Self-service psychological counseling method and system

Abstract

The invention belongs to the technical field of self-service psychological consultation, and provides a self-service psychological consultation method and system. The method comprises the steps of self-service psychological consultation equipment wearing, question and answer sheet reply data collection, physiological data and three-view video data collection, keyword feature extraction, score feature matching, physiological feature extraction, video feature recognition and input vector set construction. And analyzing the psychological health state of the user, and matching the psychological state report with the pre-intervention guidance suggestion file. According to the method, the source dimension of user state judgment is improved through the question and answer sheet reply data, the physiological data and the three-view video data, the recognition accuracy of the model is improved, meanwhile, a reference file is provided for the user by improving the corresponding pre-intervention guidance suggestion file, and psychological problems are further defended.

Description

Technical Field

[0001] The present invention relates to the technical field of self-help psychological counseling, and particularly to a self-help psychological counseling method and system. Background Art

[0002] Currently, young people are under pressure from aspects such as academics, society, and work, resulting in many psychological problems, which pose a serious threat to mental health. The sources of pressure faced by young people are diverse, including academic pressure, employment competition, economic burden, and interpersonal relationships. These pressures often lead to psychological problems such as anxiety and depression. Many studies have shown that the incidence of depression and anxiety among young people has increased significantly. Especially during the epidemic, many people experienced isolation and uncertainty, leading to an exacerbation of mental health problems. The popularity of social media, although providing a platform for communication, has also brought problems of comparison and anxiety. Young people often feel pressured on social media, affecting their self-identity and mental health. Although mental health problems are widespread, many young people still have concerns about seeking professional help, perhaps due to the stigmatization of mental health or doubts about the treatment effect. In some regions, mental health education is still insufficient, resulting in young people lacking the necessary knowledge and skills to cope with mental problems. The lack of family and social support systems makes it difficult for young people to obtain effective help and support when facing mental health problems. To solve the above problems, researchers have designed many related psychological counseling systems. For example, Patent CN113908398A provides a self-help psychological counseling device, which reduces the resistance of counseling personnel.

[0003] However, the current psychological counseling systems still have problems such as being unable to accurately and timely judge the user's emotions or having a single judgment data, resulting in a low guiding effect of the provided suggestions or a large applicable range and a low reference value, and a poor pre-intervention effect on mental problems. Summary of the Invention

[0004] In order to overcome the deficiencies of the prior art, the purpose of the present invention is to provide a self-help psychological counseling method and system, which solves the problem that the traditional method has a low accuracy in judging the user's mental state, resulting in a poor pre-intervention effect.

[0005] To achieve the above purpose, the present invention provides the following solutions:

[0006] A self-help psychological counseling method, including:

[0007] Guiding the user to wear a self-help psychological counseling device according to a preset prompt;

[0008] Collecting the user's situation statement information, providing a mental health questionnaire to the user, and collecting the questionnaire response data of the user to the mental health questionnaire; the mental health questionnaire includes: video format questions;

[0009] When collecting the case statement information and the questionnaire reply data, record the user's physiological data and three-view video data; the physiological data includes: heart rate data, respiratory rate data, muscle tension data, blood pressure data, and body temperature data;

[0010] Perform clustering analysis on the case statement information to obtain keyword features, match the questionnaire reply data according to a preset option scoring table to obtain scoring features, filter and extract features from the physiological data to obtain physiological features;

[0011] Input the three-view video data into a preset action behavior recognition model and facial expression recognition model to obtain video features;

[0012] Taking one of the keyword features as one step, combine the keyword features, the physiological features, and the video features, taking one of the scoring data in the scoring features as one step, combine the scoring features, the physiological features, and the video features to obtain an input vector set;

[0013] Input the input vector set into a pre-trained user state recognition model for state analysis to obtain the user's mental health state;

[0014] Match the user's mental health state with a mental state and early intervention plan library obtained based on expert experience and literature statistics to obtain a mental state report and a pre-intervention guidance suggestion document.

[0015] Preferably, the process of collecting the physiological data includes:

[0016] Use a preset flexible pressure sensing belt to collect the first pressure data of the user's upper arm, lower arm, thigh, and calf, and use a preset muscle tension formula to calculate all the first pressure data to obtain the muscle tension data;

[0017] Use a preset sphygmomanometer to collect the blood pressure information of the user's upper arm to obtain the blood pressure data;

[0018] Use a preset flexible heart rate and temperature measurement bracelet to collect the temperature information and the second pressure data of the user's lower arm, and perform frequency extraction on the second pressure data to obtain the heart rate data and the body temperature data;

[0019] Use a preset chest surround measurement flexible strap and an abdominal surround measurement flexible strap to collect the pressure change data of the user's chest and abdomen respectively, and perform filtering and frequency extraction on the pressure change data to obtain the respiratory rate data.

[0020] Preferably, the process of extracting the keyword features includes:

[0021] Segment the situation statement information, remove common words from the situation statement information according to a pre - counted common word list, and after the removal, use a clustering analysis algorithm to extract keywords from the situation statement information to obtain a target number of statement keywords;

[0022] Convert the statement keywords into serial number representations according to a preset keyword sequence list to obtain the keyword features.

[0023] Preferably, input the three - perspective video data into a preset action behavior recognition model and a facial expression recognition model to obtain video features, including:

[0024] Split the three - perspective video data by frame number, and determine three captured images at the same time frame in the three - perspective video data as a group;

[0025] Input the captured images of different groups into a pre - trained action mapping model in sequence according to the time frame order for preset virtual human action optimization to obtain complete virtual human mapped actions;

[0026] Extract the action data of the preset virtual human's arms and legs to obtain behavior data;

[0027] Input the behavior data into a pre - trained action behavior recognition model for behavior recognition to obtain the action features in the video features;

[0028] The training process of the action mapping model includes:

[0029] Collect three - perspective video data of volunteers of various body types performing under the guidance of a psychological expert in different psychological states;

[0030] According to an adversarial neural network, use the captured images from three perspectives of the same frame in the three - perspective video data as input and the position data of each node of a preset virtual human as output to construct a basic model architecture;

[0031] Determine the loss function of the basic model architecture; the loss function is:

[0032] L all =λ 1 L image +λ 2 L feature +λ 3 L adv ;

[0033] Wherein,

[0034]

[0035] Lall , L image , L feature , L adv are the total loss value, the virtual mapping image loss value, the feature loss, and the adversarial loss, respectively; λ 1 , λ 2 , λ 3 are the first weight, the second weight, and the third weight, respectively; are the virtual human image pixel distribution and the real human image pixel distribution, respectively; M is the number of feature layers; F(I virtual ) j , F(I real ) j are the extracted features of the virtual human image and the real human image in the j-th feature layer, respectively; respectively represent maximizing the probability of the discriminator for real images and maximizing the probability of the discriminator's wrong judgment of virtual images;

[0036] Input the captured images of the three perspectives of the same frame in the three-perspective video data into the basic model architecture to obtain virtual human node position data;

[0037] Update the virtual human node position data to the preset virtual human, and use a virtual camera to capture the three-perspective images of the preset virtual human; the angles and positions of the virtual camera are the same as those of the real camera used in the three-perspective video data;

[0038] Use the loss function to calculate the loss value between the three-perspective images and the corresponding captured images to obtain the total loss value;

[0039] Use the smooth training strategy and the total loss value to optimize the parameters of the basic model architecture until the total loss value is lower than the preset loss threshold, and then output the trained action mapping model.

[0040] Preferably, the training process of the action behavior recognition model includes:

[0041] Obtain the three-dimensional action data of the preset virtual human in different mental states through motion capture technology or by making three-dimensional virtual human animations;

[0042] Extract the action data of the arms and legs from the three-dimensional action data to obtain training data, and add action type labels to the training data;

[0043] Use a long short-term memory network to train the model with the action data in the training data as the input and the action type labels in the training data as the target data to obtain the action behavior recognition model.

[0044] Preferably, the user state recognition model adopts a Transformer model.

[0045] Preferably, the mental state and early intervention solution library includes: a mental state category column, a keyword column, a physiological feature column, a motion feature column, a scoring feature column, and a pre-intervention guidance and suggestion column.

[0046] Preferably, a self-service psychological counseling system includes: eight flexible pressure sensing bands, a blood pressure monitor, three cameras, a microphone, a touch display screen, a flexible heart rate and temperature measurement bracelet, a chest surrounding flexible strap, and an abdominal surrounding flexible strap; a plurality of pressure sensors are arranged on the inner sides of the pressure sensing bands, the chest surrounding flexible strap, and the abdominal surrounding flexible strap; a pressure sensor and a temperature sensor are embedded on the inner side of the flexible heart rate and temperature measurement bracelet;

[0047] The three cameras are respectively arranged at the front, side, and obliquely upper positions of the user; the flexible pressure sensing bands are respectively attached to the upper arms, lower arms, thighs, and calves of the user; the blood pressure monitor is strapped to the left or right arm of the user; the microphone is fixed at the edge of the touch display screen; the flexible heart rate and temperature measurement bracelet is attached to the wrist of the user; the chest surrounding flexible strap and the abdominal surrounding flexible strap are respectively strapped to the chest and abdomen of the user.

[0048] The present invention discloses the following technical effects:

[0049] The present invention provides a self-service psychological counseling method and system. By collecting and model analyzing the reply data of the questionnaire, physiological data, and three-perspective video data, the problem that the traditional method has a low accuracy in judging the user's mental state and thus a poor pre-intervention effect is solved, and accurate recognition of the user's mental state and provision of accurate guidance are achieved. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for use in the embodiments. Obviously, the drawings described below 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.

[0051] Figure 1 It is a schematic diagram of the self-service psychological counseling process provided by the embodiment of the present invention;

[0052] Figure 2 It is a schematic diagram of the training process of the motion mapping model provided by the embodiment of the present invention;

[0053] Figure 3Schematic diagrams of the pressure sensing belt, the flexible chest circumferential measurement strap, and the flexible abdominal circumferential measurement strap provided by the embodiments of the present invention. Detailed implementation manners

[0054] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0055] The object of the present invention is to provide a self-service psychological counseling method and system to solve the problem that the traditional method has a low accuracy in judging the psychological state of users, resulting in a poor pre-intervention effect.

[0056] To make the above objects, features, and advantages of the present invention more obvious and understandable, the present invention will be further described in detail below in conjunction with the accompanying drawings and specific implementation manners.

[0057] Figure 1 Schematic diagram of the self-service psychological counseling process provided by the embodiments of the present invention. As Figure 1 shown, the present invention provides a self-service psychological counseling method, including:

[0058] Step 100: Guide the user to wear the self-service psychological counseling device according to the preset prompts.

[0059] Step 200: Collect the user's situation statement information, provide the user with a mental health questionnaire, and collect the questionnaire response data of the user to the mental health questionnaire; the mental health questionnaire includes: questions in video format.

[0060] Step 300: When collecting the situation statement information and the questionnaire response data, record the user's physiological data and three-view video data; the physiological data includes: heart rate data, respiratory rate data, muscle tension data, blood pressure data, and body temperature data.

[0061] Step 400: Perform clustering analysis on the situation statement information to obtain keyword features, match the questionnaire response data according to the preset option scoring table to obtain scoring features, and filter and extract features from the physiological data to obtain physiological features.

[0062] Step 500: Input the three-view video data into the preset action behavior recognition model and facial expression recognition model to obtain video features.

[0063] Step 600: Taking one of the keyword features as one step, combining the keyword feature, the physiological feature, and the video feature, and taking one scoring data in the scoring feature as one step, combining the scoring feature, the physiological feature, and the video feature, to obtain an input vector set;

[0064] Step 700: Inputting the input vector set into a pre-trained user state recognition model for state analysis to obtain the user's mental health state;

[0065] Step 800: Matching the user's mental health state with a mental state and early intervention plan library obtained based on expert experience and literature statistics to obtain a mental state report and a pre-intervention guidance recommendation document.

[0066] Preferably, the process of collecting the physiological data includes:

[0067] Using a preset flexible pressure sensing belt to collect the first pressure data of the user's upper arm, lower arm, thigh, and calf, and using a preset muscle tension formula to calculate all the first pressure data to obtain the muscle tension data;

[0068] Using a preset sphygmomanometer to collect the blood pressure information of the user's upper arm to obtain the blood pressure data;

[0069] Using a preset flexible heart rate and temperature measurement bracelet to collect the temperature information and the second pressure data of the user's lower arm, and performing frequency extraction on the second pressure data to obtain the heart rate data and the body temperature data;

[0070] Using a preset flexible chest circumference measurement band and a flexible abdominal circumference measurement band to collect the pressure change data of the user's chest and abdomen respectively, and performing filtering and frequency extraction on the pressure change data to obtain the respiratory rate data.

[0071] Furthermore, the process of extracting the keyword feature includes:

[0072] Performing word segmentation on the situation statement information, removing common words from the situation statement information according to a pre-statistical common word list, and after the removal, using a clustering analysis algorithm to extract keywords from the situation statement information to obtain a target number of statement keywords;

[0073] Converting the statement keywords into serial number representations according to a preset keyword sequence table to obtain the keyword feature.

[0074] Specifically, inputting the three-view video data into a preset action behavior recognition model and a facial expression recognition model to obtain video features, including:

[0075] Split the three-view video data by the number of frames, and determine the three captured images in the same time frame of the three-view video data as a group;

[0076] Input the captured images of different groups into the pre-trained action mapping model in sequence according to the time frame order for preset virtual human action optimization, and obtain the complete virtual human mapped actions;

[0077] Extract the action data of the preset virtual human's arms and legs to obtain the behavior data;

[0078] Input the behavior data into the pre-trained action behavior recognition model for behavior recognition to obtain the action features in the video features;

[0079] Reference Figure 2 , the training process of the action mapping model includes:

[0080] Step 501: Collect three-view video data of volunteers of various body types performing under the guidance of a psychologist in different mental states;

[0081] Step 502: Based on the adversarial neural network, use the captured images from the three perspectives in the same frame of the three-view video data as the input and the position data of each node of the preset virtual human as the output to construct the basic model architecture;

[0082] Step 503: Determine the loss function of the basic model architecture; the loss function is:

[0083] L all =λ 1 L image +λ 2 L feature +λ 3 L adv ;

[0084] Among them,

[0085]

[0086] L all 、L image 、L feature 、L adv are the total loss value, the virtual mapping image loss value, the feature loss, and the adversarial loss respectively; λ 1 、λ 2 、λ 3 are the first weight, the second weight, and the third weight respectively; are the virtual human image pixel distribution and the real human image pixel distribution respectively; M is the number of feature layers; F(I virtual ) j 、F(Ireal ) j They are the extracted features of the virtual human image and the real human image in the j-th feature layer respectively. They respectively represent the probability of the maximized discriminator for real images and the probability of the misjudgment of the maximized discriminator for virtual images.

[0087] Step 504: Input the captured images of the three perspectives in the same frame of the three-perspective video data into the basic model architecture to obtain the virtual human node position data.

[0088] Step 505: Update the virtual human node position data to the preset virtual human, and use a virtual camera to capture the three-perspective images of the preset virtual human; the angles and positions of the virtual camera are the same as those of the real camera used in the three-perspective video data.

[0089] Step 506: Use the loss function to calculate the loss value between the three-perspective images and the corresponding captured images to obtain the total loss value.

[0090] Use the smooth training strategy and the total loss value to optimize the parameters of the basic model architecture until the total loss value is lower than the preset loss threshold, and then output the trained action mapping model.

[0091] Furthermore, the training process of the action behavior recognition model includes:

[0092] Obtain the three-dimensional action data of the preset virtual human in different mental states through motion capture technology or by making three-dimensional virtual human animations.

[0093] Extract the action data of the arms and legs from the three-dimensional action data to obtain training data, and add action type labels to the training data.

[0094] Use a long short-term memory network to train the model with the action data in the training data as the input and the action type labels in the training data as the target data to obtain the action behavior recognition model.

[0095] Optionally, the user state recognition model adopts a Transformer model.

[0096] Preferably, the mental state and early intervention plan library includes: a mental state category column, a keyword column, a physiological feature column, an action feature column, a scoring feature column, and a pre-intervention guidance and suggestion column.

[0097] Further, a self-service psychological counseling system includes: eight flexible pressure sensing bands, a blood pressure monitor, three cameras, a microphone, a touch display screen, a flexible heart rate and temperature measurement bracelet, a flexible chest circumference measurement strap, and a flexible abdominal circumference measurement strap; a plurality of pressure sensors are provided on the inner sides of the pressure sensing bands, the flexible chest circumference measurement strap, and the flexible abdominal circumference measurement strap; a pressure sensor and a temperature sensor are embedded on the inner side of the flexible heart rate and temperature measurement bracelet;

[0098] The three cameras are respectively arranged at the front, side, and obliquely upward positions of the user; the flexible pressure sensing bands are respectively attached to the upper arms, lower arms, thighs, and calves of the user; the blood pressure monitor is strapped to the left arm or the right arm of the user; the microphone is fixed to the edge of the touch display screen; the flexible heart rate and temperature measurement bracelet is attached to the wrist of the user; the flexible chest circumference measurement strap and the flexible abdominal circumference measurement strap are respectively strapped to the chest and abdomen of the user.

[0099] Referring to Table 1, the following provides an example of a psychological state and early intervention plan library in this embodiment. This plan is for reference only. In actual application, a more detailed plan library needs to be formulated in combination with research data, literature analysis, and expert experience to provide guiding suggestions for users.

[0100] Table 1

[0101]

[0102]

[0103]

[0104] Reference Figure 3 , buckles are provided on both sides of the pressure sensing bands, the flexible chest circumference measurement strap, and the flexible abdominal circumference measurement strap, and pressure sensors are evenly arranged in the middle. Working principle: Bind the device to the measured part, such as the upper arm, lower arm, or thigh, calf, tie it tightly, and then record the sensing data of each pressure sensor. According to the collected sensing data, judge the muscle tension state of the user. When the data collected by all sensors fluctuates relatively smoothly, it is judged that the user's muscles are in a relaxed state. On the contrary, if the collected data shows violent fluctuations, it is determined that the user is relatively tense. Specifically, the judgment criteria and grading are set according to actual needs. The judgment of the breathing rate mainly judges a certain process in exhalation or inhalation. After filtering the collected pressure data, it is judged that two numerical decline intervals are a cycle, set the cycle judgment duration, count the average value of the cycle duration during this period, and calculate the breathing rate within this time period based on this.

[0105] Optionally, this embodiment provides a consultation environment for users, including: a code scanning registration area, a consultation seat, and a report QR code prompt area. The three areas are guided by voice, without personnel intervention to reduce the psychological pressure of users. The devices in the three areas are connected to the background system, and the background system is used for data analysis and issuance of guiding opinions. After the user scans the code for registration, they are guided by voice to the consultation seat to start data collection and upload. The touch display screen will guide the customer to make a situation description and provide relevant psychological test questions for the user. After the customer's consultation ends, the consultation task is completed by scanning the report QR code.

[0106] The beneficial effects of the present invention are as follows:

[0107] The present invention improves the source dimension of user state judgment through questionnaire reply data, physiological data, and three-view video data, improves the recognition accuracy of the model, and at the same time provides a reference document for users by improving the corresponding pre-intervention guiding suggestion document, further preventing the occurrence of psychological problems.

[0108] In this specification, each embodiment is described in a progressive manner. The key point of each embodiment is to illustrate the differences from other embodiments. The same or similar parts among the embodiments can be referred to each other.

[0109] Specific examples are used in this article to elaborate on the principles and implementation manners of the present invention. The description of the above embodiments is only used to help understand the method of the present invention and its core idea; at the same time, for those of ordinary skill in the art, according to the idea of the present invention, there will be changes in the specific implementation manners and application scopes. In summary, the content of this specification should not be construed as a limitation to the present invention.

Claims

1. A self-help psychological counseling method, characterized in that: include: Instruct users to wear self-service psychological counseling equipment according to preset prompts; Collecting user's situation statement information, providing the user with a mental health questionnaire, and collecting the user's response data to the mental health questionnaire; Said mental health questionnaire includes: video format questions; When collecting the situation statement information and the questionnaire response data, recording the user's physiological data and three-view video data; the physiological data includes: heart rate data, respiratory rate data, muscle tension data, blood pressure data and body temperature data; Performing cluster analysis on the situation statement information to obtain keyword features, matching the question-answer form response data according to a preset option scoring table to obtain scoring features, filtering and feature extracting the physiological data to obtain physiological features; Inputting the three-view video data into a preset action behavior recognition model and a facial expression recognition model to obtain video features; Taking one of the keyword features as a step, combining the keyword feature, the physiological feature and the video feature, taking one of the scoring data of the scoring feature as a step, combining the scoring feature, the physiological feature and the video feature, to obtain an input vector set; Inputting the input vector set into a pre-trained user state recognition model to perform state analysis to obtain the user's mental health state; The user's mental health status is matched with the mental status and pre-intervention program library obtained based on expert experience and literature statistics to obtain a mental status report and pre-intervention guidance and suggestion document.

2. A self-help psychological consultation method according to claim 1, characterized in that: The physiological data collection process includes: Using a preset flexible pressure sensor belt to collect first pressure data of the user's upper arm, lower arm, thigh and calf, and using a preset muscle tension formula to calculate all of the first pressure data to obtain the muscle tension data; Using a preset blood pressure meter to collect blood pressure information of the user's upper arm to obtain the blood pressure data; Using a preset flexible heart rate and temperature measurement bracelet to collect the temperature information of the user's lower arm and the second pressure data, and performing frequency extraction on the second pressure data to obtain the heart rate data and the body temperature data; The pressure change data of the user's chest and abdomen are collected respectively by using a preset chest-surrounding flexible measurement strap and an abdomen-surrounding flexible measurement strap, and the pressure change data are filtered and frequency extracted to obtain the breathing frequency data.

3. A self-help psychological consultation method according to claim 1, characterized in that: The keyword feature extraction process includes: Segmenting the situation statement information, and removing common words from the situation statement information according to a pre-statisticed common word list, and after the removal, extracting keywords from the situation statement information using a clustering analysis algorithm to obtain a target number of statement keywords; The statement keywords are converted into serial number representations according to a preset keyword sequence table to obtain the keyword features.

4. A self-help psychological consultation method according to claim 1, characterized in that: The three-view video data is input into a preset action behavior recognition model and a facial expression recognition model to obtain video features, including: Splitting the three-view video data according to the number of frames, and determining three captured images of the same time frame in the three-view video data as a group; Inputting the captured images of different groups into the pre-trained motion mapping model in sequence according to the time frame order to optimize the preset virtual human motion, so as to obtain a complete virtual human mapping motion; Extract the motion data of the preset virtual human's arms and legs to obtain the behavior data; Inputting the behavior data into a pre-trained action behavior recognition model for behavior recognition to obtain action features in the video features; The training process of the action mapping model includes: Collect three-view video data of volunteers of various body types performing different psychological states under the guidance of psychological experts; According to the adversarial neural network, the captured images of the three perspectives of the same frame in the three-perspective video data are used as input, and the position data of each node of the preset virtual person is used as output to construct a basic model architecture; Determine the loss function of the basic model architecture; the loss function is: L all =λ1L image +λ2L feature +λ3L adv ; in, L all , L image , L feature , L adv They are the total loss value, virtual mapping image loss value, feature loss, and adversarial loss respectively; λ1, λ2, and λ3 are the first weight, the second weight, and the third weight respectively; are the pixel distribution of virtual human image and real human image respectively; M is the number of feature layers; F(I virtual ) j 、F(I real ) j are the extracted features of the virtual human image and the real human image in the jth feature layer respectively; They represent maximizing the probability of the discriminator for real images and maximizing the probability of the discriminator's misjudgment of virtual images respectively; Input the captured images of three perspectives of the same frame in the three-perspective video data into the basic model framework to obtain virtual human node position data; The virtual person node position data is updated to the preset virtual person, and a three-view image of the preset virtual person is captured by a virtual camera; the angle position of the virtual camera and the real camera used for the three-view video data are consistent; Calculating the loss values ​​of the three-view images and the captured images corresponding to the three-view images by using the loss function to obtain the total loss value; The basic model architecture is optimized by using a smooth training strategy and the total loss value until the total loss value is lower than a preset loss threshold, and then the trained action mapping model is output.

5. A self-help psychological consultation method according to claim 1, characterized in that: The training process of the action behavior recognition model includes: Obtaining 3D motion data of a preset virtual human in different psychological states through motion capture technology or making 3D virtual human animations; Extracting motion data about arms and legs from the three-dimensional motion data to obtain training data, and adding motion type labels to the training data; The action behavior recognition model is obtained by using a long short-term memory network with the action data in the training data as input and the action type label in the training data as target data for model training.

6. A self-help psychological consultation method according to claim 1, characterized in that: The user state recognition model adopts the Transformer model.

7. A self-help psychological consultation method according to claim 1, characterized in that: The psychological state and early intervention program library includes: a psychological state category column, a keyword column, a physiological feature column, an action feature column, a scoring feature column, and a pre-intervention guidance suggestion column.

8. A self-service psychological consultation system, characterized in that: include: Eight flexible pressure sensing belts, a blood pressure monitor, three cameras, a microphone, a touch screen, a flexible heart rate and temperature measuring bracelet, a flexible chest measurement strap and a flexible abdomen measurement strap; the pressure sensing belt, the flexible chest measurement strap and the flexible abdomen measurement strap are all provided with a plurality of pressure sensors on the inner side; the flexible heart rate and temperature measuring bracelet is embedded with a pressure sensor and a temperature sensor; The three cameras are respectively arranged on the front, side and upper side of the user; the flexible pressure sensing belt is respectively attached to the upper arm, lower arm, thigh and calf of the user; the blood pressure monitor is tied to the left arm or right arm of the user; the microphone is fixed on the edge of the touch display screen; The flexible heart rate and temperature measurement bracelet is attached to the wrist of the user; the flexible chest-encircling measurement strap and the flexible abdomen-encircling measurement strap are respectively tied to the chest and abdomen of the user.

Images