A teaching level evaluation method and system based on local accounts

Abstract

The present invention relates to the field of intelligent evaluation technology, and specifically discloses a teaching level evaluation method and system based on a local account. The method includes obtaining the usage process and facial data of each student based on the granted permissions, determining the concentration according to the usage process, and synchronously training the facial data to a neural network model of the concentration; regularly collecting the facial data of each student, and determining the concentration of each student based on the trained neural network model; receiving the teaching plan input by the teacher, identifying the teaching plan, determining the importance of different time periods, and determining the evaluation score according to the importance of different moments and the concentration of each student at different moments. The present invention determines the importance of each stage in the teaching process based on the information of the teacher, and determines the concentration of the students in each stage based on the information of the students. This process is carried out in real time, and the evaluation score of the teaching process is jointly determined by combining the concentration and the importance, with high globality and strong objectivity.

Description

Technical Field

[0001] The present invention relates to the technical field of intelligent evaluation, and specifically to a teaching level evaluation method and system based on local accounts. Background Art

[0002] With the progress of technology and the development of society, people have more and more desires and ways to acquire knowledge, and the forms and quantities of classes are also increasing. The intelligent classroom is a teaching interaction method based on intelligent devices, including online and offline methods.

[0003] There is an important behavior in the teaching interaction process, that is, the teaching level evaluation behavior. The existing evaluation methods are still based on manual evaluation. However, the memory of students is limited, and the data sources for evaluation often come only from some teaching segments in the final stage. In fact, on the basis of the sufficient popularization of existing intelligent devices, it is not complicated to introduce a global and objective teaching level evaluation scheme. How to build a global and objective teaching level evaluation framework is the technical problem that the technical solution of the present invention wants to solve. Summary of the Invention

[0004] The purpose of the present invention is to provide a teaching level evaluation method and system based on local accounts to solve the problems raised in the above background art.

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

[0006] A teaching level evaluation method based on local accounts, the method includes:

[0007] Sending a permission acquisition request to the user and receiving the permission granted by the user; wherein, the user includes students and teachers;

[0008] Obtaining the usage process and facial data of each student based on the granted permission, determining the concentration according to the usage process, and synchronously training the facial data to the neural network model of the concentration; wherein, one student corresponds to one neural network model;

[0009] Regularly collecting the facial data of each student and determining the concentration of each student based on the trained neural network model;

[0010] Receiving the teaching plan input by the teacher, identifying the teaching plan, determining the importance of different time periods, and determining the evaluation score according to the importance of different times and the concentration of each student at different times.

[0011] As a further solution of the present invention: the step of obtaining the usage process and facial data of each student based on the granted permission, determining the concentration according to the usage process, and synchronously training the facial data to the neural network model of the concentration includes:

[0012] Obtain the browsing content and operation parameters of the user at each moment in real time based on the granted permissions, and determine the concentration at each moment according to the browsing content and operation parameters;

[0013] Obtain the trainee image based on the granted permissions, identify the trainee image, and extract facial data;

[0014] Perform time-domain registration on the facial data and concentration, and use the time-domain registered facial data and concentration as samples to train a neural network model; the input of the neural network model is facial data, and the output is concentration.

[0015] As a further solution of the present invention: the step of obtaining the browsing content and operation parameters of the user at each moment in real time based on the granted permissions and determining the concentration at each moment according to the browsing content and operation parameters includes:

[0016] Obtain the browsing content and operation parameters of the user at each moment in real time based on the granted permissions;

[0017] Read the serial number corresponding to the operation parameter at each moment in the preset parameter conversion table; the parameter conversion table contains an operation type item and a serial number item;

[0018] Aggregate the moments according to the browsing content to obtain the time period corresponding to each browsing content;

[0019] Statistically arrange the serial numbers corresponding to the operation parameters within the time period based on the time sequence to obtain a group of actual operation serial numbers;

[0020] Read the preset standard serial number group of each browsing content, compare the group of actual operation serial numbers with the standard serial number group, calculate the similarity, and determine the concentration according to the similarity; wherein, the concentration is directly proportional to the similarity.

[0021] As a further solution of the present invention: the step of receiving the teaching plan input by the teacher, identifying the teaching plan, determining the importance of different time periods, and determining the evaluation score according to the importance of different moments and the concentration of each trainee at different moments includes:

[0022] Receive the teaching plan input by the teacher, identify the teaching plan, and determine the importance of different time periods; identifying the teaching plan includes a keyword matching process, different keywords correspond to different importance levels, and both the keywords and their importance levels are preset parameters;

[0023] Read the concentration of each trainee at different moments and insert them into the same time axis;

[0024] Intercept matching points on the time axis according to a preset time step, and read the number of people with a concentration less than a preset concentration threshold at the matching points; the matching points are time points;

[0025] Query the importance corresponding to the matching points, and determine the evaluation score according to the importance and the number of people.

[0026] As a further solution of the present invention: the step of receiving the teaching plan input by the teacher, identifying the teaching plan, and determining the importance of different time periods includes:

[0027] Receive the teaching keywords and their importance of each time period input by the teacher.

[0028] Based on the granted permissions, obtain the audio information of the teacher, locate the teaching keywords, and correct the teaching keywords and their importance of each time period according to the positioning results.

[0029] As a further solution of the present invention: the step of querying the importance corresponding to the matching points and determining the evaluation score according to the importance and the number of people includes:

[0030] Query the number threshold of people for each importance, calculate the ratio of the number of people to the number threshold, and determine the score of each matching point according to the ratio; where the score is inversely proportional to the ratio.

[0031] Determine the weight according to the importance of each matching point, and accumulate all scores based on the weight to obtain the final evaluation score; where the weight is directly proportional to the importance.

[0032] The technical solution of the present invention also provides a teaching level evaluation system based on a local account, and the system includes:

[0033] A permission acquisition module, configured to send a permission acquisition request to the user and receive the permission granted by the user; where the user includes students and teachers;

[0034] A model training module, configured to obtain the usage process and facial data of each student based on the granted permissions, determine the concentration according to the usage process, and synchronously train the neural network model of the facial data to the concentration; where one student corresponds to one neural network model;

[0035] A concentration recognition module, configured to regularly collect the facial data of each student and determine the concentration of each student based on the trained neural network model.

[0036] An evaluation score determination module, configured to receive the teaching plan input by the teacher, identify the teaching plan, determine the importance of different time periods, and determine the evaluation score according to the importance of different moments and the concentration of each student at different moments.

[0037] As a further solution of the present invention: The model training module includes:

[0038] The initial concentration determination unit is used to obtain the browsing content and operation parameters of the user at each moment in real time based on the granted permissions, and determine the concentration at each moment according to the browsing content and operation parameters;

[0039] The facial data extraction unit is used to obtain the student image based on the granted permissions, identify the student image, and extract the facial data;

[0040] The training execution unit is used to perform time-domain registration on the facial data and concentration, and use the time-domain registered facial data and concentration as samples to train the neural network model; the input of the neural network model is the facial data, and the output is the concentration.

[0041] As a further solution of the present invention: The initial concentration determination unit includes:

[0042] The data acquisition subunit is used to obtain the browsing content and operation parameters of the user at each moment in real time based on the granted permissions;

[0043] The serial number query subunit is used to read the serial number corresponding to the operation parameter at each moment in the preset parameter conversion table; the parameter conversion table contains an operation type item and a serial number item;

[0044] The time aggregation subunit is used to aggregate the moments according to the browsing content to obtain the time period corresponding to each browsing content;

[0045] The serial number arrangement subunit is used to statistically arrange the serial numbers corresponding to the operation parameters within the time period based on the time sequence to obtain the practical operation serial number group;

[0046] The comparison subunit is used to read the preset standard serial number group of each browsing content, compare the practical operation serial number group with the standard serial number group, calculate the similarity, and determine the concentration according to the similarity; wherein, the concentration is proportional to the similarity.

[0047] As a further solution of the present invention: The evaluation score determination module includes:

[0048] The solution recognition unit is used to receive the teaching solution input by the teacher, identify the teaching solution, and determine the importance of different time periods; identifying the teaching solution includes a keyword matching process, different keywords correspond to different importance levels, and both the keywords and their importance levels are preset parameters;

[0049] The concentration statistics unit is used to read the concentration of each student at different moments and insert them into the same time axis;

[0050] A number query unit is used to intercept matching points on the time axis at a preset time step and read the number of people whose concentration at the matching points is less than a preset concentration threshold; the matching points are time points.

[0051] A data application unit is used to query the importance corresponding to the matching points and determine an evaluation score according to the importance and the number of people.

[0052] Compared with the prior art, the beneficial effects of the present invention are as follows: The present invention determines the importance degree of each stage in the teaching process based on the information of the teacher, and determines the concentration of the students in each stage based on the information of the students. This process is carried out in real time, and the evaluation score of the teaching process is jointly determined by combining the concentration and the importance degree, with high globality and strong objectivity. Brief Description of the Drawings

[0053] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following will briefly introduce the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings in the following description are only some embodiments of the present invention.

[0054] Figure 1 It is a flowchart of a teaching level evaluation method based on a local account.

[0055] Figure 2 It is a composition structure diagram of a teaching level evaluation system based on a local account. Detailed Embodiment

[0056] In order to make the technical problems, technical solutions and beneficial effects to be solved by the present invention more clearly understood, the following further details the present invention in conjunction with the drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0057] Figure 1 It is a flowchart of a teaching level evaluation method based on a local account. In an embodiment of the present invention, a teaching level evaluation method based on a local account, the method includes:

[0058] Step S100: Send a permission acquisition request to the user and receive the permission granted by the user; wherein, the user includes students and teachers.

[0059] The application scenario of this application is very clear, which is a smart classroom equipped with electronic devices. Each student is equipped with a tablet computer, and the instructor also has his own computer device. In the technical solution of the present invention, the permission acquisition process is a necessary process, including obtaining the permissions granted by students and the permissions granted by the instructor. The obtained permissions include the camera capture permission and the operation information recording permission. The camera capture permission is used to obtain the video of the student or the instructor, and the operation information recording permission is used to obtain the operation behaviors of the student or the instructor on the device, including touch screen signals and keyboard and mouse signals. In the application scenario of this application, the audio signal for operation, that is, the audio control method, is relatively rare.

[0060] Step S200: Obtain the usage process and facial data of each student based on the granted permissions, determine the concentration based on the usage process, and synchronously train the facial data to the neural network model of the concentration; where one student corresponds to one neural network model;

[0061] After receiving the permissions granted by the user, obtain the usage process and facial data of each student based on the granted permissions. The usage process indicates what operations the student has performed on his tablet computer. By analyzing the usage process, the concentration degree of the student during the teaching process can be calculated, which is called the concentration. At the same time, this application also obtains facial data. There is a mapping relationship between the facial data and the concentration. Based on the existing facial data and concentration, using the facial data as features (inputs) and the concentration as labels (outputs), a sample set can be constructed. Based on the sample set, a neural network model is trained. After the training is completed, inputting the facial data can obtain the concentration.

[0062] It should be noted that in the technical solution of the present invention, the training process of each student is independent. One student corresponds to one neural network model, and the type of facial data of one student is very limited. The training process is very simple, and a simple regression model can be used.

[0063] Step S300: Regularly collect the facial data of each student, and determine the concentration of each student based on the trained neural network model;

[0064] In practical applications, during a teaching activity, regularly collect the facial data of each student, and input the facial data into the trained neural network model to obtain the concentration of each student. Regarding the concept of timing in this application, this application takes the start of the teaching activity as the zero moment. For example, if a teaching activity has 72 class hours, then the very beginning is the zero moment. Facial data is collected once every minute. The result obtained from the first collection is the result corresponding to one minute, and the result obtained from the second collection is the result corresponding to two minutes, and so on, to obtain multiple results containing time.

[0065] It is worth mentioning that in order to improve the evaluation accuracy, a shorter fixed time interval can be adopted. At this time, the detection granularity is larger and more data is collected. For example, it changes from one minute to thirty seconds, and the amount of data obtained is almost twice the original amount of data.

[0066] Step S400: Receive the teaching plan input by the instructor, identify the teaching plan, determine the importance of different time periods, and determine the evaluation score according to the importance of different times and the concentration of each student at different times;

[0067] The instructor can input the teaching plan for the entire teaching process at the beginning of teaching (suitable for recorded teaching activities), or input the teaching plan for the current class before each class (suitable for real-time teaching activities), identify the teaching plan (mainly perform text recognition on the teaching plan), and determine the importance of each moment in the entire teaching activity according to the recognition result, which is represented by the parameter of importance; in step S300, the concentration of each student at each moment has been calculated, and the teaching process is evaluated by combining the importance of different times and the concentration of each student at different times to obtain the evaluation score; the evaluation score represents how many students are concentrated on the content of different importance levels in the teaching process and the degree of concentration. Its practical significance is how much important content is attracted by students with what degree of concentration, which is a very suitable parameter for evaluating the teaching effect.

[0068] Regarding step S100, the steps of obtaining the usage process and facial data of each student based on the granted permission, determining the concentration according to the usage process, and synchronously training the neural network model of facial data to concentration include:

[0069] Obtain the browsing content and operation parameters of the user at each moment in real time based on the granted permission, and determine the concentration at each moment according to the browsing content and operation parameters;

[0070] Obtain the student image based on the granted permission, identify the student image, and extract the facial data;

[0071] Perform time-domain registration on the facial data and the concentration, and use the time-domain registered facial data and concentration as samples to train the neural network model; the input of the neural network model is the facial data, and the output is the concentration.

[0072] In an example of the technical solution of the present invention, based on the granted permissions, the browsing content and operation parameters of the trainee at each moment are obtained in real time. The browsing content represents what content the trainee is viewing on the tablet computer, and the operation parameters represent what operations are performed. The browsing content and operation parameters at each moment are analyzed to determine the trainee's degree of concentration in the class, which is represented by the parameter of concentration. At the same time, based on the granted permissions, the trainee's image is obtained, and the trainee's image is recognized to extract facial data. This process belongs to the existing facial recognition process and will not be elaborated here. Whether it is the degree of concentration or the facial data, both contain time information. Only the degree of concentration and facial data at the same moment (the time difference is small enough) can be used as the same sample. The facial data and the degree of concentration are registered in the time domain. Based on the facial data and the degree of concentration after time-domain registration as samples, a neural network model is trained. Among them, the process of time-domain registration is to group the degree of concentration and facial data at the same moment.

[0073] Specifically, the steps of obtaining the browsing content and operation parameters of the user at each moment in real time based on the granted permissions and determining the degree of concentration at each moment according to the browsing content and operation parameters include:

[0074] Obtain the browsing content and operation parameters of the user at each moment in real time based on the granted permissions;

[0075] Read the serial number corresponding to the operation parameter at each moment in the preset parameter conversion table; the parameter conversion table contains an operation type item and a serial number item;

[0076] Aggregate the moments according to the browsing content to obtain the time period corresponding to each browsing content;

[0077] Statistically arrange the serial numbers corresponding to the operation parameters within the time period based on the time sequence to obtain a group of actual operation serial numbers;

[0078] Read the preset standard serial number group of each browsing content, compare the group of actual operation serial numbers with the standard serial number group, calculate the similarity, and determine the degree of concentration according to the similarity; wherein, the degree of concentration is proportional to the similarity.

[0079] In an example of the technical solution of the present invention, a specific solution for calculating the degree of concentration is provided. Based on the granted permissions, the browsing content and operation parameters of the user at each moment are obtained in real time. The browsing content is generally the courseware pre-sent by the instructor, which belongs to known data. Regarding the operation parameters, there are only several existing operation parameters, including zooming, sliding, page turning, and marking, etc. They can all be pre-statistically recorded in the parameter conversion table and numbered, so that each operation is quantified as a serial number, and one operation corresponds to a unique serial number. That is, the parameter conversion table contains an operation type item and a serial number item.

[0080] For the registration process of browsing content and operation parameters, the browsing content generally corresponds to a time period. The top-level content is obtained on the tablet based on the granted permissions, which is the browsing content. Since the moment corresponding to each browsing content is continuous, it is necessary to aggregate the moments according to the browsing content to obtain the time period corresponding to each browsing content.

[0081] Furthermore, regarding the operation information, the time period of each browsing content is determined. The operation information within the time period is the operation process of the trainee regarding the browsing content. Based on the time sequence, the serial numbers corresponding to the operation parameters within the time period are statistically arranged to obtain the practical operation serial number group, which reflects the quantified operation process of the trainee during this period; for each browsing content, since it is known data, its standard operation can be determined in advance. For reading text, it is almost an intermittent sliding and page-turning operation. Its standard operation is pre-statistically determined (the input of the instructor can be received and used as a limitation), and it is represented in the form of a standard serial number group. At this time, the operation information corresponds to the practical operation serial number group, and the standard operation corresponds to the standard serial number group. By comparing the practical operation serial number group and the standard serial number group, the similarity is calculated, and the concentration can be determined based on the similarity.

[0082] Regarding the process of comparing the practical operation serial number group and the standard serial number group and calculating the similarity, the simplest way is to calculate the maximum intersection of the two, and determine the similarity based on the length of the maximum intersection. The longer the length of the maximum intersection, the higher the similarity.

[0083] Regarding step S400, the steps of receiving the teaching plan input by the instructor, identifying the teaching plan, determining the importance of different time periods, and determining the evaluation score based on the importance of different moments and the concentration of each trainee at different moments include:

[0084] Receiving the teaching plan input by the instructor, identifying the teaching plan, and determining the importance of different time periods; identifying the teaching plan includes a keyword matching process, where different keywords correspond to different importance levels, and both the keywords and their importance levels are preset parameters;

[0085] Reading the concentration of each trainee at different moments and inserting them into the same time axis;

[0086] Intercepting matching points on the time axis at a preset time step, and reading the number of people whose concentration at the matching points is less than the preset concentration threshold; the matching points are moment points;

[0087] Querying the importance corresponding to the matching points, and determining the evaluation score based on the importance and the number of people.

[0088] In an example of the technical solution of the present invention, the evaluation process is specifically defined. The teaching plan input by the teacher is received, and the teaching plan is identified. The identification of the teaching plan includes a keyword matching process. Different keywords correspond to different importance levels. The keywords and their importance levels are all preset parameters. The teaching content of different time periods is intercepted from the teaching plan, the keywords in the teaching content are matched, the importance levels of the matched keywords are read, and then the importance levels of the matched keywords are accumulated to obtain the final importance level. It should be noted that when intercepting the teaching content of different time periods from the teaching plan, the different time periods can be preset time periods, such as two minutes or three minutes, etc. The determination process of the importance level is carried out in segments.

[0089] The concentration levels of each student at different times are read and inserted into the same time axis. The concentration level is the concentration level at each moment. For example, the concentration level is collected every thirty seconds. All students have a concentration level. For the convenience of analysis and processing, it is inserted into the same time axis. Further, there may be a slight difference in the concentration level recognition moments of each student. For example, there is a difference of a few tenths of a second caused by the data transmission process. Therefore, when analyzing the concentration level, first intercept the matching points on the time axis. The intercepted matching points are to intercept one every thirty seconds (the same as the guiding period of the concentration level analysis process, that is, the period of the image acquisition and recognition process). Query the concentration levels of each student at the matching points (moments), and compare them with the preset concentration level threshold. When the concentration level is less than the preset concentration level threshold, it means that the student is lacking attention at the current matching point. Obtain the number of students lacking attention at the matching point (the concentration level is small enough), and the number of students who are distracted at the corresponding moment can be obtained.

[0090] Finally, query the time period where the matching point is located, read the corresponding importance level, and combine the importance level and the number of people to determine the evaluation score.

[0091] Further, the steps of receiving the teaching plan input by the teacher, identifying the teaching plan, and determining the importance levels of different time periods include:

[0092] Receive the teaching keywords and their importance levels of each time period input by the teacher;

[0093] Based on the granted permissions, obtain the audio information of the teacher, locate the teaching keywords, and correct the teaching keywords and their importance levels of each time period according to the positioning results.

[0094] The above provides a new process for determining a teaching plan. The teaching plan itself is the information uploaded by the trainees, and the instructor knows it best. Therefore, it is possible to receive the importance levels input by the instructor for each time period. At the same time, the teaching keywords for each time period are also received. The purpose of receiving the teaching keywords is to verify the importance levels. For example, if the corresponding teaching keyword does not appear in a certain time period, that time period is split into two segments and incorporated into the two adjacent time periods respectively, that is, the time period is corrected.

[0095] In addition, for the importance level corresponding to the query matching point, the steps for determining the evaluation score according to the importance level and the number of people include:

[0096] Query the number threshold for each importance level, calculate the ratio of the number of people to the number threshold, and determine the score for each matching point according to the ratio; among them, the score is inversely proportional to the ratio;

[0097] Determine the weight according to the importance level of each matching point, and accumulate all the scores based on the weight to obtain the final evaluation score; among them, the weight is directly proportional to the importance level.

[0098] Specifically, the calculation process of the evaluation score is as follows: ; In the formula, represents the evaluation score, represents the total number of matching points, represents the th ratio corresponding to the matching point, represents the th score corresponding to the matching point.

[0099] Among them, the small detail of the above calculation process is that the influence of the importance level is greater, so it uses an exponential base.

[0100] The above specifically describes the evaluation process. The evaluation score is comprehensively determined by the number of people and the importance level. For each importance level, a number threshold is first determined in advance. The number threshold indicates the maximum number of trainees in the state of lack of attention for this importance level; the number of people is the number of people lacking attention at the current matching point. The number threshold is generally a relatively large value. Calculate the ratio of the number of people to the number threshold. The ratio is generally less than 1. The smaller the ratio, the fewer people lacking attention, and a higher score is given to it; on this basis, a weight is determined by the importance level. The weight can also be understood as the unit score. The higher the importance level, the greater the unit score; finally, accumulate all the scores based on the weight to obtain the final evaluation score; the higher the score, the higher the teaching level.

[0101] Figure 2 FIG. is a structural composition diagram of a teaching level evaluation system based on a local account. In an embodiment of the present invention, a teaching level evaluation system based on a local account, the system 10 includes:

[0102] A permission acquisition module 11 for sending a permission acquisition request to a user and receiving the permissions granted by the user; wherein, the users include trainees and instructors.

[0103] A model training module 12 for obtaining the usage process and facial data of each trainee based on the granted permissions, determining the concentration based on the usage process, and synchronously training the facial data to a neural network model of the concentration; wherein, one trainee corresponds to one neural network model.

[0104] A concentration recognition module 13 for regularly collecting the facial data of each trainee and determining the concentration of each trainee based on the trained neural network model.

[0105] An evaluation score determination module 14 for receiving the teaching plan input by the instructor, identifying the teaching plan, determining the importance of different time periods, and determining the evaluation score according to the importance of different times and the concentration of each trainee at different times.

[0106] Further, the model training module 12 includes:

[0107] A preliminary concentration determination unit for obtaining the browsing content and operation parameters of the user at each moment in real time based on the granted permissions, and determining the concentration at each moment according to the browsing content and operation parameters.

[0108] A facial data extraction unit for obtaining the trainee image based on the granted permissions, identifying the trainee image, and extracting the facial data.

[0109] A training execution unit for performing time domain registration on the facial data and the concentration, and training the neural network model based on the time domain registered facial data and concentration as samples; the input of the neural network model is the facial data, and the output is the concentration.

[0110] Specifically, the preliminary concentration determination unit includes:

[0111] A data acquisition subunit for obtaining the browsing content and operation parameters of the user at each moment in real time based on the granted permissions.

[0112] A serial number query subunit for reading the serial number corresponding to the operation parameter at each moment in a preset parameter conversion table; the parameter conversion table contains an operation type item and a serial number item.

[0113] A moment aggregation subunit for aggregating the moments according to the browsing content to obtain the time period corresponding to each browsing content.

[0114] A serial number arrangement subunit for statistically arranging the serial numbers corresponding to the operation parameters within the time period based on the time sequence to obtain an actual operation serial number group.

[0115] A comparison subunit, configured to read a preset standard serial number group of each piece of browsing content, compare the actual operation serial number group with the standard serial number group, calculate a similarity, and determine a concentration degree according to the similarity; wherein, the concentration degree is directly proportional to the similarity.

[0116] Furthermore, the evaluation score determination module 14 includes:

[0117] A scheme recognition unit, configured to receive a teaching scheme input by a teacher, recognize the teaching scheme, and determine the importance degree of different time periods; recognizing the teaching scheme includes a keyword matching process, different keywords corresponding to different importance degrees, and both the keywords and their importance degrees are preset parameters;

[0118] A concentration degree statistics unit, configured to read the concentration degree of each student at different moments and insert them into the same time axis;

[0119] A number query unit, configured to intercept matching points on the time axis according to a preset time step length, and read the number of people whose concentration degree at the matching points is less than a preset concentration degree threshold; the matching points are time points;

[0120] A data application unit, configured to query the importance degree corresponding to the matching points, and determine an evaluation score according to the importance degree and the number of people.

[0121] The above are only the preferred embodiments of the present invention, and are not intended to limit the present invention. Any modifications, equivalent replacements, and improvements made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A teaching level evaluation method based on a local account, characterized in that: The method comprises: Sending a permission acquisition request to a user, and receiving the permission granted by the user; wherein the user includes a student and a teacher; Based on the granted permissions, the usage progress and facial data of each student are obtained, the concentration is determined according to the usage progress, and a neural network model of facial data to concentration is trained synchronously; wherein one student corresponds to one neural network model; Collect facial data of each student regularly and determine the concentration of each student based on the trained neural network model; Receive the teaching plan input by the teacher, identify the teaching plan, determine the importance of different time periods, and determine the evaluation score according to the importance at different times and the concentration of each student at different times; The steps of obtaining the usage process and facial data of each student based on the granted authority, determining the concentration according to the usage process, and synchronously training the facial data to the neural network model of the concentration include: Based on the granted authority, the browsing content and operation parameters of the user at each moment are obtained in real time, and the concentration at each moment is determined according to the browsing content and operation parameters; Obtain the student image based on the granted permissions, identify the student image, and extract facial data; Performing temporal registration on the facial data and the degree of concentration, and training a neural network model based on the facial data and the degree of concentration after the temporal registration as samples; the input of the neural network model is the facial data, and the output is the degree of concentration; The step of obtaining the browsing content and operation parameters of the user at each moment in real time based on the granted authority, and determining the concentration at each moment according to the browsing content and operation parameters comprises: Based on the granted permissions, the user's browsing content and operation parameters at each moment are obtained in real time; Reading the serial number corresponding to the operation parameter at each moment in a preset parameter conversion table; the parameter conversion table contains an operation type item and a serial number item; Aggregate the time according to the browsing content to obtain the time period corresponding to each browsing content; Based on the time sequence, the serial numbers corresponding to the operation parameters in the time period are counted and arranged to obtain the actual operation serial number group; The preset standard serial number group of each browsing content is read, the practical serial number group and the standard serial number group are compared, the similarity is calculated, and the concentration is determined according to the similarity; wherein the concentration is proportional to the similarity.

Images