the emotional state dynamic tracking module is used to introduce time series modeling to capture the dynamic changes of the emotional state.

[0150] If the function is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on such an understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or a part of this technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods of the various embodiments of the present invention. The aforementioned storage medium includes: various media such as USB flash drives, mobile hard disks, read-only memories (ROM, Read-Only Memory), random access memories (RAM, Random Access Memory), magnetic disks or optical discs that can store program codes.

[0151] The logic and / or steps represented in the flowchart or described in other ways herein, for example, can be considered as a definite sequence list of executable instructions for implementing logical functions, and can be specifically implemented in any computer-readable medium for use by an instruction execution system, apparatus, or device (such as a computer-based system, a system including a processor, or other systems that can fetch instructions from the instruction execution system, apparatus, or device and execute the instructions), or in combination with these instruction execution systems, apparatuses, or devices. For the purposes of this specification, a "computer-readable medium" can be any device that can contain, store, communicate, propagate, or transmit a program for use by or in combination with an instruction execution system, apparatus, or device.

[0152] More specific examples (a non-exhaustive list) of computer-readable media include the following: electrical connections (electronic devices) having one or more wirings, portable computer diskettes (magnetic devices), random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber devices, and portable compact disc read-only memory (CDROM). Additionally, the computer-readable media can even be paper or other suitable media on which a program can be printed, because the program can be obtained electronically, for example, by optically scanning the paper or other media, followed by editing, interpretation, or otherwise processing it as appropriate, and then storing it in a computer memory.

[0153] It should be understood that the various parts of the present invention can be implemented by hardware, software, firmware, or a combination thereof. In the above embodiments, multiple steps or methods can be implemented by software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented by hardware, as in another embodiment, any one or a combination of the following techniques well known in the art can be used: discrete logic circuits having logic gate circuits for implementing logical functions on data signals, application specific integrated circuits having appropriate combinational logic gate circuits, programmable gate arrays (PGA), field programmable gate arrays (FPGA), etc. It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit them. Although the present invention has been described in detail with reference to the preferred embodiments, those of ordinary skill in the art should understand that the technical solutions of the present invention can be modified or equivalently replaced without departing from the spirit and scope of the technical solutions of the present invention, and they should all be covered by the scope of the claims of the present invention.

[0154] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit them. Although the present invention has been described in detail with reference to the preferred embodiments, those of ordinary skill in the art should understand that the technical solutions of the present invention can be modified or equivalently replaced without departing from the spirit and scope of the technical solutions of the present invention, and they should all be covered by the scope of the claims of the present invention.

Claims

1. A method for monitoring emotional state based on skin resistance change, characterized in that: include: Collect multi-channel skin resistance signals; Preprocess the signal and extract adaptive features; Dynamic correction of individual baseline based on skin resistance changes; Introducing time series modeling to capture the dynamic changes of emotional states; Provide customized emotional health recommendations based on individual historical data.

2. The method for monitoring emotional state based on skin resistance change according to claim 1, characterized in that: The collecting of multi-channel skin resistance signals includes collecting skin resistance signals by using a multi-channel skin sensor layout; Sensors are placed at finger tips, palm center, inner wrist, and forearm. Multi-channel configuration uses a distributed multi-channel architecture, including main and auxiliary sensors. Sensor nodes use wireless transmission protocols to transmit collected data to the data processing center in real time. Adjust the sampling frequency according to the user's activity status. It is recommended to set the frequency between 1Hz and 10Hz.

3. The method for monitoring emotional state based on skin resistance change according to claim 2, characterized in that: The preprocessing of the signal and extracting the adaptive features includes performing signal preprocessing, including filtering and noise reduction, signal smoothing and signal standardization; Filtering and noise reduction include power frequency noise filtering, high frequency noise suppression and baseline drift correction; The power frequency noise is filtered out by using a 50Hz notch filter to remove power supply interference; High-frequency noise suppression uses a filter with a cutoff frequency of 1Hz to 10Hz to remove high-frequency interference; baseline drift correction uses an adaptive baseline correction algorithm to adjust the signal baseline in real time to eliminate drift caused by monitoring; Signal smoothing is to use wavelet transform to decompose the original signal into multiple scales, retain the characteristic signal and remove high-frequency noise; Perform normalization processing to map the signal amplitudes of different sensors to the same value range; Adaptive feature extraction includes extracting time domain features, frequency domain features, nonlinear features and adaptive feature fusion.

4. The method for monitoring emotional state based on skin resistance change according to claim 3, characterized in that: The dynamic correction individual baseline based on skin resistance change includes characteristic data including resistance change patterns of individuals in different emotional states. Due to differences in physiological characteristics of different individuals, as well as the influence of external environment and sensor contact state factors, emotional state analysis of original signal characteristics will lead to deviations; By establishing a dynamic correction model for each individual, eliminating the differences between individuals in resting state, and dynamically adjusting the baseline of real-time skin resistance data, the mathematical model of individual baseline dynamic correction based on skin resistance change is expressed as: Among them, B(t) is the individual baseline dynamic correction value at time t, which is used to adjust the current skin resistance value and eliminate individual differences. t0 is the initial time point of the correction. x(t) is the skin resistance value collected at time t. μ(t) is the baseline mean at time t, which represents the skin resistance reference value of the current individual in the resting state. α is the time attenuation coefficient, which controls the weight of historical data in the current correction. σ 2 (t) is the baseline variance at time t, reflecting the degree of data fluctuation, β is the variance weight attenuation coefficient, controlling the contribution of data in different time periods to the current correction, λ is the individual adaptive adjustment coefficient, used to balance the weight of the global baseline and real-time data, φ(x i ) is the feature conversion function, which maps the collected skin resistance data xi to the feature space. is the characteristic change rate, reflecting the contribution of instantaneous characteristic changes to the overall correction, ω i is the feature weight factor, indicating the importance of different feature dimensions, and N is the number of features in the current time window; When B(t)>0, it means that the current skin resistance baseline is higher than the resting state, reflecting that the individual is in an excited or nervous state; When B(t) < 0, it means that the current skin resistance baseline is lower than the resting state, reflecting that the individual is in a relaxed or calm state; When B(t)≈0, it means that the current state is the individual's resting state.

5. The method for monitoring emotional state based on skin resistance change according to claim 4, characterized in that: The introduction of time series modeling to capture dynamic changes in emotional states includes dynamic correction of individual baselines to eliminate differences in individual physiological baselines, ensuring that skin resistance characteristics of different individuals are compared and analyzed on the same standard. The emotional classification model construction step uses mathematical modeling to organically integrate time domain, frequency domain, statistical characteristics and nonlinear characteristics, comprehensively considers the time attenuation of historical data, feature change rate, and individual feature distribution, and realizes accurate classification of the current emotional state. The sentiment classification model is expressed as: Among them, C is the emotion classification result, which represents the optimal classification label of the current emotional state, y is a category in the emotion category set, Y is the set of all emotion categories, and x i (t) is the i-th skin resistance characteristic value collected at time t, and μ is the mean value of the skin resistance characteristic in the time window; is the sum of the variances of all features in the current time window, reflecting the degree of discreteness of the overall data. is the partial derivative of the feature mapping function, indicating the instantaneous rate of change of the kth feature, z k is the kth feature parameter, N is the number of features in the time window, M is the number of dimensions of the features in the time window, and K is the number of feature maps; If C>0, the individual is currently in an excited or anxious state; If C<0, the individual is currently in a calm or relaxed state; If C≈0, the individual is currently in a neutral emotional state, and the skin resistance characteristics do not fluctuate.

6. The method for monitoring emotional state based on skin resistance change according to claim 5, characterized in that: Providing customized emotional health advice in combination with individual historical data includes combining the individual's historical emotional data, understanding the user's emotional patterns through big data analysis, trend modeling, and risk identification, discovering emotional health problems, and providing emotional health advice; Each time the emotion classification result is obtained through monitoring, the corresponding skin resistance characteristics and time tags will be recorded and stored in the database; Individual characteristic data includes the user’s gender, age, occupation, sleep pattern, and previous mental health history; Associate emotional data with individual characteristic data, build individual emotional data archives, and store them in the form of time series; Use individual historical emotion data to conduct time series modeling, identify emotion change patterns, and predict future emotion change trends; Calculate the standard deviation and mean change rate of emotional data, analyze the frequency and amplitude of emotional fluctuations, and judge the emotional stability of individuals; Identify abnormal emotional reactions that occur at specific times or situations, and determine whether there are cyclical changes in emotional states.

7. The method for monitoring emotional state based on skin resistance change according to claim 6, characterized in that: The providing customized emotional health advice in combination with individual historical data includes setting personalized emotional fluctuation thresholds based on historical emotional data; Risk level classification includes high risk, medium risk and low risk; For high-risk emotional states, psychological counseling, relaxation training, situation avoidance, and daily emotional journaling are recommended; For medium-risk emotional states, regular work and rest, relaxation training, self-regulation techniques, and short-term emotional management are recommended; For low-risk emotional states, emotional health, regular relaxation training, and positive psychological reinforcement are recommended.

8. A system using the emotional state monitoring method based on skin resistance change as claimed in any one of claims 1 to 7, characterized in that: It includes initialization module, preprocessing module, individual baseline dynamic correction module, and emotional state dynamic tracking module; The initialization module is used to obtain the original skin resistance signal through multi-point sensor distribution; The preprocessing module is used to perform noise reduction, filtering and adaptive feature extraction on the signal; The individual baseline dynamic correction module is used to establish a personalized skin resistance baseline model to eliminate individual differences; The emotional state dynamic tracking module is used to introduce time series modeling to capture the dynamic changes of emotional states.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the method for monitoring emotional state based on skin resistance change according to any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method for monitoring emotional state based on skin resistance change according to any one of claims 1 to 7 are implemented.

Images