Personalized psychological decompression intervention system and device

Abstract

The invention provides a personalized psychological decompression intervention system and device, and relates to the technical field of psychological decompression, the system mainly comprises a data acquisition module, a data processing module and a result generation module; the data acquisition module is constructed based on an Apache Storm system and comprises a Spout component and a plurality of Bolt components; the Bolt component stores a pre-trained three-classification model and is used for outputting three correction signals including enhanced intervention, maintained intervention or weakened intervention; the data processing module comprises a pressure index unit, a VR generation unit, a feedback adjustment unit and an evaluation analysis unit; and the result generation module is used for sending out the evaluation and analysis result. According to the scheme, immersive decompression can be carried out on the testee through the VR technology; through a big data technology and machine learning, physiological data of a testee are collected in real time, a psychological decompression intervention effect is fed back, VR materials are intelligently corrected, a pressure index change curve and an intervention effect evaluation report are generated, and personalized pressure evaluation and decompression suggestions are provided.

Description

Technical Field

[0001] The present invention relates to the technical field of psychological stress reduction, and particularly to a personalized psychological stress reduction intervention system and device. Background Art

[0002] With the accelerating pace of life, everyone often faces various life pressures, leading to increasingly serious psychological problems. For this reason, many psychological stress assessment and intervention methods have emerged, such as psychological scale methods and biofeedback methods.

[0003] Although these methods can, to a certain extent, assess and intervene in the psychological stress of the subjects, they often adopt static assessment and single adjustment means, lacking real-time closed-loop feedback of the intervention effect on the adjustment means; and they cannot take into account the physical and mental characteristics of different subjects; moreover, in terms of intervention methods, most are mainly two-dimensional displays, lacking immersion, resulting in unsatisfactory psychological stress reduction intervention effects. Summary of the Invention

[0004] The purpose of the present invention is to provide a personalized psychological stress reduction intervention system and device to solve at least one of the above technical problems existing in the prior art.

[0005] In a first aspect, to solve the above technical problems, the present invention provides a personalized psychological stress reduction intervention system, including a data acquisition module, a data processing module, and a result generation module; The data acquisition module is constructed based on the Apache Storm system and includes a Spout component and several Bolt components: The Spout component is used to receive the physiological data and basic information of the subject in real time; the physiological data includes electrocardiogram data, electroencephalogram data, skin conductance data, etc.; the basic information includes age, gender, social attributes, characteristics of the stress environment, and psychological scale results (such as SAM scale, psychological resilience recovery scale, positive and negative psychological scale), etc.; The Bolt component stores a pre-trained three-classification model and is used to output three correction signals: enhanced intervention, maintenance intervention, or weakened intervention; The three-classification model specifically includes a tree model, a time series classification model, and an ensemble model; The tree model outputs a basic stress classification result based on the basic information; The time series classification model, such as RNN, LSTM, and attention model, etc., outputs an electrocardiogram stress classification result, an electroencephalogram stress classification result, and a skin conductance stress classification result respectively based on the normalized values of electrocardiogram data, electroencephalogram data, and skin conductance data; each classification result is set as an integer between 0 and 10, and the larger the value, the greater the psychological stress; The integrated model is used to calculate the sum after weighting the basic pressure classification result, electrocardiogram pressure classification result, electroencephalogram pressure classification result, and skin conductance pressure classification result, compare the calculation result with a preset intervention threshold, and output three correction signals: enhanced intervention, maintained intervention, or weakened intervention, so as to dynamically correct the psychological stress reduction intervention plan; the preset intervention threshold includes a first intervention threshold and a second intervention threshold, and the first intervention threshold is less than the second intervention threshold: when the calculation result is less than the first intervention threshold, a correction signal for weakened intervention is output; when the calculation result is greater than or equal to the first intervention threshold and less than the second intervention threshold, a correction signal for maintained intervention is output; when the calculation result is greater than or equal to the second intervention threshold, a correction signal for enhanced intervention is output; The data processing module includes a stress index unit, a database, a VR (virtual reality) generation unit, a feedback adjustment unit, and an evaluation and analysis unit; The stress index unit calculates the stress index of the subject based on the physiological data and the basic information, and is used to evaluate the psychological stress level of the subject; The database stores two-dimensional picture materials; the two-dimensional picture materials can be selected from the IAPS International Affective Picture System and the CAPS Chinese Emotion Picture System, and the two-dimensional picture materials with a valence above 6.5, an arousal below 3.5, and a dominance between 5 and 7 are selected as the reference basis for constructing the VR scene; The VR generation unit selects two-dimensional picture materials from the database based on the psychological stress reduction intervention plan, and then converts the two-dimensional picture materials into three-dimensional images through 3D modeling software, such as Meshroom open-source software, and constructs a VR scene to play for the subject; the VR scene includes images and sounds; The feedback adjustment unit adjusts the psychological stress reduction intervention plan based on the correction signal to obtain a personalized psychological stress reduction intervention plan; the psychological stress reduction intervention plan includes the category type of the material, the playback order, the stress reduction index, and the playback times; the category type can include natural scenery categories, warm family scene categories, pet animal categories, etc. that can achieve stress reduction and relaxation effects; The evaluation and analysis unit generates a stress index change curve and an intervention effect evaluation report based on the stress index; the intervention effect evaluation report determines the intervention conclusion, such as "the stress reduction intervention is effective", "the stress reduction intervention is ineffective", etc. based on the stress index change curve according to the preset stress index threshold; The result generation module is used to send out the stress index change curve and the intervention effect evaluation report.

[0006] Through the above system, immersive stress reduction can be carried out on the subjects through VR technology; and through big data technology and machine learning, the physiological data of the subjects can be collected in real time, the effects of psychological stress reduction interventions can be fed back, VR materials can be intelligently corrected, a stress index change curve and an intervention effect evaluation report can be generated, and personalized stress assessment and stress reduction suggestions can be provided.

[0007] In a feasible implementation, the Spout component initializes the data stream through the initialize method, which is used to enable the Spout component to accurately receive the physiological data and basic information of the subject; and connects to the Bolt component through the nextTuple method, which is used to enable the Spout component to effectively transfer the data stream to the next Bolt component in the topology.

[0008] In a feasible implementation, in the Bolt component, through the execute method, a three-classification model is called to perform three-classification on the physiological data and basic information of the subject in real time.

[0009] In a feasible implementation, the electroencephalogram data includes alpha waves and / or beta waves, and these two bands are related to psychological stress: alpha waves can often indicate that the brain is in a relaxed state; beta waves can often indicate that the brain is in a tense state.

[0010] In a feasible implementation, the training process of the three-classification model includes: Step a1: Collect the basic information initially input by the subject, calculate the initial stress index, and construct an initial psychological stress reduction intervention plan. Step a2: Based on the psychological stress reduction intervention plan, search for several two-dimensional picture materials from the database, construct a VR scene through 3D modeling, such as the Meshroom open-source software, and play it for the subject to start a group of psychological stress reduction interventions; collect the physiological data of the subject in real time; the materials are marked according to the stress reduction index; the stress reduction index can be the opposite of the stress index. Step a3: After playing each material, collect the SAM psychological scale results currently input by the subject, update the basic information, calculate the current stress index until all the materials in this group of psychological stress reduction intervention plans are played. Step a4: Based on the basic information and the current physiological data, input into the three-classification model and output a correction signal. Step a5: Adjust the psychological stress reduction intervention plan based on the correction signal. Step a6: Iteratively execute step a2 until the stress index is less than the preset stress index threshold.

[0011] Through the above method, a reliable three-classification model can be trained.

[0012] In a feasible implementation, the calculation formula of the pressure index can specifically be: SI = σ1 * BD + σ2 * PD; Wherein, SI represents the pressure index; BD represents the basic information index; σ1 represents the weight of the basic information index; PD represents the physiological data index; σ2 represents the weight of the physiological data index.

[0013] In a feasible implementation, the basic information index includes the SAM scale index, and the specific calculation method is: After converting the relaxed PAD value in the PAD emotion scale to the positive value range, the relaxed standard value is obtained; the SAM scale data of the subject for different materials is collected; the Euclidean distance d between the SAM scale data and the relaxed standard value is calculated as the SAM index, and the specific formula is: ; Wherein, , and respectively represent the scale data values of the subject for different materials; , and respectively represent the numerical values of the relaxed standard value.

[0014] In a feasible implementation, the basic information index further includes an age coefficient, which is used to correct the basic information index through product operation, and the specific value-taking method is: The age range of 0 - 18 years old is divided into the low-pressure area, and the value is 1; The age range of 18 - 60 years old is divided into the high-pressure area, and the value is 1.6; The age range of over 60 years old is divided into the medium-pressure area, and the value is 1.2; In this way, the characteristic that the psychological pressure of middle-aged people is significantly greater than that of the elderly and teenagers can be reflected, so as to enhance the intensity of psychological decompression intervention for middle-aged people in the subsequent process.

[0015] In a feasible implementation, the physiological data index includes an electroencephalogram index, and the specific calculation method is: Preprocess the electroencephalogram data; divide and quantify the electroencephalogram data according to the emotion type into emotion values; according to the measurement time, evenly divide the electroencephalogram data into several time periods, and sort them according to the magnitude of the emotion values in each time period, then calculate the th subject's emotion average value for the th material The specific formula is: ; Wherein, represents the maximum emotion value; Represents the median emotional value; Represents the minimum emotional value; For each piece of material, calculate the average emotional value of the subject, and map it to the interval [0, 1] through the min-max normalization method. The specific formula is: ; Wherein, Represents the original value of the current emotional value; Represents the minimum value among the average emotional values corresponding to each piece of material; Represents the maximum value among the average emotional values corresponding to each piece of material; Represents the conversion result of the current emotional value.

[0016] In a feasible implementation manner, the physiological data index further includes the coefficient of variation of heart rate, which is used to correct the physiological data index through product operation to reflect the inner stability degree of the subject. The specific calculation formula is: ; Wherein, Represents the coefficient of variation of heart rate of the th subject regarding the th piece of material; Represents the standard deviation of heart rate of the th subject regarding the th piece of material; Represents the average heart rate of the th subject regarding the th piece of material; The specific calculation formula of the is: ; Wherein, Represents the statistical quantity of the subjects; Represents the heart rate of the th subject regarding the th piece of material; Represents the preset average heart rate of the th piece of material; The specific calculation formula of the is: ; For the average coefficient of variation of heart rate of the th piece of material, the specific formula is: .

[0017] In a feasible implementation manner, the physiological data index further includes an electrocardiogram index, which is also used to perform product operation correction on the physiological data index to further reflect the inner stability degree of the subject. The specific calculation method is as follows: Preprocess the electrocardiogram data; perform HRV data analysis on the electrocardiogram data, and calculate HRV characteristic values, including SDNN value (standard deviation of all normal sinus beat intervals), RMSSD value (root mean square value of the differences between adjacent normal sinus beat intervals throughout the process), and LF / HF ratio (ratio between low-frequency power and high-frequency power); since the SDNN value, RMSSD value, and LF / HF ratio are related to the stress degree of emotions, a high value can reflect the enhancement of sympathetic nerve activity, corresponding to emotional classification characteristics such as tension and anxiety, and a decreased value can correspond to emotional classification characteristics such as calmness and tranquility, with the anxiety feeling significantly reduced. Therefore, based on the corresponding relationship of the above emotional classification characteristics, construct an emotion model: quantify the HRV characteristic values into emotion values; use the min-max normalization method to map the emotion values into the interval [0, 1] to quantify the emotional state and provide reliable physiological data support for emotion analysis. The specific formula is: ; where, represents the current HRV characteristic value; represents the minimum value among all HRV characteristic values; represents the maximum value among all HRV characteristic values; represents the normalized result.

[0018] In a feasible implementation manner, the Spout component is also used to receive the eye movement data of the subject, and the eye movement data includes binocular coordinates and eye movement speed; in the Bolt component, the eye movement data of the subject is also transmitted backward in real time through the execute method; the feedback adjustment unit also collects the image playback coordinate range in the VR generation unit, converts the eye movement data into the human eye fixation coordinate range in the display area, then compares the image playback coordinate range with the human eye fixation coordinate range, calculates the concentration of the subject, and adjusts the psychological decompression intervention plan based on the concentration; in this way, the eye-tracing technology can be used to perform correlation analysis on the attention position of the human eye and the position of the image material in the VR scene, obtain the attraction of the image material to the subject and the current concentration degree of the subject, and synchronously adjust the psychological decompression intervention plan, which is beneficial to obtaining a more personalized psychological decompression intervention plan suitable for the subject.

[0019] In a feasible implementation manner, the specific calculation method of the concentration is as follows: Step b1: Collect the image playback coordinate range D1 of the current frame; based on the standard time sequence, collect the current binocular coordinates, and calculate the current human eye fixation coordinate range D2 in the display area through the principle of solid geometry; Step b2: Calculate the number MT of the same coordinates in D1 and D2; calculate the number MX of the coordinates in the smaller one of D1 and D2; take the ratio of MT to MX as the coincidence degree CH, and make a judgment: if the coincidence degree CH is greater than the preset coincidence degree threshold, it is determined that the human eye is fixated on the image, indicating that the image attracts the attention of the subject and the psychological decompression intervention effect is effective; if the coincidence degree CH is less than or equal to the preset coincidence degree threshold, it is determined that the human eye is not fixated on the image, indicating that the subject may have been distracted at this time and has not entered the psychological decompression intervention state, the image does not attract the attention of the subject, and the psychological decompression intervention effect is weakened or invalid; Step b3: Collect the image playback coordinate range D1' of the image after time t, and calculate the image displacement speed VS between D1 and D1'. The specific calculation formula can be: VS = L / t; where L represents the relative displacement between D1 and D1'; Step b4: Based on the standard time sequence, collect the eye movement speed VY within time t; Step b5: Take the ratio of VS to VY as the follow-up degree SD, and make a judgment: if the follow-up degree SD is greater than the preset follow-up degree threshold, it is determined that the human eye always follows the image movement, indicating that the image is attractive to the subject and the psychological decompression intervention effect is effective; if the follow-up degree SD is less than or equal to the preset follow-up degree threshold, it is determined that the human eye does not always follow the image movement, indicating that the image is not attractive to the subject and the psychological decompression intervention effect fails; Step b6: Based on the coincidence degree CH and the follow-up degree SD, calculate the concentration degree ZZ. The specific formula can be: ZZ = q1*CH + q2*SD; where q1 represents the preset weight of the coincidence degree; q2 represents the preset weight of the follow-up degree.

[0020] In a feasible implementation manner, the specific method for adjusting the psychological decompression intervention plan based on the concentration degree is: When the concentration degree ZZ is greater than the preset concentration degree threshold, it indicates that the current image material can attract the attention of the subject, and the playback times of the current image material and the image materials with the same decompression index are increased according to a preset ratio; when the concentration degree ZZ is less than or equal to the preset concentration degree threshold, it indicates that the current image material is not very attractive to the subject, and the playback times of the current image material and the image materials with the same decompression index are reduced according to a preset ratio.

[0021] In a feasible implementation, the Spout component is further configured to receive the subject's facial expression data, where the facial expression data includes infrared images of glabellar lines and crow's feet lines; the three-classification model further includes a convolutional neural network model (CNN), which inputs the infrared images of glabellar lines and crow's feet lines and outputs the classification result of facial expression stress; the integrated model is further configured to assign weights to the basic stress classification result, electrocardiogram stress classification result, electroencephalogram stress classification result, skin conductance stress classification result, and facial expression stress classification result, and then output three correction signals, namely enhanced intervention, maintenance intervention, or weakened intervention, so as to dynamically correct the psychological stress reduction intervention plan; The glabellar line, also known as the inter-brow line, is a wrinkle formed between the two eyebrows due to the active corrugator supercilii muscle and depressor supercilii muscle, and this wrinkle is commonly seen in cases of high psychological stress; The crow's feet line refers to the radial wrinkles at the outer corners of the eyes, and this wrinkle is commonly seen in cases of a happy mood; In this way, the characteristics that facial expression features such as glabellar lines and crow's feet lines are easy to reflect the subject's psychological state can be utilized. Compared with physiological data such as electrocardiogram, electroencephalogram, and skin conductance data, facial expression features are basically not interfered by other factors such as the body state and self-diseases, and can effectively and directly feedback the effect of the psychological stress reduction intervention plan.

[0022] In a feasible implementation, the training process of the convolutional neural network model includes: collecting infrared images of glabellar lines and crow's feet lines of several subjects in different psychological states and marking them according to the classification result of facial expression stress as a data set; the psychological states include joy, sadness, relaxation, tension, fatigue, and attention, etc.; randomly dividing the data set into a training set and a test set; inputting the training set and the test set into the convolutional neural network model respectively, outputting the classification result of facial expression stress, and performing iterative training and testing until the requirements of the evaluation index are met; the evaluation index includes mean average precision (mAP), etc.

[0023] In a second aspect, based on the same inventive concept, the present application also provides a personalized psychological stress reduction intervention device for implementing the above personalized psychological stress reduction intervention system, including a VR headset, an ECG device, an EEG device, and a computing platform that are electrically connected to each other; The VR headset is used to display a VR scene to enhance the subject's immersive experience; The ECG device refers to a wearable electrocardiogram monitoring device, which is used to real-time monitor heart rate variability (HRV) to evaluate the cardiac stress response; The EEG device refers to a wearable electroencephalogram monitoring device, which is used to real-time record brain waves; The computing platform includes a processor, a memory, and a bus. The memory stores instructions and data read by the processor. The processor is used to call the instructions and data in the memory. The bus is connected between the functional components to transmit information.

[0024] In a feasible implementation, the VR headset includes an eye tracking sensor for collecting the eye movement data of the subject.

[0025] In a feasible implementation, the VR headset further includes an infrared camera. The infrared camera is disposed inside the VR headset and is used to capture infrared images of the glabellar lines and crow's feet regions of the subject. In this way, real-time images of the glabellar lines and crow's feet regions can be collected without affecting the VR display effect, so as to feedback the psychological stress reduction effect and correct the psychological stress reduction intervention plan.

[0026] Adopting the above technical solution, the present invention has the following beneficial effects: A personalized psychological stress reduction intervention system and device provided by the present invention can perform immersive stress reduction on the subject through VR technology; and through big data technology and machine learning, collect the physiological data of the subject in real time, feedback the psychological stress reduction intervention effect, intelligently correct VR materials, generate a stress index change curve and an intervention effect evaluation report, and provide personalized stress assessment and stress reduction suggestions; it can also perform correlation analysis on the attention position of the human eye and the position of the image materials in the VR scene through eye tracking to obtain the attractiveness of the image materials to the subject and the current concentration degree of the subject; and it can further correct the psychological stress reduction intervention plan by adding facial expression recognition. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] In order to more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the following will briefly introduce the drawings required for the description of the specific embodiments or the prior art. Obviously, the drawings in the following description are 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.

[0028] Figure 1 It is a diagram of a personalized psychological stress reduction intervention system provided by an embodiment of the present invention; Figure 2 It is a flowchart of the training process of a three-classification model provided by an embodiment of the present invention; Figure 3 It is a flowchart of the specific calculation method of the concentration degree provided by an embodiment of the present invention; Figure 4Example diagram of expression data provided by the embodiments of the present invention: Among them, Figure a is the image of crow's feet in a fatigued state; Figure b is the image of crow's feet in a relaxed state; Figure c is the image of crow's feet in a joyful state; Figure d is the image of frown lines in a sad state; Figure e is the image of frown lines in a state of attention; Figure f is the image of frown lines in a tense state; Figure 5 Schematic diagram of the main structure of the VR headset provided by the embodiments of the present invention; Figure 6 Schematic diagram of the structure of the ECG device provided by the embodiments of the present invention; Reference numerals: 1 - VR headset; 11 - First acquisition unit; 12 - Second acquisition unit; 2 - ECG device; 21 - Reference electrode; 22 - Active electrode. Detailed implementation manners

[0029] Next, the technical solutions of the present invention will be clearly and completely described in conjunction with the accompanying drawings. Obviously, the described embodiments are part of the embodiments of the present invention, rather than all of them. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts belong to the scope of protection of the present invention.

[0030] In the description of the present invention, it should be noted that the orientation or positional relationship indicated by the terms "center", "upper", "lower", "left", "right", "vertical", "horizontal", "inner", "outer", etc. is based on the orientation or positional relationship shown in the accompanying drawings. It is only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and thus cannot be understood as a limitation to the present invention. In addition, the terms "first", "second", "third" are only used for descriptive purposes and cannot be understood as indicating or implying relative importance.

[0031] In the description of the present invention, it should be noted that unless otherwise clearly defined and limited, the terms "installed", "connected", "connected" should be understood in a broad sense. For example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be directly connected or indirectly connected through an intermediate medium, and it can be the communication inside two elements. For those of ordinary skill in the art, the specific meanings of the above terms in the present invention can be understood according to specific situations.

[0032] It should also be noted that the following specific embodiments or specific implementation manners are a series of optimized setting methods listed by the present invention to further explain the specific invention content, and these setting methods can be combined with each other or used in association with each other.

[0033] The present invention will be further explained and illustrated below in conjunction with specific implementation manners.

[0034] Example 1: As Figure 1 shown, a personalized psychological stress reduction intervention system provided in this embodiment includes a data acquisition module, a data processing module, and a result generation module; The data acquisition module is constructed based on the Apache Storm system and includes a Spout component and several Bolt components: The Spout component is used to receive the physiological data and basic information of the subject in real time; the physiological data includes electrocardiogram data, electroencephalogram data, skin conductance data, etc.; the basic information includes age, gender, social attributes, characteristics of the stress environment, and psychological scale results (such as SAM scale, psychological resilience recovery scale, positive and negative psychology scale), etc.; The Bolt component stores a pre-trained three-classification model and is used to output three correction signals: enhanced intervention, maintenance intervention, or weakened intervention; The three-classification model specifically includes a tree model, a time series classification model, and an ensemble model; The tree model outputs a basic stress classification result based on the basic information; The time series classification model, such as RNN, LSTM, and attention model, etc., outputs an electrocardiogram stress classification result, an electroencephalogram stress classification result, and a skin conductance stress classification result respectively based on the normalized values of electrocardiogram data, electroencephalogram data, and skin conductance data; each classification result is set as an integer between 0 and 10, and the larger the value, the greater the psychological stress; The ensemble model is used to sum the basic stress classification result, the electrocardiogram stress classification result, the electroencephalogram stress classification result, and the skin conductance stress classification result after weighting, compare the calculation result with a preset intervention threshold, and output three correction signals: enhanced intervention, maintenance intervention, or weakened intervention, so as to dynamically correct the psychological stress reduction intervention plan; the preset intervention threshold includes a first intervention threshold and a second intervention threshold, and the first intervention threshold is less than the second intervention threshold: when the calculation result is less than the first intervention threshold, a correction signal for weakened intervention is output; when the calculation result is greater than or equal to the first intervention threshold and less than the second intervention threshold, a correction signal for maintenance intervention is output; when the calculation result is greater than or equal to the second intervention threshold, a correction signal for enhanced intervention is output; The data processing module includes a stress index unit, a database, a VR (virtual reality) generation unit, a feedback adjustment unit, and an evaluation and analysis unit; The stress index unit calculates the stress index of the subject based on the physiological data and the basic information, and is used to evaluate the psychological stress level of the subject; The database stores two-dimensional picture materials. The two-dimensional picture materials are selected from the IAPS International Affective Picture System and the CAPS Chinese Emotion Picture System, and are two-dimensional picture materials with a valence above 6.5, an arousal below 3.5, and a dominance between 5 and 7, so as to be used as a reference for constructing VR scenes. The VR generation unit selects two-dimensional picture materials from the database based on the psychological stress reduction intervention plan, and then uses Meshroom open-source software to convert the two-dimensional picture materials into three-dimensional images and construct a VR scene for playback to the subjects. The VR scene includes images and sounds. The feedback adjustment unit adjusts the psychological stress reduction intervention plan based on the correction signal, so as to obtain a personalized psychological stress reduction intervention plan. The psychological stress reduction intervention plan includes the category type of materials, the playback order, the stress reduction index, and the number of playbacks. The evaluation and analysis unit generates a stress index change curve and an intervention effect evaluation report based on the stress index. The intervention effect evaluation report determines the intervention conclusion, such as "the stress reduction intervention is effective", "the stress reduction intervention is ineffective", etc., based on the stress index change curve and the preset stress index threshold. The result generation module is used to send out the stress index change curve and the intervention effect evaluation report.

[0035] Through the above system, immersive stress reduction can be carried out on the subjects through VR technology; and through big data technology and machine learning, the physiological data of the subjects can be collected in real time, the psychological stress reduction intervention effect can be feedback, the VR materials can be intelligently corrected, a stress index change curve and an intervention effect evaluation report can be generated, and personalized stress assessment and stress reduction suggestions can be provided.

[0036] Further, the Spout component initializes the data stream through the initialize method, which is used to enable the Spout component to accurately receive the physiological data and basic information of the subjects; and connects to the Bolt component through the nextTuple method, which is used to enable the Spout component to effectively transfer the data stream to the next Bolt component in the topology.

[0037] Further, in the Bolt component, the three-classification model is called through the execute method to perform three-classification on the physiological data and basic information of the subjects in real time.

[0038] Further, the electroencephalogram data includes α waves and / or β waves, and these two bands are related to psychological stress: α waves can often indicate that the brain is in a relaxed state; β waves can often indicate that the brain is in a tense state.

[0039] Further, as Figure 2As shown, the training process of the three-classification model includes: Step a1: Collect the basic information initially input by the subject, calculate the initial stress index, and construct an initial psychological stress reduction intervention plan. Step a2: Based on the psychological stress reduction intervention plan, search for several two-dimensional picture materials from the database, construct a VR scene through the Meshroom open-source software, play it for the subject, and start a group of psychological stress reduction interventions; collect the physiological data of the subject in real time; the materials are marked according to the stress reduction index; the stress reduction index can be the opposite of the stress index. Step a3: After playing each material, collect the results of the SAM psychological scale currently input by the subject, update the basic information, calculate the current stress index until all the materials in this group of psychological stress reduction intervention plans are played. Step a4: Based on the basic information and the current physiological data, input them into the three-classification model and output a correction signal. Step a5: Adjust the psychological stress reduction intervention plan based on the correction signal. Step a6: Iteratively execute Step a2 until the stress index is less than the preset stress index threshold.

[0040] Through the above method, a reliable three-classification model can be trained.

[0041] Furthermore, the specific calculation formula of the stress index can be: SI = σ1 * BD + σ2 * PD; where SI represents the stress index; BD represents the basic information index; σ1 represents the weight of the basic information index; PD represents the physiological data index; σ2 represents the weight of the physiological data index.

[0042] Furthermore, the basic information index includes the SAM scale index, and the specific calculation method is: After converting the relaxation PAD value in the PAD emotion scale to the positive value range, the relaxation standard value is obtained; collect the SAM scale data of the subject for different materials; calculate the Euclidean distance d between the SAM scale data and the relaxation standard value as the SAM index, and the specific formula is: ; where , and respectively represent the scale data values of the subject for different materials; , and respectively represent the relaxation standard value data.

[0043] Further, the basic information index also includes an age coefficient, which is used to correct the basic information index through multiplication. The specific value-taking method is as follows: The age range from 0 to 18 years old is divided into a low-stress area, with a value of 1; The age range from 18 to 60 years old is divided into a high-stress area, with a value of 1.6; The age range above 60 years old is divided into a medium-stress area, with a value of 1.2; In this way, the characteristic that the psychological pressure of middle-aged people is significantly greater than that of the elderly and teenagers can be reflected, so as to enhance the intensity of psychological stress reduction intervention for middle-aged people in the subsequent process.

[0044] Further, the physiological data index includes an electroencephalogram index. The specific calculation method is as follows: Preprocess the electroencephalogram data; divide the electroencephalogram data according to the emotion type and quantitatively process it into emotion values; according to the measurement time, evenly divide the electroencephalogram data into several time periods, and sort them according to the size of the emotion values in each time period. Then calculate the emotion average value of the th subject regarding the th material. The specific formula is: ; Among them, represents the maximum emotion value; represents the median emotion value; represents the minimum emotion value; For each material, calculate the average emotion value of the subject, and through the min-max normalization method, map it to the interval [0,1]. The specific formula is: ; Among them, represents the original value of the current emotion value; represents the minimum value among the average emotion values corresponding to each material; represents the maximum value among the average emotion values corresponding to each material; represents the conversion result of the current emotion value.

[0045] Further, the physiological data index also includes a heart rate variability coefficient, which is used to correct the physiological data index through multiplication to reflect the inner stability degree of the subject. The specific calculation formula is: ; Among them, represents the heart rate variability coefficient of the th subject regarding the th material; represents the th subject regarding the The standard deviation of heart rate of a material; represents the th subject's average heart rate for the th material; The specific calculation formula of the above is: ; wherein, represents the statistical quantity of the subjects; represents the th subject's heart rate for the th material; represents the th preset average heart rate of the material; The specific calculation formula of the above is: ; For the average heart rate variability coefficient of the th material, the specific formula is: .

[0046] Furthermore, the physiological data index further includes an electrocardiogram index, which is also used to perform product operation correction on the physiological data index to further reflect the inner stability of the subject. The specific calculation method is as follows: Preprocess the electrocardiogram data; perform HRV data analysis on the electrocardiogram data, and calculate HRV characteristic values, including SDNN value (standard deviation of all normal sinus beat intervals), RMSSD value (root mean square value of the differences between adjacent normal sinus beat intervals throughout the process), and LF / HF ratio (ratio of low-frequency power to high-frequency power); since the SDNN value, RMSSD value, and LF / HF ratio are related to the stress level of emotions, a high value can reflect an increase in sympathetic nerve activity, corresponding to emotional classification characteristics such as tension and anxiety, and a decrease in value can correspond to emotional classification characteristics such as calmness and tranquility, with a significant reduction in anxiety. Therefore, based on the corresponding relationship of the above emotional classification characteristics, construct an emotional model: quantify the HRV characteristic values into emotional numerical values; use the min-max normalization method to map the emotional numerical values to the interval [0,1] to quantify the emotional state and provide reliable physiological data support for emotional analysis. The specific formula is: ; wherein, represents the current HRV characteristic value; represents the minimum value among all HRV characteristic values; represents the maximum value among all HRV characteristic values; represents the normalized result.

[0047] Further, the Spout component is also used to receive the eye movement data of the subject, and the eye movement data includes binocular coordinates and eye movement speed; in the Bolt component, the eye movement data of the subject is also transmitted backward in real time through the execute method; the feedback adjustment unit also collects the image playback coordinate range in the VR generation unit, converts the eye movement data into the human eye fixation coordinate range in the display area, then compares the image playback coordinate range with the human eye fixation coordinate range, calculates the concentration of the subject, and adjusts the psychological decompression intervention plan based on the concentration; in this way, the eye-tracing technology can be used to perform correlation analysis on the attention position of the human eye and the position of the image material in the VR scene, obtain the attraction of the image material to the subject and the current concentration of the subject, and synchronously adjust the psychological decompression intervention plan, which is beneficial to obtaining a more personalized psychological decompression intervention plan suitable for the subject.

[0048] Further, as Figure 3 shown, the specific calculation method of concentration is as follows: Step b1: Collect the image playback coordinate range D1 of the current frame; based on the standard time sequence, collect the current binocular coordinates, and calculate the current human eye fixation coordinate range D2 in the display area through the principle of solid geometry; Step b2: Calculate the number MT of the same coordinates in D1 and D2; calculate the number MX of the coordinates of the smaller one of D1 and D2; take the ratio of MT to MX as the coincidence degree CH, and make a judgment: if the coincidence degree CH is greater than the preset coincidence degree threshold, it is determined that the human eye is fixated on the image, indicating that the image attracts the attention of the subject and the psychological decompression intervention effect is effective; if the coincidence degree CH is less than or equal to the preset coincidence degree threshold, it is determined that the human eye is not fixated on the image, indicating that the subject may have been distracted at this time and has not entered the psychological decompression intervention state, the image does not attract the attention of the subject, and the psychological decompression intervention effect is weakened or invalid; Step b3: Collect the image playback coordinate range D1' of the image after time t, and calculate the image displacement speed VS between D1 and D1', and the specific calculation formula can be: VS = L / t; wherein, L represents the relative displacement between D1 and D1'; Step b4: Based on the standard time sequence, collect the eye movement speed VY within time t; Step b5: Take the ratio of VS to VY as the follow-up degree SD, and make a judgment: if the follow-up degree SD is greater than the preset follow-up degree threshold, it is determined that the human eye always follows the image movement, indicating that the image is attractive to the subject and the psychological decompression intervention effect is effective; if the follow-up degree SD is less than or equal to the preset follow-up degree threshold, it is determined that the human eye does not always follow the image movement, indicating that the image is not attractive to the subject and the psychological decompression intervention effect is invalid; Step b6: Calculate the concentration degree ZZ based on the coincidence degree CH and the follow-up degree SD. The specific formula can be: ZZ = q1 * CH + q2 * SD; where q1 represents the preset weight of the coincidence degree; q2 represents the preset weight of the follow-up degree.

[0049] Furthermore, the specific method for adjusting the psychological decompression intervention plan based on the concentration degree is as follows: When the concentration degree ZZ is greater than the preset concentration degree threshold, it indicates that the current image material can attract the attention of the subject. Then, increase the playback times of the current image material and the image materials with the same decompression index according to a preset ratio; when the concentration degree ZZ is less than or equal to the preset concentration degree threshold, it indicates that the current image material is less likely to attract the attention of the subject. Then, reduce the playback times of the current image material and the image materials with the same decompression index according to a preset ratio.

[0050] Furthermore, the Spout component is also used to receive the facial expression data of the subject. The facial expression data includes the infrared image of the glabellar lines and the infrared image of the crow's feet lines; the three-classification model also includes a convolutional neural network model (CNN). Input the infrared image of the glabellar lines and the infrared image of the crow's feet lines, and output the facial expression stress classification result; the integrated model is also used to assign weights to the basic stress classification result, the electrocardiogram stress classification result, the electroencephalogram stress classification result, the skin conductance stress classification result, and the facial expression stress classification result, and then output three correction signals: enhanced intervention, maintenance intervention, or weakened intervention, so as to dynamically correct the psychological decompression intervention plan; The glabellar lines, also known as the inter-brow lines, are the wrinkles formed between the two eyebrows due to the activation of the corrugator supercilii muscle and the depressor supercilii muscle. These wrinkles are common in cases of high psychological stress; The crow's feet lines refer to the radial wrinkles at the outer corners of the eyes. These wrinkles are common in cases of a happy mood; In this way, the characteristics that the facial expression features such as the glabellar lines and the crow's feet lines are easy to reflect the psychological state of the subject can be utilized. Compared with physiological data such as electrocardiogram, electroencephalogram, and skin conductance data, the facial expression features are basically not interfered by other factors such as the body state and self-diseases, and can effectively and directly feedback the effect of the psychological decompression intervention plan.

[0051] Furthermore, the training process of the convolutional neural network model includes: Collect the infrared images of the glabellar lines and the infrared images of the crow's feet lines of several subjects in different psychological states and mark them according to the facial expression stress classification results as the data set; The psychological states include joy, sadness, relaxation, tension, fatigue, and attention, etc.; Exemplarily, such as Figure 4As shown in the figure, where Figure a is the image of crow's feet under fatigue; Figure b is the image of crow's feet under a relaxed state; Figure c is the image of crow's feet under a happy state; Figure d is the image of the vertical frown lines under a sad state; Figure e is the image of the vertical frown lines under an attentive state; Figure f is the image of the vertical frown lines under a tense state; Randomly divide the data set into a training set and a test set; Input the training set and the test set into the convolutional neural network model respectively, output the classification results of expression stress, and perform iterative training and testing until the requirements of the evaluation index are met; The evaluation index includes mean average precision (mAP), etc.

[0052] Embodiment 2: This embodiment provides a personalized psychological stress reduction intervention device for implementing the above-mentioned personalized psychological stress reduction intervention system, including a VR headset, an ECG device, an EEG device, and a computing platform; The VR headset is used to display a VR scene to enhance the immersive experience of the subject; The ECG device refers to a wearable electrocardiogram monitoring device. Electrodes are attached to the chest and other parts of the subject to continuously monitor heart rate variability (HRV) to evaluate the cardiac stress response; The EEG device refers to a wearable electroencephalogram monitoring device, which is worn on the head of the subject to continuously record brain waves and can use conventional devices; The computing platform includes a processor, a memory, and a bus. The memory stores instructions and data read by the processor. The processor is used to call the instructions and data in the memory, and the bus is connected between each functional component to transmit information.

[0053] Further, as Figure 5 shown, a first acquisition part 11 is provided at the center of the front end of the VR headset 1; second acquisition parts 12 are provided on both sides of the VR headset 1; An eye movement sensor and a first infrared camera are provided in the first acquisition part 11: the eye movement sensor is used to collect the eye movement data of the subject; the first infrared camera is used to capture the infrared image of the area of the vertical frown lines of the subject; A second infrared camera is provided in the second acquisition part 12, which is used to capture the infrared image of the area of the crow's feet of the subject; In this way, the eye movement data, the infrared image of the area of the vertical frown lines, and the infrared image of the area of the crow's feet of the subject can be collected in real time without affecting the VR display effect, which is beneficial to feedback the psychological stress reduction effect and correct the psychological stress reduction intervention plan.

[0054] Further, as Figure 6As shown, a reference electrode 21 and an active electrode 22 are arranged on the surface of the ECG device 2 for real-time acquisition of the heart rate electrical signals of the subject; the housing material of the ECG device 2 is medical-grade silicone to provide a comfortable touch.

[0055] Finally, it should be noted that: the above embodiments are only used to illustrate the technical solutions of the present invention, rather than limiting it; although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that: they can still modify the technical solutions recorded in the foregoing embodiments, or perform equivalent replacements on some or all of the technical features; and these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. A personalized psychological stress reduction intervention system, characterized in that, It includes a data acquisition module, a data processing module, and a result generation module; The data acquisition module is built based on the Apache Storm system and includes a Spout component and several Bolt components: The Spout component is used to receive the physiological data and basic information of the subject in real time; the physiological data includes electrocardiogram data, electroencephalogram data, and skin conductance data; the basic information includes age, gender, social attributes, characteristics of the stress environment, and psychological scale results; The Bolt component stores a pre-trained three-classification model and is used to output three correction signals: enhanced intervention, maintenance intervention, or weakened intervention; The three-classification model specifically includes a tree model, a time series classification model, and an ensemble model; The tree model outputs a basic stress classification result based on the basic information; The time series classification model outputs an electrocardiogram stress classification result, an electroencephalogram stress classification result, and a skin conductance stress classification result respectively based on the normalized values of electrocardiogram data, electroencephalogram data, and skin conductance data; Each classification result is set as an integer between 0 and 10, and the larger the value, the greater the psychological stress; The ensemble model is used to sum the basic stress classification result, the electrocardiogram stress classification result, the electroencephalogram stress classification result, and the skin conductance stress classification result after weighting, compare the calculation result with a preset intervention threshold, and output three correction signals: enhanced intervention, maintenance intervention, or weakened intervention, so as to dynamically correct the psychological stress reduction intervention plan; the preset intervention threshold includes a first intervention threshold and a second intervention threshold, and the first intervention threshold is less than the second intervention threshold: when the calculation result is less than the first intervention threshold, a correction signal for weakened intervention is output; when the calculation result is greater than or equal to the first intervention threshold and less than the second intervention threshold, a correction signal for maintenance intervention is output; when the calculation result is greater than or equal to the second intervention threshold, a correction signal for enhanced intervention is output; The data processing module includes a stress index unit, a database, a VR generation unit, a feedback adjustment unit, and an evaluation and analysis unit; The stress index unit calculates the stress index of the subject based on the physiological data and the basic information, and is used to evaluate the psychological stress level of the subject; The database stores two-dimensional picture materials; The VR generation unit selects two-dimensional picture materials from the database based on the psychological stress reduction intervention plan, then converts the two-dimensional picture materials into three-dimensional images through 3D modeling, and constructs a VR scene to play for the subject; the VR scene includes images and sounds; The feedback adjustment unit adjusts the psychological stress reduction intervention plan based on the correction signal to obtain a personalized psychological stress reduction intervention plan; the psychological stress reduction intervention plan includes the category type of the materials, the playing order, the stress reduction index, and the playing times; The evaluation and analysis unit generates a stress index change curve and an intervention effect evaluation report based on the stress index; The result generation module is used to send out the stress index change curve and the intervention effect evaluation report.

2. The system according to claim 1, wherein The Spout component initializes the data stream through the initialize method, which is used to enable the Spout component to accurately receive the physiological data and basic information of the subject; it connects to the Bolt component through the nextTuple method, which is used to enable the Spout component to effectively transfer the data stream to the next Bolt component in the topology structure.

3. The system according to claim 1, wherein The training process of the three-classification model includes: Step a1: Collect the basic information initially input by the subject, calculate the initial stress index, and construct an initial psychological stress reduction intervention plan. Step a2: Based on the psychological stress reduction intervention plan, search for several two-dimensional picture materials from the database, construct a VR scene through 3D modeling, play it to the subject, and start a group of psychological stress reduction interventions; collect the physiological data of the subject in real time; the materials are marked according to the stress reduction index; the stress reduction index is the opposite of the stress index. Step a3: After playing each material, collect the SAM psychological scale results currently input by the subject, update the basic information, calculate the current stress index until all the materials in this group of psychological stress reduction intervention plans are played. Step a4: Based on the basic information and the current physiological data, input them into the three-classification model and output a correction signal. Step a5: Adjust the psychological stress reduction intervention plan based on the correction signal. Step a6: Iteratively execute Step a2 until the stress index is less than the preset stress index threshold.

4. The system according to claim 1, wherein The specific calculation formula of the stress index is: SI = σ1 * BD + σ2 * PD; Among them, SI represents the stress index; BD represents the basic information index; σ1 represents the weight of the basic information index; PD represents the physiological data index; σ2 represents the weight of the physiological data index.

5. The system according to claim 1, characterized in that, The basic information index includes the SAM scale index, and the specific calculation method is: After converting the relaxation PAD value in the PAD emotion scale to the positive value range, obtain the relaxation standard value; collect the SAM scale data of the subject for different materials; calculate the Euclidean distance d between the SAM scale data and the relaxation standard value as the SAM index, and the specific formula is: ； Among them, , and respectively represent the scale data values of the subjects for different materials; , and respectively represent the relaxation standard value.

6. The system according to claim 1, characterized in that, The physiological data index includes the electroencephalogram index, and the specific calculation method is: Preprocess the EEG data; divide the EEG data according to the emotion type and quantitatively process it into emotion values; according to the measurement time, evenly divide the EEG data into several time periods, sort them according to the size of the emotion values in each time period, and then calculate the th subject's emotion average value for the th material The specific formula is: ； Among them, represents the maximum emotion value; represents the median emotion value; represents the minimum emotion value; For each material, calculate the average value of the subject's emotion values, and map it to the [0,1] interval through the min-max normalization method. The specific formula is: ; Among them, represents the original value of the current emotion value; represents the minimum value among the average emotion values corresponding to each material; represents the maximum value among the average emotion values corresponding to each material; represents the conversion result of the current emotion value; The physiological data index also includes the electrocardiogram index, and the specific calculation method is: Preprocess the electrocardiogram data; perform HRV data analysis on the electrocardiogram data, calculate HRV characteristic values, including SDNN value, RMSSD value, and LF / HF ratio; high HRV characteristic values correspond to the emotion classification characteristics of tension and anxiety, while low HRV characteristic values correspond to the emotion classification characteristics of calmness and tranquility. Based on the corresponding relationship of the emotion classification characteristics, construct an emotion model: quantify the HRV characteristic values into emotion values; use the min-max normalization method to map the emotion values to the [0,1] interval. The specific formula is: ； Among them, represents the current heart rate variability eigenvalue; represents the minimum value among all heart rate variability eigenvalues; represents the maximum value among all heart rate variability eigenvalues; represents the result after normalization.

7. The system according to claim 1, characterized in that The Spout component is also used to receive the eye movement data of the subject, and the eye movement data includes binocular coordinates and eye movement speed. In the Bolt component, the eye movement data of the subject is also transmitted backward in real time through the execute method. The feedback adjustment unit also collects the image playback coordinate range in the VR generation unit, converts the eye movement data into the human eye fixation coordinate range in the display area, then compares the image playback coordinate range with the human eye fixation coordinate range, calculates the concentration of the subject, and adjusts the psychological stress reduction intervention plan based on the concentration.

8. The system according to claim 7, wherein The specific calculation method of the concentration is as follows: Step b1: Collect the image playback coordinate range D1 of the current frame; based on the standard time sequence, collect the current binocular coordinates, and calculate the current human eye fixation coordinate range D2 in the display area through the principle of solid geometry. Step b2: Calculate the number MT of the same coordinates in D1 and D2; calculate the number MX of the coordinates in the smaller one of D1 and D2; take the ratio of MT to MX as the coincidence degree CH, and make a judgment: if the coincidence degree CH is greater than the preset coincidence degree threshold, it is determined that the human eye is fixated on the image; if the coincidence degree CH is less than or equal to the preset coincidence degree threshold, it is determined that the human eye is not fixated on the image. Step b3: Collect the image playback coordinate range D1' of the image after time t, and calculate the image displacement speed VS between D1 and D1', and the specific calculation formula is: VS = L / t; where L represents the relative displacement between D1 and D1'. Step b4: Based on the standard time sequence, collect the eye movement speed VY within time t. Step b5: Take the ratio of VS to VY as the follow-up degree SD, and make a judgment: if the follow-up degree SD is greater than the preset follow-up degree threshold, it is determined that the human eye always follows the image movement; if the follow-up degree SD is less than or equal to the preset follow-up degree threshold, it is determined that the human eye does not always follow the image movement. Step b6: Based on the coincidence degree CH and the follow-up degree SD, calculate the concentration ZZ, and the specific formula is: ZZ = q1 * CH + q2 * SD; where q1 represents the preset weight of the coincidence degree; q2 represents the preset weight of the follow-up degree.

9. The system according to claim 1, characterized in that, The Spout component is also used to receive the facial expression data of the subject, and the facial expression data includes the infrared image of the frown line and the infrared image of the crow's feet; the three-classification model also includes a convolutional neural network model, which inputs the frown line image and the infrared image of the crow's feet and outputs the facial expression stress classification result; the integrated model is also used to assign weights to the basic stress classification result, the electrocardiogram stress classification result, the electroencephalogram stress classification result, the skin conductance stress classification result and the facial expression stress classification result, and then outputs three correction signals: enhanced intervention, maintenance intervention or weakened intervention.

10. A personalized psychological stress reduction intervention device for implementing the system described in any one of claims 1-9, characterized in that, It includes a VR headset, an ECG device, an EEG device and a computing platform that are electrically connected to each other. The VR headset is used to display the VR scene. The ECG device refers to a wearable electrocardiogram monitoring device, which is used to monitor the heart rate variability in real time. The EEG device refers to a wearable electroencephalogram monitoring device, which is used to record the brain waves in real time. The computing platform includes a processor, a memory, and a bus. The memory stores instructions and data that are read by the processor. The processor is used to call the instructions and data in the memory. The bus is connected between the functional components to transmit information.

Images