AI dynamic emotion recognition multi-dimensional directional training system

Abstract

The invention provides an AI dynamic emotion recognition multi-dimensional directional training system, and relates to the technical field of computer vision and artificial intelligence, the AI dynamic emotion recognition multi-dimensional directional training system comprises a plurality of emotion collection and training terminals at the front end and a background server, the emotion collection and training terminals are used for carrying out emotion detection on a detected person, the terminals analyze face images, and the face images are sent to the background server; obtaining multi-dimensional emotion data of the tested person; the multi-dimensional emotion data is uploaded to a background server, and the background server analyzes the current emotion state of the tested person according to the received multi-dimensional emotion data, and pushes different directional training tasks to the corresponding emotion acquisition and training terminals for the tested person in different states; after the emotion acquisition and training terminal receives the directional training task, the tested person completes corresponding training content according to guidance; the psychological quality of the tested person is improved in a targeted mode or the bad emotional state is improved.

Description

Technical Field

[0001] The present invention proposes an AI dynamic emotion recognition multi-dimensional directional training system, which relates to the fields of computer vision technology and artificial intelligence technology. Background Art

[0002] In the current social environment, people live and work at a fast pace and with great pressure. Affected by factors such as work environment, life pressure, and social relationships, people are prone to negative emotions, such as irritability, irritability, depression, and negativity. Bad emotions can affect people's physical health and mental state, and may even lead to violent conflicts, suicide and other serious incidents. If we can quickly identify people's negative emotional states and conduct timely intervention and guidance, we can improve people's negative mental states and sudden emotional problems in a targeted manner, predict or even avoid sudden events in advance, and reduce the possibility of causing serious consequences.

[0003] Traditional emotion assessment methods often require face-to-face diagnosis by psychologists, or filling out specific scales for assessment, which takes a long time and is not suitable for scenarios that require large-scale rapid screening. Emotional intervention methods often require venues, props or on-site counseling by psychologists, require active cooperation from personnel, and are not timely. Among the existing new technologies, there are related technical means for psychological assessment through video analysis of micro-expressions and micro-movements or using sound features. Non-contact detection can be achieved, but the existing technology generally requires the person being tested to stand still in front of the camera for about 1-3 minutes, or answer certain specific questions to collect voiceprints. The cooperation of the person being tested is still required. It takes a certain amount of collection time, cannot achieve real-time monitoring, and is difficult to detect sudden emotional problems in a timely manner. In addition, the existing technology lacks supporting means for intervention and treatment after detection, and it is difficult to fundamentally adjust the emotions and psychological state of the person being tested. Summary of the invention

[0004] In order to overcome the problems existing in the related technologies, this application provides an AI dynamic emotion recognition multi-dimensional directional training system, which can quickly and non-contactly detect the dynamic emotions and psychological states of people from multiple dimensions, and propose relevant directional training and guidance content for the different states of the tested people, thereby improving bad emotional states and improving mental health levels. It can realize rapid emotional screening, dynamic emotional and mental health regulation in large-scale scenarios, and analyze and display data and provide information warnings.

[0005] An AI dynamic emotion recognition multi-dimensional directional training system, comprising: multiple emotion collection and training terminals at the front end and a back-end server;

[0006] The emotion collection and training terminal is used to perform emotion detection on the person being tested, and the terminal analyzes the facial image to obtain multi-dimensional emotion data of the person being tested;

[0007] The multi-dimensional emotion data is uploaded to the backend server, and the backend server analyzes the current emotional state of the measured person according to the received multi-dimensional emotion data, and pushes different directional training tasks to the corresponding emotion collection and training terminal for the measured persons in different states;

[0008] After the emotion collection and training terminal receives the directional training task, the person being tested completes the corresponding training content according to the guidance.

[0009] Furthermore, the emotion collection and training terminal includes a camera, a display, an emotion data calculation module, and a control and data transmission module;

[0010] The camera is used to collect the facial image of the person being tested;

[0011] The display is used to display the images captured by the camera and the orientation training content;

[0012] The emotion data calculation module identifies the feature points of the human face and obtains the dynamic emotion feature data of the measured person within a fixed time;

[0013] The control and data transmission module uploads the dynamic emotion feature data to the background server, and receives the directional training content and control instructions returned by the background server.

[0014] Furthermore, the emotion data calculation module identifies the feature points of the face, obtains the positions of the facial feature points in each frame image and the relative changes between frames, represents the change rules of the facial feature points in the form of spatial vectors, and forms a facial feature change matrix; the facial feature change matrix is ​​input into the emotion recognition feature model for calculation to obtain the emotional feature data of the person to be tested.

[0015] Furthermore, a coordinate system is established with the pixel point at the lower left corner of the face image as the origin. The coordinate value of the nth feature point in the face image is: (x n ,y n );

[0016] The coordinate values ​​of the feature points are standardized to obtain the relative position coordinates (x zn ,y zn ):

[0017]

[0018]

[0019] Among them, xzn ,y zn are the normalized horizontal and vertical coordinate values; mean(x) and mean(y) are the average values ​​of the horizontal and vertical coordinates of all feature points; std(x) and std(y) are the standard deviations of the horizontal and vertical coordinates of all feature points.

[0020] Furthermore, the relative position coordinates of the 78 feature points of the current frame image are selected to form a 156-dimensional feature point space vector Ve:

[0021] Ve={(x z1 ,y z1 ),(x z2 ,y z2 ),…,(x z78 ,y z78 )}

[0022] For the previous frame image, the feature point space vector V is processed in the same way e ';

[0023] Subtract the feature point space vectors of the previous and next frames to obtain the face feature change matrix V c :

[0024] V c =V e '-V e ={(x z1 '-x z1 ,y z1 '-y z1 ),…,(x z78 '-x z78 ,y z78 '-y z78 )};

[0025] The face feature change matrix V c The emotion recognition feature model is input for calculation to obtain the emotion feature data of the person being tested.

[0026] Furthermore, the emotion feature recognition model includes a spatial feature model, a change feature model, and a regional feature model:

[0027] The input of the spatial feature model is: the relative position coordinate matrix Ve of facial feature points extracted from each frame image; the output is: spatial feature emotion label

[0028] The input of the variation feature model is: the variation matrix V of the facial features between each frame image and its previous frame image c ; Output: Change feature emotion label

[0029] The input of the regional feature model is: the regional images of each frame image segmented according to the distribution of facial features, including the left eye image I eyel , right eye image I eyer , nose image I nos , left cheek image I facl , right cheek image I facr , mouth image I mou ;

[0030] In the feature layer of the regional feature model, the output features of each segmentation image model network are spliced ​​to obtain the splicing result ConFea:

[0031] ConFea=[F(I eyel ),F(I eyer ),F(I nos ),F(I facl ),F(I facr ),F(I mou )]

[0032] Where F(·) is the shallow convolutional neural network of each region image;

[0033] Output region feature emotion labels

[0034]

[0035] Among them, FC(·) is a fully connected layer network, and softmax(·) is a normalized exponential function.

[0036] Furthermore, the emotion label of the current frame is obtained through weighted cumulative calculation

[0037]

[0038] Among them, W s , W c , W a They are spatial feature weight value, change feature weight value and regional feature weight value respectively.

[0039] Furthermore, the emotion data calculation module collects multiple frames of facial images within a period of time, obtains multiple emotion labels corresponding to the multiple frames through the emotion recognition feature model, accumulates the multiple emotion labels and takes the average value to obtain the dynamic emotion feature data of the person being tested within the period of time.

[0040] Furthermore, the emotion label is a collection of 12-dimensional emotion feature parameters, expressed in the form of 12-dimensional vectors, representing: aggressiveness, pressure, tension, suspicion, balance, self-confidence, energy, self-regulation, depression, neuroticism, inhibition, and happiness.

[0041] The technical solution provided by this application may include the following beneficial effects: This application can quickly screen the dynamic psychological and emotional state of personnel in a non-contact manner, quickly intervene and conduct targeted training for negative emotions and psychological problems, and timely warn of abnormal emotions. It can quickly adjust the psychological level and emotional state of personnel in a large range, and to a certain extent avoid sudden malicious incidents caused by negative emotions. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings in the following description are only some embodiments of the present invention. For ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without paying creative labor.

[0043] Figure 1 This is the overall architecture diagram of the AI ​​dynamic emotion recognition multi-dimensional directional training system of the present invention;

[0044] Figure 2 is a front structural schematic diagram of the emotion collection and training terminal of the present invention;

[0045] Figure 3 is a schematic diagram of the back structure of the emotion collection and training terminal of the present invention;

[0046] Figure 4 It is a schematic diagram of the workflow of the AI ​​dynamic emotion recognition multi-dimensional directional training system of the present invention. DETAILED DESCRIPTION

[0047] In order to make the purpose, technical solution and advantages of the embodiments of the present application clearer, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0048] In the drawings of the specific embodiments of the present invention, in order to better and more clearly describe the working principles of the various components in the system, the connection relationship of the various parts in the device is shown, which only clearly distinguishes the relative position relationship between the various components, and cannot constitute a limitation on the signal transmission direction, connection sequence and structural size, size and shape of each part within the component or structure.

[0049] Figure 1This is the overall architecture diagram of an AI dynamic emotion recognition multi-dimensional directional training system of the present invention. The AI ​​dynamic emotion recognition multi-dimensional directional training system includes multiple emotion collection and training terminals at the front end and a back-end server. The multiple emotion collection and training terminals at the front end and the back-end server form a local area network through Ethernet and network switching equipment to transmit data and control information.

[0050] The backend server is equipped with data service and management programs, which are used to analyze and process the emotion feature data uploaded by the emotion collection and training terminal. According to the distribution of scores of different dimensions in the emotion feature data, the current emotional state of the tested person is analyzed, and the corresponding targeted training content is generated according to the analysis results. The targeted training content is pushed to the emotion collection and training terminal, and displayed on the display to guide the tested person to complete the training task.

[0051] like Figure 2 and Figure 3 As shown, they are schematic diagrams of the front and back structures of the emotion collection and training terminal, respectively. The emotion collection and training terminal includes a housing 1, a camera 2, a display 3, an emotion data calculation module 4, and a control and data transmission module 5.

[0052] The camera 2 is installed at the front end of the housing, illuminating the direction of the person being measured, and is used to collect the facial image of the person being measured.

[0053] The display 3 is installed on the front of the housing and is used to display the images captured by the camera and the orientation training content.

[0054] The emotion data calculation module 4 and the control and data transmission module 5 are installed in the same host and can be composed of a device with data access and calculation functions such as a tablet computer, a computer host, and the application software installed therein.

[0055] The emotion data calculation module 4 identifies the facial feature points, obtains the positions of the facial feature points in each frame image and the relative changes between frames, and represents the change rules of the facial feature points in the form of space vectors to form a facial feature change matrix.

[0056] Specifically, for each frame of the image collected, a face detection algorithm is used to obtain a rectangular face area in the image, assuming that the area is a rectangular image area with X pixels in the horizontal direction and Y pixels in the vertical direction. The face area is used as the detection object to detect the position of the facial feature points.

[0057] First, the position of facial contour feature points is detected and a deep convolutional neural network is constructed. The input of the deep convolutional neural network is the face area image and the output is the position of facial contour feature points. A total of 17 facial contour feature points are taken. After obtaining the facial contour feature points, the face is segmented into six areas according to the position of the facial features: left eye, right eye, nose, left cheek, right cheek, and mouth.

[0058] Six different deep neural networks were constructed. The input of the deep neural network was the cropped image of each region (left eye, right eye, nose, left cheek, right cheek, and mouth), and the output was the position coordinates of the facial contour feature points in the region. For each eye region, 5 feature points were taken on the eyebrows and 6 feature points on the eyes; 5 feature points were taken on each cheek region; 9 feature points were taken on the nose; and 20 feature points were taken on the mouth. A total of 61 feature points of the facial features were obtained.

[0059] For each frame of the face image, after feature point detection, a total of 78 facial feature point position coordinate values ​​are obtained, and each feature point is numbered from 1 to 78. A coordinate system is established with the pixel point at the lower left corner of the face image as the origin. The coordinate value of the nth feature point in the face image is: (x n ,y n );

[0060] In order to represent the relative positions of the feature points of the face, the Z-socore method is used to standardize the coordinate values ​​of the feature points.

[0061]

[0062] Among them, x zn ,y zn is the normalized horizontal and vertical coordinate value, and the relative position coordinate (x zn ,y zn );

[0063] mean(x) and mean(y) are the average values ​​of the horizontal and vertical coordinates of all feature points; std(x) and std(y) are the standard deviations of the horizontal and vertical coordinates of all feature points.

[0064] The relative position coordinates of the 78 feature points of the current frame image are selected to form a 156-dimensional feature point space vector Ve:

[0065] Ve={(x z1 ,y z1 ),(x z2 ,y z2 ),…,(x z78 ,y z78 )}

[0066] For the previous frame image, the feature point space vector V is processed in the same way e ';

[0067] Subtracting the feature point space vectors of the previous and next frames can obtain the face feature change matrix V c .

[0068] V c =Ve '-V e ={(x z1 '-x z1 ,y z1 '-y z1 ),…,(x z78 '-x z78 ,y z78 '-y z78 )}

[0069] The obtained facial feature change matrix V c The emotion recognition feature model is input for calculation to obtain the emotion feature data of the person. The emotion data calculation module has a built-in emotion recognition feature model obtained through an artificial intelligence algorithm. It is obtained by inputting a large number of facial images with different emotion features through annotation verification and model training.

[0070] Specifically, the emotion feature recognition model includes three sub-models: spatial feature model, change feature model, and regional feature model. The input and output of each model are:

[0071] (1) Spatial feature model:

[0072] Input: relative position coordinate matrix Ve of facial feature points extracted from each frame image;

[0073] Output: Spatial feature emotion label

[0074] (2) Change feature model:

[0075] Input: The change matrix V of facial features between each frame image and the previous frame image c ;

[0076] Output: Change feature emotion label

[0077] (3) Regional characteristic model:

[0078] Input: The regional images of each frame image segmented according to the distribution of facial features, including the left eye image I eyel , right eye image I eyer , nose image I nos , left cheek image I facl , right cheek image I facr , mouth image I mou

[0079] At the feature level of the model, the output features of each segmentation image model network are spliced:

[0080] ConFea=[F(I eyel ),F(I eyer ),F(Inos ),F(I facl ),F(I facr ),F(I mou )]

[0081] Where F(·) is the shallow convolutional neural network of each region image;

[0082] Then, the regional feature emotion label is obtained by feature concatenation:

[0083]

[0084] Among them, FC(·) is a fully connected layer network, and softmax(·) is a normalized exponential function.

[0085] Output: Regional feature sentiment labels

[0086] After obtaining the emotion label output by each sub-model, the emotion label of the current frame is finally obtained through weighted accumulation.

[0087]

[0088] Among them, W s , W c , W a They are spatial feature weight value, change feature weight value and regional feature weight value respectively.

[0089] W s +W c +W a =1.

[0090] The above emotional labels are a collection of 12-dimensional emotional feature parameters, expressed in the form of 12-dimensional vectors. Each dimension uses an integer score of 0-100 to represent the intensity of the emotional dimension. They represent: aggressiveness, pressure, tension, suspicion, balance, confidence, energy, self-regulation, depression, neuroticism, inhibition, and happiness.

[0091] The emotion data calculation module collects facial images within a period of time, and obtains the emotion labels of each frame through the above process. After accumulating the emotion labels, the average value is taken to obtain the dynamic emotion feature data of the person being tested during the period of time. It is also represented in the form of a 12-dimensional vector.

[0092] The control and data transmission module 5 controls the overall workflow of the terminal, uploads the dynamic emotion feature data to the background server, receives the directional training content and control instructions returned by the background server, controls the display to display the training content on the current screen, and guides the tested person to perform the directional emotion training task.

[0093] For example, the emotional characteristic data is multi-dimensional data represented by a score of 0-100, with E1-E 12 The higher the score, the deeper the emotional or psychological state in this dimension. The specific meanings of the data in each dimension are as follows:

[0094] E1: Aggression. When E1>70, it means the person being tested is in a state of irritability.

[0095] E2: stress. When E2>60, it means the person being tested is at a high stress level.

[0096] E3: Tension. When E3>70, it means that the person being tested is nervous and anxious.

[0097] E4: Suspicious. When E4>70, it means the person being tested lacks a sense of security.

[0098] E5: Balance. When E5 < 40, it means that the person being tested is in a state of physical and mental imbalance.

[0099] E6: Confidence. When E6 < 40, it means that the person being tested has insufficient self-confidence and low self-recognition.

[0100] E7: Energy. When E7 < 50, it means that the person being tested lacks vitality and is physically tired.

[0101] E8: Self-regulation. When E8 is less than 40, it means that the person being tested has low self-control.

[0102] E9: Depressed. When E9>60, it means that the person being tested is in a depressed and lost state.

[0103] E 10 : Neurotic, when E 10 >80, indicating that the person being tested has obsessive-compulsive tendencies;

[0104] E 11 :Inhibit, when E 11 >70, indicating that the person being tested is in a state of depression and self-isolation;

[0105] E 12 : Happiness, when E 12 <40 means that the person being tested has low happiness and is dissatisfied with his or her life.

[0106] The content of targeted training can be multifaceted, including encouraging texts, physical exercise programs, communication tasks, reading, listening to music, etc. When the psychological and emotional data of the tested person is in a relatively stable and healthy state, the training content can also be sharing joy, helping others, rewards, etc.

[0107] Table 1 exemplarily lists the contents of targeted training corresponding to different emotional feature data states. In practical applications, a variety of different targeted training contents can be added as needed.

[0108]

[0109]

[0110] Figure 4 It is a schematic diagram of the workflow of an AI dynamic emotion recognition multi-dimensional directional training system shown in an embodiment of the present application.

[0111] See also Figure 4 , the system workflow is as follows:

[0112] The person being tested stands in front of the emotion collection and training terminal and starts the emotion detection;

[0113] The terminal uses a camera to collect facial images for 5 seconds. During the collection process, the display prompts the collection time and the position of the face in the picture to prevent the person being measured from leaving early or shifting his position.

[0114] Obtain the identity information of the person being tested through face recognition;

[0115] The terminal analyzes the facial image and obtains the multi-dimensional emotional data of the person being tested;

[0116] Upload the identity information and multi-dimensional emotional data of the tested person to the backend server;

[0117] The backend server receives the multi-dimensional emotional data of the personnel, organizes and stores the data, and can also display the existing data in the form of charts and other forms as needed;

[0118] The backend server analyzes the current psychological and emotional state of the person being tested based on the received multi-dimensional emotional data, and pushes different directional training tasks to the emotion collection and training terminal for people in different states.

[0119] After receiving the directional training task, the terminal displays the detailed requirements of the task on the display in the form of a text pop-up window;

[0120] The person being tested completes the corresponding training content according to the text instructions on the display.

[0121] In the above embodiments, it can be implemented in whole or in part by software, hardware, firmware or any combination thereof. When implemented using software, it can be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, the process or function described in the embodiment of the present application is generated in whole or in part. The computer may be a general-purpose computer, a special-purpose computer, a computer network or other programmable device. The computer instructions may be stored in a computer-readable storage medium or transmitted through the computer-readable storage medium. The computer-readable storage medium may be any available medium that a computer can access or a data storage device such as a server or a data center that includes one or more available media integrated. The available medium may be a magnetic medium, (e.g., a floppy disk, a hard disk, a tape), an optical medium (e.g., a DVD), or a semiconductor medium (e.g., a solid state disk (SSD)), etc.

[0122] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any technician familiar with the technical field can easily think of various equivalent modifications or replacements within the technical scope disclosed in the present application, and these modifications or replacements should be included in the protection scope of the present application. Therefore, the protection scope of the present application shall be based on the protection scope of the claims.

Claims

1. An AI dynamic emotion recognition multi-dimensional directional training system, characterized in that: include: Multiple emotion collection and training terminals at the front end and back-end servers; The emotion collection and training terminal is used to perform emotion detection on the person being tested, and the terminal analyzes the facial image to obtain multi-dimensional emotion data of the person being tested; The multi-dimensional emotion data is uploaded to the backend server, and the backend server analyzes the current emotional state of the measured person according to the received multi-dimensional emotion data, and pushes different directional training tasks to the corresponding emotion collection and training terminal for the measured persons in different states; After the emotion collection and training terminal receives the directional training task, the person being tested completes the corresponding training content according to the guidance.

2. According to the AI ​​dynamic emotion recognition multi-dimensional directional training system described in claim 1, it is characterized in that: The emotion collection and training terminal includes a camera, a display, an emotion data calculation module, and a control and data transmission module; The camera is used to collect the facial image of the person being tested; The display is used to display the images collected by the camera and the orientation training content; The emotion data calculation module identifies the feature points of the human face and obtains the dynamic emotion feature data of the person being measured within a fixed time; The control and data transmission module uploads the dynamic emotion feature data to the background server, and receives the directional training content and control instructions returned by the background server.

3. According to the AI ​​dynamic emotion recognition multi-dimensional directional training system described in claim 2, it is characterized in that: The emotion data calculation module identifies the feature points of the face, obtains the positions of the feature points in each frame image and the relative changes between frames, and represents the change rules of the feature points in the form of space vectors to form a face feature change matrix; The facial feature change matrix is ​​input into the emotion recognition feature model for calculation to obtain the emotion feature data of the person to be tested.

4. According to the AI ​​dynamic emotion recognition multi-dimensional directional training system described in claim 3, it is characterized in that: The coordinate system is established with the pixel point at the lower left corner of the face image as the origin. The coordinate value of the nth feature point in the face image is: (x n ,y n ); The coordinate values ​​of the feature points are standardized to obtain the relative position coordinates (x zn ,y zn ): Among them, x zn ,y zn are the normalized horizontal and vertical coordinate values; mean(x) and mean(y) are the average values ​​of the horizontal and vertical coordinates of all feature points; std(x) and std(y) are the standard deviations of the horizontal and vertical coordinates of all feature points.

5. The AI ​​dynamic emotion recognition multi-dimensional directional training system according to claim 4, characterized in that: The relative position coordinates of the 78 feature points of the current frame image are selected to form a 156-dimensional feature point space vector Ve: Ve={(x z1 ,y z1 ),(x z2 ,y z2 0,…,(x z78 ,y z78 )} For the previous frame image, the feature point space vector V is processed in the same way e '; Subtract the feature point space vectors of the previous and next frames to obtain the face feature change matrix V c : V c =V e '-V e ={(x z1 '-x z1 ,and z1 '-and z1 ),…,(x z78 '-x z78 ,and z78 '-and z78 )}; The face feature change matrix V c The emotion recognition feature model is input for calculation to obtain the emotion feature data of the person being tested.

6. The AI ​​dynamic emotion recognition multi-dimensional directional training system according to claim 5, characterized in that: The emotion feature recognition model includes a spatial feature model, a change feature model, and a regional feature model: The input of the spatial feature model is: the relative position coordinate matrix Ve of facial feature points extracted from each frame image; the output is: spatial feature emotion label The input of the variation feature model is: the variation matrix V of the facial features between each frame image and its previous frame image c ; Output: Change feature emotion label The input of the regional feature model is: the regional images of each frame image segmented according to the distribution of facial features, including the left eye image I eyel , right eye image I eyer , nose image I nos , left cheek image I facl , right cheek image I facr , mouth image I mou ; In the feature layer of the regional feature model, the output features of each segmentation image model network are spliced ​​to obtain the splicing result ConFea: ConFea=[F(I eyel ),F(I eyer ),F(I nos ),F(I facl ),F(I facr ),F(I mou )] Where F(·) is the shallow convolutional neural network of each region image; Output region feature emotion labels Among them, FC(·) is a fully connected layer network, and softmax(·) is a normalized exponential function.

7. The AI ​​dynamic emotion recognition multi-dimensional directional training system according to claim 6, characterized in that: Through weighted cumulative calculation, the emotion label of the current frame is obtained Among them, W s , W c , W a They are spatial feature weight value, change feature weight value and regional feature weight value respectively.

8. The AI ​​dynamic emotion recognition multi-dimensional directional training system according to claim 7, characterized in that: The emotion data calculation module collects multiple frames of facial images within a period of time, obtains multiple emotion tags corresponding to the multiple frames through the emotion recognition feature model, accumulates the multiple emotion tags and takes the average value to obtain the dynamic emotion feature data of the person being tested during the period of time.

9. The AI ​​dynamic emotion recognition multi-dimensional directional training system according to claim 7, characterized in that: The emotion label is a collection of 12-dimensional emotion feature parameters, expressed in the form of 12-dimensional vectors, representing: aggressiveness, pressure, tension, suspicion, balance, self-confidence, energy, self-regulation, depression, neuroticism, inhibition, and happiness.

Images