Teacher emotion recognition intervention method and system based on interaction behavior

Abstract

The application discloses a teacher emotion recognition intervention method and system based on interactive behavior, relates to the field of intelligent education technology, and comprises the following steps: step one, data preparation, including: identifying the identity of a target teacher, calling the database of a teaching management system according to the identity of the teacher, searching, and determining matched teaching data; step two, emotion analysis and intervention, including: defining the current moment as t1, the starting moment and the ending moment of a course of the target teacher as t2 and t3 respectively, and the generation time of the latest historical examination result as t4; judging whether the time difference obtained by t1-t4 is greater than a preset answer period, if yes, executing a course tracking process; if no, calling the corresponding historical examination result and analyzing the teacher emotion; if the current emotion of the teacher meets preset bad conditions, analyzing the teaching plan of the class corresponding to the historical examination result, and generating a teaching suggestion. The application has the effects of relieving bad emotions in the teaching activities of teachers and improving the teaching quality.

Description

Technical Field

[0001] This application relates to the field of smart education technology, and in particular to a teacher emotion recognition and intervention method and system based on interactive behavior. Background Technology

[0002] With the introduction of the new generation of electronic blackboards, the teaching model has been changed: In the traditional classroom, teachers often cannot keep track of all students' progress in a timely manner, resulting in some students falling behind and affecting the teaching effect; however, now, with the help of cameras and AI algorithms, the system can collect, identify, and analyze video during class, provide prompts to teachers during the teaching process, and promptly detect situations such as students' lack of concentration, so that teachers can adjust their teaching strategies and improve the teaching effect; moreover, the electronic blackboard can also be used with microphones to actively summarize the teacher's class content, forming teaching records and helping teachers simplify teaching summaries and other tasks.

[0003] However, most of the electronic blackboards and other components of smart classrooms mentioned above focus on students, with relatively little attention paid to teachers.

[0004] Today, teachers' teaching pressure has not been significantly reduced with the introduction of smart classrooms and other technologies. In addition to teaching tasks, various activities, inspections at all levels, evaluations, and writing materials make many teachers feel overworked, leading to anxiety and other emotional problems that affect teaching quality and are not conducive to long-term work. In severe cases, it has even caused some teachers to choose to change careers. Summary of the Invention

[0005] To alleviate negative emotions among teachers during teaching activities and improve teaching quality, this application provides a teacher emotion recognition and intervention method and system based on interactive behavior.

[0006] Firstly, this application provides a teacher emotion recognition and intervention method based on interactive behavior, employing the following technical solution:

[0007] A teacher emotion recognition intervention method based on interactive behavior includes:

[0008] Step 1, Data Preparation, includes:

[0009] The system identifies the target teacher and accesses the academic affairs system's database based on that teacher's identity to search for and determine matching teaching data. This teaching data includes the class taught, the class schedule, the student's historical assessment scores, and the teaching plan for any given class.

[0010] Step two, emotion analysis and intervention, includes:

[0011] Let the current time be t1, the start time and end time of a certain course of the target teacher be t2 and t3 respectively, and the generation time of the latest historical assessment score be t4;

[0012] Determine if the time difference obtained from t1 to t4 is greater than the preset Q&A period. If yes, execute the course tracking process; otherwise, retrieve the corresponding historical assessment results and analyze the teacher's emotions.

[0013] If the teacher's current mood meets the preset negative conditions, then analyze the teaching plan of the class corresponding to the historical assessment results and generate teaching suggestions;

[0014] The course tracking process includes: if the relationship between t1 and a certain course is t2 < t1 < t3, then the on-site emotion analysis method and the on-site emotion intervention method are executed;

[0015] The on-site emotion analysis method includes:

[0016] Identify the classes taught by the target teacher based on the course schedule, and request access to the monitoring data of the corresponding classes;

[0017] Teachers' emotions are analyzed based on monitoring data, and on-site emotional intervention methods are used to address the issues based on the analysis results.

[0018] Optionally, the step of retrieving corresponding historical assessment scores and analyzing teacher emotions includes:

[0019] If the current analysis is the first to the n1st time based on performance to analyze emotions, the pre-stored emotion rating table is output and sent to the target teacher, requesting them to actively conduct emotion rating, receive emotion rating feedback, and update the emotion rating table; the emotion rating table at least records the mathematical relationship between performance and emotion rating.

[0020] If the current analysis is the n2th time that emotion is analyzed based on performance, then a similarity analysis is performed based on the emotion rating table, and the record with the highest similarity is selected and the corresponding emotion rating is output; where n2 > n1, and n1 is a preset value.

[0021] Optionally, the step of analyzing the teaching plans for classes corresponding to historical assessment scores and generating teaching suggestions includes:

[0022] If the teaching plan includes teaching arrangements and lesson plans, then determine whether the next class for the class corresponding to the grade should be an assessment Q&A session based on the teaching arrangements.

[0023] If the next lesson is an assessment and Q&A session, then:

[0024] The system retrieves the assessment data from the corresponding class, identifies and summarizes the incorrect answers to form a data set.

[0025] Based on the analysis of incorrect question data, the knowledge points of the questions and their positions in the textbook are determined by searching the pre-set textbook database. Reference data for answering questions is generated as suggestions and sent to designated student representatives.

[0026] Optionally, the step of analyzing the teaching plans for classes corresponding to historical assessment scores and generating teaching suggestions further includes:

[0027] If the next session is not an assessment Q&A session, then:

[0028] The course theme will be determined according to the teaching schedule;

[0029] Based on the course theme, the pre-designed textbook database and lesson plans are searched to obtain the teaching content;

[0030] Analyze and determine whether there are knowledge points in the teaching content that correspond to the wrong question data. If so, segment the teaching content, exchange the teaching content that matches the wrong question data with the teaching content of subsequent courses, update the teaching content as a suggestion, and use the teaching content that matches the wrong question data as a suggestion when answering questions during the assessment.

[0031] Optionally, the teaching plan can be obtained by: establishing a work-accompanying AI, setting the hardware entity on which the AI ​​is mounted, and configuring accompanying logic for the hardware entity;

[0032] The accompanying logic includes:

[0033] Receive user binding instructions and bind the teacher to the hardware entity;

[0034] If the current time meets the preset audio acquisition conditions, then acquire the audio data in the environment; wherein, the audio acquisition conditions include the current time being a preset interaction period;

[0035] If new historical assessment results appear, and the current time is determined to be a non-class time according to the course schedule, then the current time is determined to meet the preset audio collection conditions, and an assessment Q&A planning inquiry and historical assessment results are sent to the user.

[0036] The system identifies and analyzes audio data to generate assessment and Q&A plans and teaching arrangements, and updates the teaching arrangements based on these plans.

[0037] Optionally, the accompanying logic further includes:

[0038] Access audio feedback data from teaching following historical assessment Q&A sessions and link it to the corresponding historical assessment scores;

[0039] Access teacher sentiment data obtained from historical performance analysis based on a sentiment rating scale;

[0040] Acoustic feature analysis is performed on the audio feedback data, and the teacher's emotions are linked to generate a set of acoustic features under the emotion, which is used to train the emotion recognition model.

[0041] If the current environment meets the preset emotion model verification conditions, the audio data is analyzed using the trained and pre-imported emotion recognition model to generate emotion analysis results, and the teacher is requested to evaluate the results. The set of acoustic features under emotion is then updated based on the evaluation.

[0042] Optionally, the step of analyzing teachers' emotions based on monitoring data and then addressing them using on-site emotional intervention methods according to the analysis results includes:

[0043] Audio data during teaching is acquired through a personal electronic device worn by the teacher.

[0044] The audio data is analyzed using an emotion recognition model to generate emotion analysis results.

[0045] If the emotional result meets the preset negative conditions, the image is extracted for facial expression recognition and behavior recognition, and an emotional analysis result is generated based on the recognition results.

[0046] The on-site emotion intervention method includes: if the wearable electronic device meets the preset physical contact wearing conditions and has a vibration function, then when the emotion analysis result meets the preset adverse conditions, a vibration trigger message is sent.

[0047] Optionally, the accompanying logic further includes:

[0048] Establish pairing relationships with other hardware entities;

[0049] Obtain the work tasks and corresponding deadlines of the bound teachers, and issue timely warnings for due dates;

[0050] Identify work tasks and assess workload.

[0051] If the current workload exceeds the preset threshold, the system queries other paired hardware entities to see if there are any identical task records, and sends an experience sharing request if identical task records are found.

[0052] Secondly, this application provides a teacher emotion recognition and intervention system based on interactive behavior, which adopts the following technical solution:

[0053] A teacher emotion recognition intervention system based on interactive behavior includes a cloud backend and a hardware unit for placement on a teacher's desk. The cloud backend is used to load and execute a computer program for a teacher emotion recognition intervention method based on interactive behavior as described above, and the hardware unit is configured with accompanying logic.

[0054] In summary, this application includes the following beneficial technical effects: On the one hand, after each student assessment, the teacher's emotions can be analyzed and evaluated through the assessment results, and the teacher can be guided to adjust the teaching plan and alleviate negative emotions during the teaching process when these emotions are negative; on the other hand, in daily classes, the monitoring of the classroom can be used to analyze and evaluate the teacher's emotions and intervene accordingly. Thus, this method can alleviate negative emotions in teachers' teaching activities and improve teaching quality. Attached Figure Description

[0055] Figure 1 This is a schematic diagram of the main process flow of the method in this application;

[0056] Figure 2 This is a schematic diagram of the architecture of this application. Detailed Implementation

[0057] The following is in conjunction with the appendix Figure 1-2 This application will be described in further detail.

[0058] This application discloses a teacher emotion recognition intervention method based on interactive behavior.

[0059] Reference Figure 1 This method, based on interactive behavior-based teacher emotion recognition and intervention, aims to utilize smart classrooms and smart campuses to collect daily interactive behaviors of teachers and students. It then uses artificial intelligence to analyze and assess teachers' daily emotions, flexibly adjusting their teaching plans, providing suggestions, and guiding them to alleviate negative emotions and improve teaching quality. This method includes:

[0060] Step 1: Data Preparation;

[0061] Step 2: Emotion Analysis and Intervention.

[0062] Step one above, data preparation, includes:

[0063] Identify the target teacher's identity and, based on that identity, access the academic affairs system's database to search for and determine matching teaching data.

[0064] In this embodiment, the method is based on smart classrooms and smart campuses. It should be able to obtain videos of relevant teaching behaviors, use them for facial recognition, and combine them with pre-established teacher profile data to obtain the identity of the target teacher.

[0065] The teaching data mentioned above includes the classes, timetables, students' historical assessment scores, and the teaching plan for any class. The teaching plan here includes: the teaching arrangements for each class and the corresponding lesson plans.

[0066] Step two above, emotion analysis and intervention, includes:

[0067] S1. Define the current time as t1, the start time and end time of a certain course of the target teacher as t2 and t3 respectively, and the generation time of the latest historical assessment results as t4.

[0068] It is understandable that a teacher can teach multiple classes a day, thus resulting in multiple sets of t2 and t3 throughout the day.

[0069] S2. Determine whether the time difference obtained from t1 to t4 is greater than the preset Q&A period. If yes, execute the course tracking process; if no, retrieve the corresponding historical assessment results and analyze the teacher's emotions.

[0070] In this embodiment, the aforementioned historical assessment results refer to exam scores. Therefore, the aforementioned Q&A period refers to the latest interval between explanations and Q&A sessions on the exam paper, and this value is set by the teacher or relevant school personnel.

[0071] It is important to note that if a class has already gone through the explanation of the test paper corresponding to a certain historical assessment score, the course tracking process will still be implemented even if the time difference is less than the preset Q&A period. That is, since the test paper has been taken out and analyzed, the impact of the test results on the teacher's emotions will not be directly analyzed again.

[0072] S31. If the teacher's current mood meets the preset negative conditions, analyze the teaching plan of the class corresponding to the historical assessment results and generate teaching suggestions.

[0073] S32. The course tracking process includes: if the relationship between t1 and a certain course is t2 < t1 < t3, then the on-site emotion analysis method and the on-site emotion intervention method are executed.

[0074] The methods for analyzing on-site emotions include:

[0075] S321. Determine the class taught by the target teacher based on the timetable, and request the corresponding class's monitoring data; the monitoring data comes from the cameras in the classroom and can be a component of the smart classroom.

[0076] S322. Analyze teachers' emotions based on monitoring data and handle them using on-site emotional intervention methods according to the analysis results.

[0077] Based on the above, this method has two main advantages: First, after each student assessment, the teacher's emotions can be analyzed and evaluated using the assessment results, and the teacher can be guided to adjust the teaching plan and alleviate negative emotions during the teaching process when these emotions are negative. Second, during daily classes, classroom surveillance can be used to analyze and evaluate the teacher's emotions and intervene accordingly. Thus, this method can alleviate negative emotions in teachers' teaching activities and improve teaching quality.

[0078] Regarding the aforementioned retrieval of historical performance evaluation scores and analysis of teacher sentiment, there are two scenarios:

[0079] The first approach involves analyzing emotions based on performance from the first to the n1st time. If this is the first time, the pre-stored emotion rating table is output and sent to the target teacher, requesting them to actively provide an emotion rating, receive feedback on the emotion rating, and update the emotion rating table.

[0080] First, the assessment results here do not refer to the scores of a single student, but rather the average score after removing the highest and lowest scores. The emotion rating table is shown in Table 1 below, with Table 1 using a maximum score of 100 and a maximum rating of 10 as an example.

[0081] Table 1: Emotion Rating Scale

[0082] score score Score Calculation mood 0-50 1-3 The score is divided into three levels, with the highest score corresponding to the highest percentage of each level being used as the final score. Difference 51-70 4-6 Same as above Poor 71-85 7-8 Same as above ordinary 85-100 9-10 Same as above excellent

[0083] Table 1 above is a pre-set table, and its matching degree varies for each teacher. Therefore, in the first n1 times, this method will actively ask teachers to report their emotional expression regarding the grades and update the emotional rating table to improve the accuracy of the rating; where n1 can be 4.

[0084] The second approach is to perform a similarity analysis based on the performance score if the current analysis is the n2th time. Then, the analysis is performed based on the emotion rating table, and the record with the highest similarity is selected and the corresponding emotion score is output.

[0085] Among them, similarity analysis is to calculate the difference between the assessment score and the score of each level, and the pair with the smallest difference has the highest similarity; n2>n1, and n1 is a preset value.

[0086] Based on the above settings, this method can quantify teachers' emotions with data, providing support for subsequent emotion interventions; at the same time, because different teachers will have different emotion rating scales, the assessment is relatively accurate.

[0087] Given an emotional score and Table 1 above, the aforementioned adverse conditions can be: the emotional score corresponds to an emotion of "poor" or "not good"; if the teacher's current emotion meets the adverse conditions, then the teaching plan for the class corresponding to the historical assessment scores is analyzed, and teaching suggestions are generated, specifically:

[0088] Based on the teaching schedule, determine whether the next class for the corresponding class will be an assessment Q&A session. If so, that is, the teacher will arrange for the next class to provide Q&A and explanation of the exam paper.

[0089] 1) Call up the corresponding class's assessment data (i.e., images of electronic and paper test papers), identify and summarize the incorrect questions data.

[0090] Among them, electronic test paper recognition includes text recognition, error symbol recognition, and score recognition; while paper test paper recognition first recognizes the image, and can only perform OCR recognition and extraction on questions with error symbol features, and then recognize the extracted content.

[0091] Error data can be linked to the student's name on the corresponding test paper, and the error data of the same student can be summarized to form an individual error set; the statistics of data generated from each assessment are determined by the teacher to be distributed to students.

[0092] 2) Based on the analysis of incorrect question data, search the pre-set textbook database to determine the knowledge points of the questions and their location in the textbook, generate answer reference data as suggestions, and send them to designated student representatives.

[0093] The methods for identifying the above knowledge points include: finding the corresponding reference answers to incorrect questions, identifying the knowledge points based on keywords in the reference answers, and then finding the location in the electronic textbook based on the knowledge points. An example of how data is sent to student agents is as follows:

[0094] Send the data to the computer in the class corresponding to the student representative. The student representative (who may be the class representative) checks the computer notifications in the class every day. After receiving the data, they can ask the teacher and, with the teacher's consent, print out the relevant data and distribute it to their classmates.

[0095] Based on the above settings, this method can help teachers organize the data on incorrect questions and knowledge points for the assessment before they prepare to explain the test paper, saving some preparation work; moreover, it can send relevant data to students in advance so that they can prepare in advance, reduce the difficulty of communication and understanding during the Q&A process, alleviate teachers' negative emotions, and improve teaching quality.

[0096] In another embodiment of this method, if it is determined from the teaching schedule that the next lesson will not be an assessment or Q&A session, then:

[0097] 1) Determine the course theme based on the teaching arrangement, such as the chapter titles in the textbook.

[0098] 2) Based on the course theme, search the pre-designed textbook database and lesson plans to obtain the teaching content.

[0099] 3) Analyze and determine whether there are knowledge points in the teaching content that correspond to the wrong question data. If so, when there is no continuous relationship between the teaching content of subsequent courses, split the teaching content, exchange the teaching content that matches the wrong question data with the teaching content of subsequent courses, update the teaching content as a suggestion, and use the teaching content that matches the wrong question data as a suggestion when answering questions during the assessment; if not, then end.

[0100] The above analysis can be done by comparing knowledge points with teaching content, and if there are any identical items, then they exist; while the determination of a continuous relationship can be done by whether the subsequent teaching content is based on the knowledge points of the previous teaching content.

[0101] Based on the above, firstly, this method can analyze whether there are any identical knowledge points in the next lesson to the incorrect questions. If so, on the one hand, the order of the lesson content can be changed to reduce the chances of teachers being affected by poor test scores; on the other hand, the teaching content can be paired with the incorrect questions, providing teachers with suggestions on how to explain the teaching content when reviewing the test papers, reducing the repetition of knowledge points while combining practical application, allowing students to better understand and improving teaching quality.

[0102] Reference Figure 2 In one embodiment of this method, the teaching plan is obtained by: establishing a work-accompanying AI, setting the hardware body on which the AI ​​is mounted, and configuring the accompanying logic on the hardware body.

[0103] The aforementioned hardware can be a computer or laptop, but more preferably, a standalone smart ornament. The ornament's casing can be cartoon-shaped, etc., and its interior contains an integrated circuit board with data analysis and processing capabilities. This integrated circuit board includes a Wi-Fi module, an audio module, and a battery to power the circuit board, which also has a charging circuit. The ornament is placed on the teacher's desk and used as a "companion spirit."

[0104] The accompanying logic mentioned above includes:

[0105] 1) Receive user binding instructions and bind the teacher to the hardware entity.

[0106] Understandably, for data security reasons, it is necessary to bind the aforementioned "Accompanying Genie" to the corresponding teacher. The binding method can be to register an account through the corresponding App and enter the teacher's identity. After binding, the audio collected by the "Accompanying Genie" is stored locally and can only be copied, deleted, or modified by the teacher with the corresponding permissions.

[0107] 2) If the current time meets the preset audio acquisition conditions, then acquire the audio data in the environment.

[0108] The audio acquisition conditions include the current time being a preset interactive time period, such as: the current time is a weekday and the time is 9:00-17:00; it can be understood that the audio acquisition conditions are not only met in terms of time, but can also be other environmental conditions, user behavior conditions, etc.; for example, if the teacher feels that the current time is relatively inappropriate, they can actively turn it off; that is, in this method, the "accompaniment sprite" is not constantly acquiring audio.

[0109] 3) If new historical assessment results appear, and the current time is determined to be a non-class time according to the timetable, then the current time is determined to meet the preset audio collection conditions, and an assessment Q&A planning inquiry and historical assessment results are sent to the user.

[0110] The inquiry can be made via corresponding audio; if the hardware has an integrated display screen, the inquiry and assessment results can be displayed on the screen.

[0111] 4) Identify and analyze audio data to obtain assessment and Q&A plans and teaching arrangements, and update the teaching arrangements based on the assessment and Q&A plans.

[0112] The assessment and Q&A plan includes, for example, when to conduct the assessment and Q&A sessions and how many lessons to complete them.

[0113] The teaching arrangements are designed with in mind that if this method required teachers to manually input information, it would increase their workload. Therefore, this method allows for the input of teaching arrangements via voice, in a daily, interactive manner. It should be noted that if teaching arrangements have already been prepared and plans for assessment and Q&A sessions are proposed at this time, the teaching arrangements need to be updated.

[0114] In another embodiment of this method, the accompanying logic further includes:

[0115] Access audio feedback data from teaching following historical assessment Q&A sessions and link it to the corresponding historical assessment scores;

[0116] Access teacher sentiment data obtained from historical performance analysis based on a sentiment rating scale;

[0117] Acoustic feature analysis is performed on the audio feedback data, and the teacher's emotions are linked to generate an acoustic feature set under the emotion, which is used to train the emotion recognition model. It is understood that audio recognition and speech emotion recognition are existing technologies, so they will not be elaborated on further.

[0118] If the current environment meets the preset emotion model verification conditions, such as the verification time selected by the teacher and the model analysis of the teacher's previous dialogue being excellent, then the audio data is analyzed using the trained and pre-imported emotion recognition model to generate emotion analysis results, and the teacher is asked to evaluate the results. Based on the evaluation, the acoustic feature set under emotion is updated.

[0119] Based on the above settings, the hardware entity (or companion sprite) in this method can utilize the feedback audio obtained from inquiries after each assessment result is released, and use it in conjunction with the aforementioned emotion rating standard table to automatically generate sample data for training the model; at the same time, it also cross-validates the accuracy of the emotion rating table.

[0120] In this method, teachers' emotions are analyzed based on monitoring data, and the analysis results are processed using on-site emotional intervention methods. This can involve extracting monitoring data, identifying teachers, recognizing teachers' facial expressions, and recognizing teachers' behaviors. Based on the facial expression and behavior recognition results, the teacher's emotions are determined. When the emotions meet preset negative conditions, such as showing an angry expression or exhibiting bad behavior, on-site emotional intervention methods are used to address the issue.

[0121] One method for on-site emotional intervention is to display a pop-up notification on the computer in the classroom. It is important to note that the pop-up notification should not be displayed when the classroom projector is turned on.

[0122] Based on the above, in another embodiment, this method involves: analyzing teachers' emotions based on monitoring data, and processing these emotions using on-site emotional intervention methods according to the analysis results, including:

[0123] Audio data during teaching is acquired through a personal electronic device worn by the teacher. That is, the teacher needs to wear a personal electronic device, such as a smartwatch, smart bracelet, or smart ring, and use the electronic device to keep the audio data collected in the classroom and send it to the aforementioned "accompanying spirit" via the network.

[0124] The audio data is analyzed using an emotion recognition model to generate emotion analysis results.

[0125] If the emotional outcome meets the preset negative conditions, the image is extracted for facial expression recognition and behavior recognition, and an emotion analysis result is generated based on the recognition results. For example, an angry expression indicates a bad mood; throwing books or slamming the table indicates a poor mood.

[0126] The on-site emotion intervention method includes: if the wearable electronic device meets the preset physical contact wearing conditions and has a vibration function, then when the emotion analysis result meets the preset adverse conditions, a vibration trigger message is sent.

[0127] Based on the above settings, this application can provide more accurate and flexible, subtle prompts when teachers are in a bad mood; at the same time, it reduces the amount of monitoring data processed.

[0128] The accompanying logic mentioned above also includes:

[0129] Establish pairing relationships with other hardware entities; for example, teacher A and teacher B establish a friendship relationship, and the devices corresponding to each other grant access permissions for distributing certain data (such as task names and lists as described below).

[0130] (By recognizing audio data) the system obtains the teacher's assigned tasks (e.g., activities, evaluations) and corresponding deadlines, and issues timely warnings for due dates. Teachers can set the advance warning time themselves.

[0131] Identify work tasks and assess workload, such as task quantity statistics. More accurately, developers can pre-set the workload values ​​represented by various tasks, and then accumulate the workload values ​​corresponding to all tasks.

[0132] If the current workload exceeds the preset threshold, the system queries other paired hardware devices to see if there are any records of the same tasks (the device access history list of users). If there are records of the same tasks, the system sends an experience sharing request. The request can be displayed in the form of a pop-up window for the corresponding teacher to view the response.

[0133] The above settings can automatically analyze the historical records of other teachers and seek help from other experienced teachers when a teacher experiences excessive work pressure (i.e., heavy workload). This is especially important for new teachers and newly hired teachers, as it can accelerate their integration into the job and significantly reduce their work pressure.

[0134] This application also discloses a teacher emotion recognition intervention system based on interactive behavior.

[0135] Reference Figure 2 A teacher emotion recognition intervention system based on interactive behavior includes a cloud backend and a hardware unit for placement on a teacher's desk. The cloud backend is used to load and execute a computer program for a teacher emotion recognition intervention method based on interactive behavior as described above. The hardware unit is configured with accompanying logic.

[0136] The above are all preferred embodiments of this application, and are not intended to limit the scope of protection of this application. Therefore, all equivalent changes made in accordance with the structure, shape and principle of this application should be covered within the scope of protection of this application.

Claims

1. A teacher emotion recognition intervention method based on interaction behavior, characterized in that, Comprise: Identify the identity of the target teacher, and call the database of the educational administration system according to the identity of the teacher to determine the matching teaching data; wherein, the teaching data includes the class, the timetable, the historical assessment results of the students and the teaching plan of any class; Define the current time as t1, the starting time and the ending time of a course of the target teacher as t2 and t3 respectively, and the generation time of the latest historical assessment result as t4; Determine whether the time difference obtained by t1-t4 is greater than the preset answer period, if yes, execute the course tracking process; if no, call the corresponding historical assessment results and analyze the teacher's emotion, if the current emotion of the teacher meets the preset bad condition, analyze the teaching plan of the class corresponding to the historical assessment results, and generate teaching suggestions, which include: If the teaching plan includes teaching arrangement and teaching plan, determine whether the next course of the class corresponding to the results is an examination and answer according to the teaching arrangement; If the next course is an examination and answer, then: Call the examination data of the corresponding class, identify, and summarize to form the wrong question data, wherein the examination data includes electronic test paper and image of paper test paper; Based on the analysis of the wrong question data, find the knowledge points and the location in the preset textbook database to generate answer reference data as suggestions, and send to the designated student representative; If the next course is not an examination and answer, then: Determine the course theme according to the teaching arrangement; Find the teaching content from the preset textbook database and teaching plan according to the course theme; Analyze and determine whether the teaching content contains the knowledge points corresponding to the wrong question data, if yes, segment the teaching content, exchange the teaching content matched with the wrong question data with the teaching content of the subsequent course, update the teaching content as suggestions, and use the teaching content matched with the wrong question data as suggestions for examination and answer; The teaching plan can also be obtained by: establishing a work companion AI, setting up the hardware main body carried by the AI, configuring the companion logic for the hardware main body, and the companion logic comprises: Receive user binding instructions and bind the teacher for the hardware main body; If the current time meets the preset audio collection condition, obtain the audio data in the environment, wherein the audio collection condition includes that the current time is a preset interaction companion period; If a new historical assessment result appears, and according to the timetable, it is determined that the current time is not a class time, it is determined that the current time meets the preset audio collection condition, and the user is initiated with an examination and answer planning inquiry and historical assessment results; Identify and analyze the audio data to obtain the examination and answer planning and teaching arrangement, and update the teaching arrangement based on the examination and answer planning; The course tracking process comprises: If the relationship between t1 and a course is t2 < t1 < t3, execute the on-site emotion analysis method, which comprises: Determine the class of the target teacher according to the timetable, and request to call the monitoring data of the corresponding class; Analyze the emotion of the teacher based on the monitoring data, and handle it according to the analysis result by the on-site emotion intervention method. 2.The teacher emotion recognition intervention method based on interaction behavior according to claim 1, characterized in that: The calling of the corresponding historical assessment results and the analysis of the teacher's emotion comprise: If the current is the first to the n1th time of analyzing the emotion based on the score, a pre-stored emotion score table is output, and a request for active emotion scoring is sent to the target teacher, emotion score feedback is received, and the emotion score table is updated; wherein the emotion score table records at least the mathematical relationship between the score and the emotion score; If the current is the n2th time of analyzing the emotion based on the score, similarity analysis is performed based on the emotion score table, the record with the highest similarity is taken, and the corresponding emotion score is output; wherein n2>n1, and n1 is a preset value. 3.The teacher emotion recognition intervention method based on interaction behavior according to claim 2, characterized in that, The accompanying logic further comprises: Audio feedback data of the teaching after the inquiry of the historical examination answer question planning is called, and the corresponding historical examination score is bound; The teacher emotion obtained by analyzing the historical score based on the emotion score table is called; Acoustic feature analysis is performed on the audio feedback data, and the teacher emotion is bound to generate an acoustic feature set under the emotion, which is used for training the emotion recognition model; If the current environment meets the preset emotion model verification condition, the audio data is analyzed by the trained and pre-imported emotion recognition model to generate an emotion analysis result, and the teacher is requested to evaluate, and the acoustic feature set under the emotion is updated based on the evaluation. 4.The teacher emotion recognition intervention method based on interaction behavior according to claim 3, characterized in that, The emotion of the teacher is analyzed based on the monitoring data, and the analysis result is processed by a live emotion intervention method, which comprises: Audio data during teaching is obtained through a wearable electronic device pre-worn by the teacher; The audio data is analyzed by the emotion recognition model to generate an emotion analysis result; If the emotion result meets the preset adverse condition, facial expression recognition and behavior recognition are performed on the image, and an emotion analysis result is generated according to the recognition result; If the wearable electronic device meets the preset body contact wearing condition and has a vibration function, a vibration trigger information is sent when the emotion analysis result meets the preset adverse condition. 5.The teacher emotion recognition intervention method based on interaction behavior according to claim 4, characterized in that, The accompanying logic further comprises: A pairing relationship with other hardware subjects is established; The work tasks of the bound teacher and the corresponding deadlines are obtained, and an expiration warning is given by timing; The work tasks are identified, and the task amount is evaluated; If the current task amount exceeds the preset threshold, the other paired hardware subjects are inquired whether there are the same task records, and an experience sharing request is sent when there are the same task records.

6. An interactive behavior based teacher emotion recognition intervention system, characterized by: A cloud backend and a hardware subject for placing on the teacher's desk are included, the cloud backend is used to load and execute the computer program of the teacher emotion recognition intervention method based on the interactive behavior as claimed in any one of claims 1-5, and the hardware subject is configured with the accompanying logic.

Images