CN116629647A

CN116629647A - Classroom advanced assessment method and device, storage medium and electronic equipment

Info

Publication number: CN116629647A
Application number: CN202210131227.4A
Authority: CN
Inventors: 李波
Original assignee: Guangzhou Shiyuan Electronics Thecnology Co Ltd; Guangzhou Shirui Electronics Co Ltd
Current assignee: Guangzhou Shiyuan Electronics Thecnology Co Ltd; Guangzhou Shirui Electronics Co Ltd
Priority date: 2022-02-11
Filing date: 2022-02-11
Publication date: 2023-08-22

Abstract

The application discloses a classroom advanced assessment method, a device, a storage medium and electronic equipment, wherein the method comprises the following steps: acquiring a classroom teaching plan text of a target classroom and a recorded broadcast video of the target classroom; acquiring the occurrence times of each word belonging to the target entity category in the classroom teaching plan text, and generating a first probability vector based on each occurrence time; acquiring the number of devices belonging to different device categories in the recorded broadcast video, and generating a second probability vector based on each number; and determining an advanced result of the target class based on the first probability vector and the second probability vector. By adopting the embodiment of the application, the evaluation of the target class advanced result is realized.

Description

Classroom advanced assessment method and device, storage medium and electronic equipment

Technical Field

The application relates to the technical field of computers, in particular to a classroom advance assessment method, a classroom advance assessment device, a storage medium and electronic equipment.

Background

Education is a century old of a country, and is a national culture talents, and the education of a country is strong and comprehensive. The teaching means is a tool, a medium or a device for transmitting information with students in the teaching process, and the teaching means goes through five stages of oral speech, characters and books, printed teaching materials, electronic audio-visual equipment and multimedia network technology along with the development of scientific technology, and whether digital terminal equipment is used in the teaching process or not reflects the advancement of teaching design to a certain extent along with the continuous optimization of the teaching means.

Disclosure of Invention

The embodiment of the application provides a classroom advanced evaluation method, a classroom advanced evaluation device, a storage medium and electronic equipment, which realize evaluation of target classroom advanced results.

In a first aspect, an embodiment of the present application provides a classroom advance assessment method, where the method includes:

acquiring a classroom teaching plan text of a target classroom and a recorded broadcast video of the target classroom;

acquiring the occurrence times of each word belonging to the target entity category in the classroom teaching plan text, and generating a first probability vector based on each occurrence time;

acquiring the number of devices belonging to different device categories in the recorded broadcast video, and generating a second probability vector based on each number;

and determining an advanced result of the target class based on the first probability vector and the second probability vector.

In a second aspect, an embodiment of the present application provides a class advance assessment apparatus, including:

the data acquisition module is used for acquiring a classroom teaching plan text of a target classroom and a recorded broadcast video of the target classroom;

the first vector generation module is used for obtaining the occurrence times of each word belonging to the target entity category in the classroom teaching plan text and generating a first probability vector based on each occurrence time;

The second vector generation module is used for acquiring the number of the devices belonging to different device categories in the recorded broadcast video and generating a second probability vector based on each number;

and the advanced evaluation module is used for determining an advanced result of the target class based on the first probability vector and the second probability vector.

In a third aspect, embodiments of the present application provide a computer storage medium storing a plurality of instructions adapted to be loaded by a processor and to perform the method steps of the first aspect described above.

In a fourth aspect, an embodiment of the present application provides an electronic device, which may include: a processor and a memory; wherein the memory stores a computer program adapted to be loaded by the processor and to perform the method steps of the first aspect described above.

The technical scheme provided by the embodiments of the application has the beneficial effects that at least:

in the embodiment of the application, the number of times of each word belonging to the target entity class in the text of the classroom is obtained by obtaining the text of the classroom teaching plan of the target classroom and the recorded broadcast video of the target classroom, a first probability vector is generated based on each number of times, the number of devices belonging to different device classes in the recorded broadcast video is obtained, a second probability vector is generated based on each number, and the advanced result of the target classroom is determined based on the first probability vector and the second probability vector. The method comprises the steps of generating a first probability vector by counting the occurrence times of each word belonging to a target entity class in a class teaching plan text of a target class, generating a second probability vector by counting the number of devices of different device classes in recorded broadcast video of the target class, and realizing evaluation of an advanced result of the target class by combining the first probability vector and the second probability vector.

Drawings

In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings that are required to be used in the embodiments will be briefly described below, and it is apparent that the drawings in the following description are only some embodiments of the present application, and other drawings may be obtained according to these drawings without inventive effort for a person skilled in the art.

FIG. 1 is a system architecture diagram of a class advance assessment system according to an embodiment of the present application;

FIG. 2 is a schematic diagram of a process for class advance assessment according to an embodiment of the present application;

FIG. 3 is a schematic diagram illustrating an example of generating a statistical probability vector according to an embodiment of the present application;

FIG. 4 is a schematic diagram illustrating an example of calculating an advanced score according to an embodiment of the present application;

FIG. 5 is a schematic flow chart of a class advance assessment method according to an embodiment of the present application;

FIG. 6 is an exemplary diagram of obtaining probability vectors corresponding to words according to an embodiment of the present application;

FIG. 7 is an exemplary schematic diagram of labeling devices in a frame of a picture according to an embodiment of the present application;

FIG. 8 is a schematic diagram illustrating an example of overlapping of the same positions of devices in a frame of a picture according to an embodiment of the present application;

FIG. 9 is a schematic diagram illustrating an example of overlapping of the same positions of devices in a frame of a picture according to an embodiment of the present application;

FIG. 10 is a schematic diagram illustrating an example of acquiring different device classes according to an embodiment of the present application;

FIG. 11 is a schematic diagram illustrating an example of calculating an advanced score according to an embodiment of the present application;

fig. 12 is a schematic structural diagram of a class advance assessment device according to an embodiment of the present application;

fig. 13 is a schematic structural diagram of a class advance assessment device according to an embodiment of the present application;

fig. 14 is a schematic structural diagram of an electronic device according to an embodiment of the present application.

Detailed Description

The technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present application.

In the following description, the terms "first," "second," and "first," are used for descriptive purposes only and are not to be construed as indicating or implying relative importance. The following description provides various embodiments of the application that may be substituted or combined between different embodiments, and thus the application is also to be considered as embracing all possible combinations of the same and/or different embodiments described. Thus, if one embodiment includes feature A, B, C and another embodiment includes feature B, D, then the present application should also be considered to include embodiments that include one or more of all other possible combinations including A, B, C, D, although such an embodiment may not be explicitly recited in the following.

The following description provides examples and does not limit the scope, applicability, or examples set forth in the claims. Changes may be made in the function and arrangement of elements described without departing from the scope of the application. Various examples may omit, replace, or add various procedures or components as appropriate. For example, the described methods may be performed in a different order than described, and various steps may be added, omitted, or combined. Furthermore, features described with respect to some examples may be combined into other examples.

Education is a century old of a country, and is a national culture talents, and the education of a country is strong and comprehensive. In the information age of rapid development, education has entered digitization, networking, intellectualization and multimedia, and has achieved openness, sharing, interaction and collaboration. In the age of education informatization, the modern information technology is comprehensively and deeply applied to education management, education teaching and education scientific research to promote the process of education reform and development, and the teaching means of a teacher can directly or indirectly reflect the teaching value. The teaching means is a tool, a medium or a device for the teacher to communicate information with the students in the teaching process, and whether the digital terminal device is used in the teaching process can reflect the advancement of teaching design to a certain extent. At present, school administrators have difficulty in monitoring teaching levels of teachers in tens or hundreds of classes daily in a school, and evaluating the advanced level of the class.

Based on the above, the embodiment of the application provides a class advance evaluation method, which is characterized in that through acquiring a class teaching plan text of a target class and a recorded broadcast video of the target class, the occurrence times of words belonging to a target entity class in the class teaching plan text are acquired, a first probability vector is generated based on the occurrence times, the number of devices belonging to different device classes in the recorded broadcast video is acquired, a second probability vector is generated based on the number, and an advance result of the target class is determined based on the first probability vector and the second probability vector. The method comprises the steps of generating a first probability vector by counting the occurrence times of each word belonging to a target entity class in a class teaching plan text of a target class, generating a second probability vector by counting the number of devices of different device classes in recorded broadcast video of the target class, and realizing evaluation of an advanced result of the target class by combining the first probability vector and the second probability vector.

Referring to fig. 1, a system architecture diagram of a classroom advance assessment method is provided in an embodiment of the present application.

The system architecture includes terminal devices, which may include, but are not limited to, smart interactive tablets, smart phones, personal computers, desktop computers, tablet computers, palmtops, laptops, computer-in-one, vehicle-mounted multimedia, and the like.

The method comprises the steps of inputting a classroom teaching plan text into a language processing model through obtaining the classroom teaching plan text of a target classroom and a recorded broadcast video of the target classroom, outputting probability vectors corresponding to words in the classroom teaching plan text, determining entity categories corresponding to maximum probability values in the probability vectors, determining the times at least comprising equipment names and function names in the entity categories corresponding to the maximum probability, and generating a first probability vector based on the times of occurrence of each target entity; inputting recorded broadcast video into an image equipment detection model, outputting probability vectors of equipment belonging to different equipment categories in the recorded broadcast video, determining the maximum probability in the probability vectors, determining the number corresponding to the equipment categories based on the equipment categories corresponding to the maximum probability, generating falling probability vectors based on the number corresponding to the equipment categories, splicing the first probability vectors and the second probability vectors to obtain statistical probability vectors, and inputting the statistical probability vectors into an inference model to obtain the advanced score. If the advanced score is greater than or equal to the score threshold, determining that the advanced of the target class reaches the standard; and if the advanced score is smaller than the score threshold, determining that the target class advanced does not reach the standard.

The method comprises the steps of inputting a classroom teaching plan text of a target classroom into a language processing model to obtain the occurrence times of words belonging to each target device in the classroom teaching plan text, obtaining a first probability vector, obtaining recorded broadcast video of the target classroom, inputting the recorded broadcast video into an image device detection model to obtain the number of devices belonging to different devices in the recorded broadcast video, obtaining a second probability vector, splicing the first probability vector and the second probability vector to obtain a statistical probability vector, inputting the statistical probability vector into an inference model to obtain an advanced score of the target classroom, and comparing the advanced score with a scoring threshold value to realize evaluation of advanced results of the target classroom.

A large number of classroom teaching plan texts are formulated before the classroom teaching plan texts of the target class are input into the language processing modelThe language processing model is trained. Wherein omega is a model parameter obtained by training a language processing model, phi (x, y) is the sum of probabilities that all words of each sentence in a classroom teaching plan text belong to a target entity class, and h _i [y _i ]Probability vectors for each word belonging to different entity classes in each sentence +.>The probability vector is a probability vector that adjacent words of each word in the text of the classroom teaching plan belong to different entity categories. Inputting a text of a classroom teaching plan into a language processing model, predicting the probability sum of all words belonging to each target entity category of each sentence in the text of the classroom teaching plan, and comparing the predicted probability sum with the formula +. >Comparing the calculated actual values, and updating the model parameter omega of the model so as to accurately identify the model when the model is usedThe number of occurrences of each term belonging to each target entity category in the classroom teaching plan text of the target classroom to generate a first probability vector.

Before inputting recorded video of a target class into an image equipment detection model, training the image equipment detection model by using a plurality of recorded video input into the image equipment detection model through a loss function, wherein the loss function is L (p, u, t ^u ,v)＝L _cls (p,u)+λ[u≥1]L _loc (t ^u ,v)，L _cls (p,u)＝-log[pu+(1-p)(1-u)]，L _loc (t ^u ,v)＝∑smooth _L1 (t ^u -v)，p is the probability that each device in the recorded video belongs to the target device class, u is the probability that each device in the recorded video does not belong to the target device class, t ^u In order to label the predicted coordinates and the vectors corresponding to the predicted width and height of the rectangular frames of each device, v is the actual coordinates and the vectors corresponding to the actual width and height of the rectangular frames of each device, lambda is the weight for classifying and predicting the rectangular frames of each device, x is the difference value between the actual coordinates and the predicted coordinates, L _cls (p,u)＝-log[pu+(1-p)(1-u)]To classify the rectangular boxes labeling each device, loss value, L _loc (t ^u ,v)＝∑smooth _L1 (t ^u -v) is a predicted coordinate loss value for a rectangular box labeling each device, smooth _L1 Is a smooth loss function. The recorded video is input into an image equipment detection model, equipment belonging to different equipment categories in the recorded video is detected, the accuracy of identifying the equipment categories corresponding to the equipment in the recorded video by the model is improved through the calculation of loss values, and therefore the number of the equipment corresponding to the equipment categories in the recorded video is accurately identified when the recorded video in a target class is input into the image equipment detection model, and a second probability vector is generated.

Based on the system architecture shown in fig. 1, the classroom advanced assessment method provided by the embodiment of the present application will be described in detail below with reference to fig. 2 to 11.

Referring to fig. 2, a schematic flow chart of a classroom advance assessment method is provided in an embodiment of the present application. As shown in fig. 2, the classroom advance assessment method may include the steps of:

s101, acquiring a classroom teaching plan text of a target classroom and recording and broadcasting video of the target classroom.

In one embodiment, when the advanced performance of the target classroom is required to be evaluated, inputting the classroom teaching plan text and recorded video of the target classroom into the terminal equipment, and extracting the classroom teaching plan text of the target classroom from the classroom teaching plan text database according to the classroom identification through natural language processing (Natural Language Processing, NLP) related technology; and extracting the recorded video of the target class from the recorded video database according to the class identification.

The text of the classroom teaching plan and the recorded video input to the terminal equipment are required to be the text of the classroom teaching plan and the recorded video corresponding to the same class.

The class identifier may include, but is not limited to, a time of lesson, a class of lesson, a teacher of lesson, a name of teacher of lesson, and the like.

The text of the classroom teaching plan may include, but is not limited to, teaching design process, topics, teaching purposes, emphasis and difficulty, scheduling of lessons, selection of tutorials and teaching media, board design, post-lesson recall, and the like.

NLP is an artificial intelligence for professional analysis of human language, and refers to a computing technology for automatic analysis and representation of human language, which is driven by a series of theories. The working principle of the device is as follows: receiving natural language, the final language is evolved through natural use by humans, we communicate with it every day; translating natural language, typically by a probability-based algorithm; analyzing the natural language and outputting the result. NLP enables computers to perform a number of natural language related tasks such as sentence structure parsing, part-of-speech tagging, machine translation, and dialogue systems. The scheme uses NLP to classify text of classroom teaching plan according to content.

S102, obtaining the occurrence times of each word belonging to the target entity category in the classroom teaching plan text, and generating a first probability vector based on each occurrence time.

In one embodiment, a user inputs a classroom identifier of a target classroom to be scored in a terminal device, the terminal device extracts a classroom teaching plan text of the target classroom from a classroom teaching plan text according to the classroom identifier, the step S101 can know that the classroom teaching plan text of the target classroom is an object class classroom teaching plan text extracted from a classroom teaching plan text database according to the classroom identifier by using an NLP related technology, then a word segmentation algorithm is used for segmenting the classroom teaching plan text to obtain each word corresponding to the classroom teaching plan text, the classroom teaching plan text is input in a language processing model, the language processing model outputs probability vectors of each word belonging to each preset entity class, a maximum probability value corresponding to each word is determined from each probability vector, the entity class corresponding to the maximum probability value is determined as the entity class of the word, the number of occurrences of each word belonging to each target entity class in the classroom teaching plan text is determined according to the entity class of each word, and a first probability vector is generated according to the number of occurrences of each word belonging to each target entity class.

The preset entity categories may include, but are not limited to, device names, function names, person names, place names, organization names, and others.

The target entity categories include, but are not limited to, device names, function names, and the like.

The device name may include, but is not limited to, a smart terminal, IPad, learning tablet, answering machine, and the like.

The function name may include, but is not limited to, on-screen transmission, electronic classroom activity, and the like.

The language processing model may include, but is not limited to, a bi-directional recurrent neural network + conditional random field algorithm, a deep recurrent neural network, a recurrent neural network, and the like.

Each dimension of the first probability vector is the number of occurrences of words belonging to each target entity category in the text of the classroom teaching plan. As shown in table 1, the number of times that the words of each target entity category in a certain classroom teaching plan text respectively correspond to each other appears in the classroom teaching plan text.

TABLE 1

Words of target entity class	Number of occurrences in classroom teaching plan text
		Class a words	8 times
Class B words	3 times

As can be seen from Table 1, class A words appear 8 times in the text of the classroom teaching plan, class B words appear 3 times in the text of the classroom teaching plan, and thus the corresponding first probability vector is Or->

Methods of word segmentation of classroom teaching text may include, but are not limited to, dictionary-based methods, statistical-based methods, rule-based methods, word-labeling-based chinese word segmentation methods, and the like.

The dictionary-based method (character string matching, mechanical word segmentation method) is to match the character string to be analyzed with the entry in a large machine dictionary according to a certain strategy, and if a certain character string is found in the dictionary, the matching is successful. The method can be divided into forward matching and reverse matching according to different scanning directions; the difference in length can be divided into maximum matching and minimum matching.

The more times adjacent words occur simultaneously, the more likely a word is composed in a context based on statistical word segmentation (no dictionary word segmentation). Therefore, the probability or frequency of the adjacent occurrence of the characters can better reflect the credibility of the words. The main statistical models include an N-gram model and a hidden Markov model (Hidden Markov Model, HMM).

The rule-based method (based on semantics) achieves the effect of recognizing words by simulating the understanding of human sentences, the basic idea is semantic analysis and syntactic analysis, the text is segmented by utilizing syntactic information and semantic information, and automatic reasoning is performed.

The Chinese word segmentation method based on the word labeling is a word-forming method in practice. Namely, the word segmentation process is regarded as a labeling problem of the words in the word strings. Since each word occupies a certain word forming position (i.e., word position) when constructing a particular word, if it is specified that there are at most four word forming positions per word: namely B (word head), M (word middle), E (word tail) and S (word alone). The word is not limited to Chinese characters. Considering that the real text of Chinese inevitably contains a certain number of non-kanji characters, the "word" referred to herein also includes foreign letters, arabic numerals, punctuation marks and the like. All of these characters are basic units of word formation. Of course, chinese characters remain the most numerous characters in this set of units.

And S103, acquiring the number of devices belonging to different device categories in the recorded broadcast video, and generating a second probability vector based on each number.

In one embodiment, a user inputs a class identifier corresponding to a target class to be subjected to advanced scoring into a terminal device, the terminal device extracts recorded broadcast video of the target class from a recorded broadcast video database according to the class identifier, the recorded broadcast video is input into an image device detection model, the image device detection model divides the recorded broadcast video into at least one picture frame according to a preset period, the number of devices belonging to different device categories in each picture frame is identified, and a second probability vector is generated based on the number.

The device categories may include, but are not limited to, smart interactive tablets, smart phones, answering machines, smart watches, personal computers, projectors, desktop computers, tablet computers, palm top computers, laptop computers, computer all-in-one computers, and the like.

Image device identification models may include, but are not limited to, fast regional convolutional neural networks (Fast R-CNN and Fast R-CNN), exhaustive searches, regional convolutional neural networks (Region-CNN), and the like.

The method for intercepting the recorded and broadcast video picture frames can include, but is not limited to, intercepting the recorded and broadcast video picture frames at the same time interval, or splitting the video into a plurality of videos with equal duration, and then capturing the picture frames from the plurality of videos, or splitting the video into a plurality of videos with equal size, and then capturing the picture frames from the plurality of videos.

And intercepting the picture frames at the same time interval, for example, the duration of one recorded video is 2 minutes, and if the time interval is 1 second, 120 picture frames corresponding to the recorded video can be obtained. If the recorded video of a target class is cut off according to a time interval of 1 second, 40 x 60 or 45 x 60 image frames corresponding to the recorded video can be obtained.

After the recorded video is intercepted into a plurality of picture frames, the number of the devices corresponding to each device category in all the picture frames is identified, and a second probability vector is generated according to the number of the devices corresponding to each device category, as shown in table 2, which is the number of the devices corresponding to each device category in the recorded video.

TABLE 2

Device class	Number of devices corresponding to each device class
		Class a device	4 pieces of
Class B device	6 pieces of

As can be seen from table 2, there are 4 class a devices in the recorded video and 6 class B devices in the recorded video, so that the corresponding second probability vector isOr->

And S104, determining an advanced result of the target class based on the first probability vector and the second probability vector.

In one embodiment, the first probability vector and the second probability vector generated in the step S102 and the step S103 are spliced and input into an inference model, the inference model scores the advanced performance of the target class according to the first probability vector and the second probability vector, and the advanced performance result of the target class is evaluated according to the advanced performance score output by the inference model.

The stitching of the first probability vector and the second probability vector includes, but is not limited to, that the first probability vector is placed in front of the second probability vector, and the values of the first probability vector and the second probability vector are randomly fused to form a new probability vector.

As shown in fig. 3, the first probability vector generated in the step S102 is calculatedAnd the second probability vector generated in step S103 +.>And splicing to obtain a statistical probability vector.

The inference model may include, but is not limited to, a deep neural network.

For example, as shown in fig. 4, a statistical probability vector formed by splicing a first probability vector and a second probability vector is input into a deep neural network, each neuron corresponds to a probability value corresponding to each dimension of the statistical probability vector, an advanced score of a target class is obtained by calculating parameters in the deep neural network, and the advanced of the target class is evaluated by the advanced score of the target class.

Optionally, if the advanced result does not reach the standard, determining the substandard equipment in the target class, and sending prompt information for replacing the equipment to the substandard equipment.

It will be appreciated that the advanced result may be represented by an advanced score, when the advanced score fails to reach the standard, indicating that there may be at least one device that fails to reach the standard in the target class, determining the devices that fail to reach the standard in the target class, and sending corresponding prompt information to the devices for device replacement, so that the target class meets the advanced requirement.

In the embodiment of the application, the number of times of occurrence of each word belonging to each target device in the text of the classroom is obtained by obtaining the text of the classroom teaching plan of the target classroom and the recorded broadcast video of the target classroom, a first probability vector is generated based on each number of times, the number of devices belonging to different device categories in the recorded broadcast video is obtained, a second probability vector is generated based on each number, and the advanced result of the target classroom is determined based on the first probability vector and the second probability vector. The user inputs the classroom teaching plan text of the target classroom to be evaluated and the recorded broadcast video of the target classroom into the terminal equipment, and generates a first probability vector and a second probability vector respectively corresponding to the number of occurrences of each word belonging to each target equipment in the classroom teaching plan text and the number of equipment belonging to different equipment categories in the recorded broadcast video, and the evaluation of the target classroom advanced result is realized by combining the first probability vector and the second probability vector.

Referring to fig. 5, a flow chart of a classroom advance assessment method is provided in an embodiment of the present application. As shown in fig. 5, the classroom advance assessment method may include the steps of:

S201, training a language processing model through the probability sum of all words belonging to target entity categories of all sentences in the text of the classroom teaching plan, the probability vector of each word belonging to different entity categories in all the sentences and the probability vector of adjacent words belonging to different entity categories in each word in the text of the classroom teaching plan, and determining the trained model parameters of the language processing model.

In one embodiment, prior to model use, the classroom teaching plan text is input to a language processing model, and the number of occurrences of each word for each target entity category in the classroom teaching plan text is extracted.

By the formulaTraining a language processing model to adjust parameters omega in the model, wherein omega is a model parameter obtained by training the language processing model, phi (x, y) is a probability sum of all words of each sentence in a classroom teaching plan text belonging to a target entity class, and h _i [y _i ]Probability vectors for each word belonging to different entity classes in each sentence +.>The probability vector is a probability vector that adjacent words of each word in the text of the classroom teaching plan belong to different entity categories. Inputting the text of the classroom teaching plan into a language processing model, predicting the probability sum of all words belonging to each target entity class of each sentence in the text of the classroom teaching plan, and comparing the predicted probability sum phi (x, y) with the passing formula The calculated actual probability sum phi (x, y) is compared, and the model parameter omega of the model is updated, so that the occurrence times of words belonging to the target entity categories in the classroom teaching plan text of the target classroom can be accurately identified when the model is used, and a first probability vector is generated.

S202, training an image equipment detection model through a loss function, wherein the loss function is obtained based on the probability that each equipment in the recorded broadcast video belongs to a target equipment category, the probability that each equipment in the recorded broadcast video does not belong to the target equipment category, a predicted coordinate of a rectangular frame of each equipment and a vector corresponding to the predicted width and height, an actual coordinate of the rectangular frame of each equipment and a vector corresponding to the actual width and height, a preset weight and a difference value between the actual coordinate and the predicted coordinate.

In one embodiment, while training the language processing model, the recorded video of the target class is input to the image device detection model, and the number of devices belonging to different device categories in the recorded video is extracted.

And updating parameters of the model through calculation of the loss function so as to improve detection of the model on equipment categories of all equipment in the recorded broadcast video.

Before inputting recorded video of a target class into an image equipment detection model, training the image equipment detection model by using a plurality of recorded video input into the image equipment detection model through a loss function, wherein the loss function is L (p, u, t ^u ,v)＝L _cls (p,u)+λ[u≥1]L _loc (t ^u ,v)，L _cls (p,u)＝-log[pu+(1-p)(1-u)]，L _loc (t ^u ,v)＝∑smooth _L1 (t ^u -v)，p is the probability that each device in the recorded video belongs to the target device class, u is the probability that each device in the recorded video does not belong to the target device class, t ^u In order to label the predicted coordinates and the vectors corresponding to the predicted width and height of the rectangular frames of each device, v is the actual coordinates and the vectors corresponding to the actual width and height of the rectangular frames of each device, lambda is the weight for classifying and predicting the rectangular frames of each device, x is the difference value between the actual coordinates and the predicted coordinates, L _cls (p,u)＝-log[pu+(1-p)(1-u)]To classify the rectangular boxes labeling each device, loss value, L _loc (t ^u ,v)＝∑smooth _L1 (t ^u -v) is a prediction of the rectangular box labeling each deviceCoordinate loss value, smooth _L1 Is a smooth loss function. The recorded video is input into an image equipment detection model, equipment belonging to different equipment categories in the recorded video is detected, the accuracy of identifying the equipment categories corresponding to the equipment in the recorded video by the model is improved through the calculation of loss values, and therefore the number of the equipment corresponding to the equipment categories in the recorded video is accurately identified when the recorded video in a target class is input into the image equipment detection model, and a second probability vector is generated.

Where lossiloc is a position loss, and optimization of the position loss is performed only in the non-background (u=0 is background, u≡1 represents a non-background category). The functional form of the position loss is smoothL1 loss, i.e., the square term when the argument is less than 1, and L1 loss when it is greater than 1. The L1 loss is adopted because the prediction of regression has no range limitation, and the L1 function can well inhibit the influence of outliers.

S203, acquiring a classroom teaching plan text of a target classroom and recording and broadcasting video of the target classroom.

In this step, reference may be made to step S101, which is not described herein.

S204, inputting the text of the classroom teaching plan into a language processing model, and outputting probability vectors corresponding to each word in the text of the classroom teaching plan.

In one embodiment, a user inputs a classroom teaching plan text corresponding to a target classroom which wants to be subjected to advanced evaluation into a terminal device, the classroom teaching plan text is processed through an NLP (non-linear projection) related technology, the classroom teaching plan text of the target classroom is extracted from a classroom teaching plan text database according to a classroom identifier, word segmentation processing is performed on the classroom teaching plan text, words corresponding to the classroom teaching plan text are obtained, the classroom teaching plan text is input into a trained language processing model, and the language processing model outputs probability vectors of the words belonging to preset entity categories.

If the language processing model is a bidirectional cyclic neural network, each preset entity category is a device name, a function name, and others, as shown in fig. 6, each word obtained by word segmentation processing of the text of the classroom teaching plan is input into the bidirectional cyclic neural network, and the probability vector corresponding to each preset entity category is output. The corresponding words of the text of the classroom teaching plan are respectively a probability vector corresponding to each word which is output after each word is input into the bidirectional cyclic neural network, the probability vector corresponding to the word is [0.09, 005,0.86], "the probability vector corresponding to the answer machine" is [0.95,0.03,0.02], and the probability vector corresponding to the answer is [0.1,0.13,0.7].

The method of word segmentation of the text of the classroom teaching plan may be, but is not limited to, dictionary-based methods, statistical-based methods, rule-based methods, word-labeling-based Chinese word segmentation methods, and the like.

The Chinese word segmentation method based on the word labeling is a word formation method in practice. Namely, the word segmentation process is regarded as a labeling problem of the words in the word strings. Since each word occupies a certain word forming position (i.e., word position) when constructing a particular word, if it is specified that there are at most four word forming positions per word: namely B (word head), M (word middle), E (word tail) and S (word alone). The word is not limited to Chinese characters. Considering that the real text of Chinese inevitably contains a certain number of non-kanji characters, the "word" referred to herein also includes foreign letters, arabic numerals, punctuation marks and the like. All of these characters are basic units of word formation. Of course, chinese characters remain the most numerous characters in this set of units.

S205, determining the maximum probability vector in the probability vectors, determining the times of each target entity class in entity classes corresponding to the maximum probability, and generating a first probability vector based on the times, wherein the target entity class at least comprises a device name and a function name.

In one embodiment, a maximum probability value is determined in probability vectors corresponding to each word, an entity class corresponding to the maximum probability value is determined as an entity class to which the corresponding word belongs, the number of times of occurrence of the word belonging to each target class is determined from the entity class of each word, and a first probability vector is generated according to the number of times of occurrence of the word of each target class.

As shown in fig. 5, each word is input to a bi-directional cyclic neural network (language processing model), and probability vectors belonging to each preset entity class corresponding to each word are output. In step S203, for example, each preset entity category is a device name, a function name, and other words in the text of the classroom teaching plan, the probability vectors of each word corresponding to each preset entity category are [0.09, 005,0.86], [0.95,0.03,0.02], and [0.1,0.13,0.7], respectively, so that the entity category corresponding to the maximum probability value of "use" is "other", the entity category corresponding to the maximum probability value of "answer" is 0.95 is "device name", the entity category corresponding to the maximum probability value of "answer" is 0.7 is "other", and the word corresponding to "other" appears 2 times, and the word corresponding to "device name" appears 1 time.

According to the above example, the entity class of each word in the text of the classroom teaching plan is determined according to the probability vector output by each word after the word passes through the language processing model, the entity class of each word determines the maximum probability value from the corresponding probability vectors, and the entity class corresponding to the maximum probability value is the entity class corresponding to the word.

If the preset entity category has the equipment name, the function name, the person name, the place name and other entity categories, the target entity category has the equipment name and the function name, the equipment name appears 4 times in the text of the classroom teaching plan, the function name appears 6 times in the text of the classroom teaching plan, the person name appears 5 times in the text of the classroom teaching plan, the place name appears 1 time in the text of the classroom teaching plan, the other name appears 9 times in the text of the classroom teaching plan, and the generated first probability vector comprises the times of the appearance of the equipment name and the function name in the text of the classroom teaching plan, which is thatOr->

S206, inputting the recorded video into an image equipment detection model, and outputting probability vectors of different equipment categories of all equipment in the recorded video.

In one embodiment, a user inputs a class identification frequency of a target class to be subjected to advanced evaluation into a terminal device, the terminal device extracts recorded broadcast video of the target class from a recorded broadcast video database according to the class identification, and a trained image device detection model detects devices of each device class in the recorded broadcast video to obtain probability vectors of each device belonging to each device class.

The class identifier may include, but is not limited to, a class time, a time of a lesson, a class of a lesson, a teacher of a lesson, a name of a teacher of a lesson, and the like.

The device categories may include, but are not limited to, smart interactive tablets, smart phones, answering machines, smart watches, personal computers, projectors, desktop computers, tablet computers, palm top computers, laptop computers, computer all-in-one machines, and the like.

As shown in fig. 7, by identifying all picture frames of the recorded broadcast video, determining the device category existing in the recorded broadcast video, labeling each device to obtain rectangular frames corresponding to each device, obtaining a 4-dimensional vector corresponding to each rectangular frame by obtaining the coordinates of the same position of each rectangular frame and the width and height of each rectangular frame, if the coordinates of the same position of one rectangular frame and other rectangular frames are (3, 4), the width is 6, and the height is 8, the vectors corresponding to the rectangular frames are [3,4,6,8], and comparing the vectors corresponding to each rectangular frame to determine whether the rectangular frames overlap.

Dividing recorded video into at least one picture frame according to a preset period, identifying the number of devices belonging to different device categories in each picture frame by an image device detection model, marking the devices in each picture frame to obtain rectangular frames corresponding to the devices, determining the positions of the rectangular frames and the widths and heights corresponding to the rectangular frames respectively by acquiring the coordinates of the same positions of the rectangular frames, judging whether the two rectangular frames are in the same device category according to the widths and heights of the two rectangular frames when the coordinates of the same positions of the two rectangular frames are repeated, judging the device category corresponding to the two rectangular frames as the same device category when the widths and heights of the two rectangular frames are the same, reserving one rectangular frame, determining the probability of the rectangular frame belonging to the device category according to the widths and heights of the rectangular frame, and determining the device category corresponding to the two rectangular frames as the different device category when one of the widths or the heights of the two rectangular frames is different, and determining the probability of the rectangular frames belonging to the device category according to the widths and the heights of the two rectangular frames when the two rectangular frames are different.

As shown in fig. 8, two rectangular frames are repeated at the same position in the figure, it is determined that the two rectangular frames respectively correspond to different device categories by identifying the width and the height of the two rectangular frames, and then the probability that the two rectangular frames respectively belong to each device category is determined according to the values of the width and the height of the two rectangular frames.

As shown in fig. 9, two rectangular frames are repeated at the same position in the figure, the two rectangular frames are judged to be in the same equipment category by identifying the width and the height of the two rectangular frames, only one rectangular frame is reserved, and then the probability that the rectangular frame belongs to each equipment category is determined according to the value of the width and the height of the rectangular frame.

S207, determining the maximum probability in each probability vector, determining the number corresponding to each equipment category based on the equipment category corresponding to the maximum probability, and generating a second probability vector based on the number.

In one embodiment, a recorded broadcast video of a target class is input to an image equipment detection model, probability vectors of the rectangular frames, which are obtained by labeling the equipment in each picture frame, belonging to the equipment types are obtained, a maximum probability value is determined from the probability vectors corresponding to the rectangular frames, the equipment category corresponding to the maximum probability value is determined as the equipment category to which the corresponding equipment belongs, and a second probability vector is generated according to the equipment quantity corresponding to the equipment category.

As shown in fig. 10, the device a, the device B, the device C, the device D, and the device E are labeled to obtain corresponding devices respectivelyThe coordinates of the upper left corner position of each rectangular frame are obtained, the coordinates of the rectangular frame corresponding to the equipment A at the point a are (2, 2), the coordinates of the rectangular frame corresponding to the equipment B at the point B are (10, 2), the coordinates of the rectangular frame corresponding to the equipment C at the point C are (18, 2), the coordinates of the rectangular frame corresponding to the equipment D at the point D are (1, 10), the coordinates of the rectangular frame corresponding to the equipment E at the point E are (12, 11), the width and the height of each rectangular frame are obtained, the widths corresponding to the equipment A, the equipment B and the equipment C are all 4, the heights are all 6, the widths corresponding to the equipment D and the equipment E are all 8, the heights are all 5, no overlapping is determined according to the coordinates of each equipment, the equipment A, the equipment B and the equipment C are all X equipment, the equipment D and the equipment E are all Y equipment, and the generated second probability vector isOr->

And S208, splicing the first probability vector and the second probability vector to obtain a statistical probability vector, and inputting the statistical probability vector into an inference model to obtain an advanced score.

In one embodiment, the first probability vector and the second probability vector generated in the step S204 and the step S207 are spliced to obtain a statistical probability vector, the statistical probability vector is input into an inference model, and the advanced score of the target class is calculated.

The manner in which the first probability vector and the second probability vector are stitched may include, but is not limited to, the first probability vector being placed in front of the second probability vector, and the values of the first probability vector and the second probability vector being randomly fused to form a new probability vector.

As shown in fig. 3, a statistical probability vector is obtained by stitching the first probability vector and the second probability vector.

The inference model may include, but is not limited to, a deep neural network.

As shown in fig. 11, a statistical probability vector formed by splicing a first probability vector and a second probability vector is input into a deep neural network, each neuron corresponds to a probability value corresponding to each dimension of the statistical probability vector, the input statistical probability vector is multiplied by a parameter of an inference model to obtain a corresponding matrix, and each value in the matrix is summed to obtain an advanced score.

And S209, if the advanced score is greater than or equal to a score threshold, determining that the advanced of the target class reaches the standard.

In one embodiment, the advanced score is compared to a scoring threshold, and when the advanced score is greater than or equal to the scoring threshold, then the target class advanced is determined to meet the criteria.

The scoring threshold is preset by the user, and may be a scoring threshold of a percentage system or a scoring threshold of a ten system, which is not limited in this scheme.

If the scoring threshold is 90 minutes of the percentile, the first probability vector obtained through the language processing model and the second probability vector obtained through the image equipment detection model are spliced, and then input into the reasoning model to calculate the advanced degree of the target class, and if the scoring threshold is greater than the scoring threshold, the advanced degree of the target class reaches the standard.

And S210, if the advanced score is smaller than a score threshold, determining that the target class advanced does not reach the standard.

In one embodiment, the advancement score is compared to a scoring threshold, and when the advancement score is less than the scoring threshold, it is determined that the advancement of the target class does not meet the criteria.

If the scoring threshold is 8 minutes of ten, the first probability vector obtained through the language processing model and the second probability vector obtained through the image equipment detection model are spliced, the first probability vector and the second probability vector are input into the reasoning model to calculate the advanced degree of the target class, the advanced degree score of the target class is 7.9 minutes, and if the scoring threshold is smaller than the scoring threshold, the advanced degree of the target class is determined to be less than the standard.

In the embodiment of the application, through obtaining a classroom teaching plan text of a target classroom and a recorded video of the target classroom, inputting the classroom teaching plan text into a language processing model, outputting probability vectors corresponding to words in the classroom teaching plan text, determining the maximum probability in the probability vectors, determining the number of times of each target entity category in entity categories corresponding to the maximum probability, generating a first probability vector based on the number of times, wherein the target entity at least comprises a device name and a function name, inputting the recorded video of the target classroom into an image device detection model, outputting the probability vectors of each device belonging to different device categories in the recorded video, determining the maximum probability vector in the probability vectors, determining the number corresponding to each device category based on the devices corresponding to the maximum probability vector, generating a second probability vector based on the number, splicing the first probability vector and the second probability vector to obtain a statistical probability vector, inputting the statistical probability vector into an inference model, obtaining an advanced score, and if the advanced score of the target classroom reaches the advanced score standard or is smaller than the threshold, determining that the advanced score of the target classroom reaches the advanced score standard. Through corresponding classroom teaching plan texts and recorded broadcast videos of target classes input by users, inputting the classroom teaching plan texts into a language processing model to obtain probability vectors of all words belonging to all preset entity categories in the classroom teaching plan texts, determining entity categories corresponding to maximum probability values from all the probability vectors as entity categories of all the words, determining the occurrence times of the words of all the target categories from all the entity categories, generating a first probability vector based on the occurrence times of the words of all the target categories, inputting the recorded broadcast videos into an image equipment detection model to obtain probability vectors of all the equipment belonging to different equipment categories in the recorded broadcast videos, determining equipment categories corresponding to the maximum probability values from all the probability vectors as equipment categories of all the equipment, generating a second probability vector based on the equipment number of all the equipment categories, splicing the first probability vector and the second probability vector, inputting the first probability vector and the second probability vector into an inference model to obtain an advanced score of the target class, comparing the advanced score with a scoring threshold, and judging whether the advanced grade of the target class reaches a standard or not. And the evaluation of the advanced result of the target classroom is realized by combining the text of the classroom teaching plan of the target classroom and the recorded video.

The classroom advance assessment device according to the embodiment of the present application will be described in detail below with reference to fig. 12. Note that, the present application is not limited to the above-described embodiments. The classroom advanced assessment device in fig. 12 is used to execute the method of the embodiment of the present application shown in fig. 2 to 11, and for convenience of explanation, only the relevant parts of the embodiment of the present application are shown, and specific technical details are disclosed, please refer to the method embodiment of the present application shown in fig. 2 to 11.

Referring to fig. 12, a schematic diagram of a class advance assessment device is provided in the present application. As shown in fig. 12, the class advance evaluation apparatus 1 according to the embodiment of the present application includes: a data acquisition module 11, a first vector generation module 12, a second vector generation module 13, an advanced evaluation module 14.

The data acquisition module 11 is used for acquiring a classroom teaching plan text of a target classroom and a recorded broadcast video of the target classroom;

a first vector generation module 12, configured to obtain the number of occurrences of each term belonging to the target entity category in the text of the classroom teaching plan, and generate a first probability vector based on each of the number of occurrences;

a second vector generation module 13, configured to obtain the numbers of devices belonging to different device classes in the recorded broadcast video, and generate a second probability vector based on each of the numbers;

An advanced assessment module 14, configured to determine an advanced result of the target class based on the first probability vector and the second probability vector.

Optionally, as shown in fig. 13, the apparatus further includes:

and the information sending module 15 is configured to determine an substandard device in the target class if the advanced result does not reach the standard, and send a prompt message for replacing the device to the substandard device.

Optionally, the first vector generation module 12 is specifically configured to:

inputting the text of the classroom teaching plan into a language processing model, and outputting probability vectors corresponding to each word in the text of the classroom teaching plan;

determining the maximum probability in each probability vector, determining the times of each target entity class in entity classes corresponding to the maximum probability, and generating a first probability vector based on the times, wherein each target entity class at least comprises a device name and a function name.

Optionally, as shown in fig. 13, the apparatus further includes a language model training module 16 for:

training a language processing model through the probability sum of all words belonging to target entity categories of all sentences in the classroom teaching plan text, the probability vector of each word belonging to different entity categories in all sentences and the probability vector of adjacent words belonging to different entity categories in all the sentences in the classroom teaching plan text, and determining the model parameters of the trained language processing model.

Optionally, the second vector generation module 13 is specifically configured to:

inputting the recorded video into an image equipment detection model, and outputting probability vectors of different equipment categories of all equipment in the recorded video;

determining the maximum probability in each probability vector, determining the quantity corresponding to each equipment category based on the equipment category corresponding to the maximum probability, and generating a second probability vector based on the quantity.

Optionally, as shown in fig. 13, the apparatus further includes a detection module training module 17 for:

training the image equipment detection model through a loss function, wherein the loss function is obtained based on the probability that each equipment in the recorded broadcast video belongs to a target equipment category, the probability that each equipment in the recorded broadcast video does not belong to the target equipment category, the predicted coordinates of rectangular frames of each equipment and vectors corresponding to the predicted width and height, the actual coordinates of rectangular frames of each equipment and vectors corresponding to the actual width and height, the preset weight and the difference value between the actual coordinates and the predicted coordinates.

Optionally, the advanced evaluation module 14 is specifically configured to:

splicing the first probability vector and the second probability vector to obtain a statistical probability vector, and inputting the statistical probability vector into an inference model to obtain an advanced score;

If the advanced score is greater than or equal to a score threshold, determining that the advanced of the target class meets a standard;

and if the advanced score is smaller than a score threshold, determining that the target class advanced does not reach the standard.

The embodiment of the present application further provides a computer storage medium, where the computer storage medium may store a plurality of instructions, where the instructions are adapted to be loaded by a processor and execute the classroom advance assessment method according to the embodiment shown in fig. 2 to 11, and the specific execution process may refer to the specific description of the embodiment shown in fig. 2 to 11, which is not repeated herein.

Referring to fig. 14, a schematic structural diagram of an electronic device is provided in an embodiment of the present application. As shown in fig. 14, the terminal device 1000 may include: at least one processor 1001, at least one network interface 1002, at least one input output interface 1003, at least one display unit 1004, at least one memory 1005, at least one communication bus 1006. Wherein the processor 1001 may include one or more processing cores. Processor 1001 utilizes various interfaces and lines to connect various portions of the overall electronic device 1000, and performs various functions of terminal 1000 and processes data by executing or executing instructions, programs, code sets, or instruction sets stored in memory 1005, and invoking data stored in memory 1005. The memory 1005 may be a high-speed RAM memory or a non-volatile memory (non-volatile memory), such as at least one disk memory. The memory 1005 may also optionally be at least one storage device located remotely from the processor 1001. The network interface 1002 may optionally include a standard wired interface, a wireless interface (e.g., WI-FI interface, bluetooth interface), among others. The communication bus 1006 is used to enable connected communications between these components. The display unit 1004 may be a touch panel. As shown in fig. 13, an operating system, a network communication module, an input-output interface module, and a class-advanced evaluation program may be included in the memory 1005 as one type of storage medium.

In the electronic device 1000 shown in fig. 14, the input/output interface 1003 is mainly used to provide an interface for a user and an access device, and obtain data input by the user and the access device.

In one embodiment, the processor 1001 may be configured to call a class advance assessment program stored in the memory 1005, and specifically perform the following operations:

In one embodiment, the processor 1001 also performs the following:

if the advanced result does not reach the standard, determining the substandard equipment in the target class, and sending prompt information for replacing the equipment to the substandard equipment.

In one embodiment, the processor 1001, when executing the obtaining the number of occurrences of each term belonging to different entity categories in the text of the classroom teaching plan, generates the first probability vector based on each of the number of occurrences, specifically performs the following operations:

In one embodiment, the processor 1001 also performs the following:

In one embodiment, when executing the obtaining the number of devices belonging to different device classes in the recorded broadcast video, the processor 1001 specifically executes the following operations when generating the second probability vector based on each of the numbers:

In one embodiment, the processor 1001 also performs the following:

In one embodiment, the processor 1001, when executing the evaluation of the advanced class of the target class based on the first probability vector and the second probability vector, specifically performs the following operations:

In the embodiment of the application, a classroom teaching plan text of a target classroom and recorded broadcast video of the target classroom are obtained, the classroom teaching plan text is input into a language processing model, probability vectors corresponding to words in the classroom teaching plan text are output, maximum probability in the probability vectors is determined, the number of times of each target entity category is determined in entity categories corresponding to the maximum probability, a first probability vector is generated based on the number of times, wherein the target entity at least comprises a device name and a function name, recorded broadcast video of the target classroom is input into an image device detection model, probability vectors of different device categories in the recorded broadcast video are output, the maximum probability vector in the probability vectors is determined, the number corresponding to the device categories is determined based on the maximum probability vectors, the first probability vector and the second probability vector are spliced to obtain a statistical probability vector, the statistical probability vector is input into an inference model, and if the advanced score is greater than or equal to a score threshold, the advanced score of the target classroom reaches a score standard, and if the advanced score of the target classroom is less than the score threshold is not determined. Through corresponding classroom teaching plan texts and recorded broadcast videos of target classes input by users, inputting the classroom teaching plan texts into a language processing model to obtain probability vectors of all words belonging to all preset entity categories in the classroom teaching plan texts, determining entity categories corresponding to maximum probability values from all the probability vectors as entity categories of all the words, determining the occurrence times of the words of all the target categories from all the entity categories, generating a first probability vector based on the occurrence times of the words of all the target categories, inputting the recorded broadcast videos into an image equipment detection model to obtain probability vectors of all the equipment belonging to different equipment categories in the recorded broadcast videos, determining equipment categories corresponding to the maximum probability values from all the probability vectors as equipment categories of all the equipment, generating a second probability vector based on the equipment number of all the equipment categories, splicing the first probability vector and the second probability vector, inputting the first probability vector and the second probability vector into an inference model to obtain an advanced score of the target class, comparing the advanced score with a scoring threshold, and judging whether the advanced grade of the target class reaches a standard or not. And the target classroom advanced evaluation is realized by combining the classroom teaching plan text and recorded video of the target classroom.

It will be clear to a person skilled in the art that the solution according to the application can be implemented by means of software and/or hardware. "Unit" and "module" in this specification refer to software and/or hardware capable of performing a specific function, either alone or in combination with other components, such as Field programmable gate arrays (Field-ProgrammaBLE Gate Array, FPGAs), integrated circuits (Integrated Circuit, ICs), etc.

It should be noted that, for simplicity of description, the foregoing method embodiments are all described as a series of acts, but it should be understood by those skilled in the art that the present application is not limited by the order of acts described, as some steps may be performed in other orders or concurrently in accordance with the present application. Further, those skilled in the art will also appreciate that the embodiments described in the specification are all preferred embodiments, and that the acts and modules referred to are not necessarily required for the present application.

In the foregoing embodiments, the descriptions of the embodiments are emphasized, and for parts of one embodiment that are not described in detail, reference may be made to related descriptions of other embodiments.

In the several embodiments provided by the present application, it should be understood that the disclosed apparatus may be implemented in other manners. For example, the apparatus embodiments described above are merely illustrative, such as the division of the units, merely a logical function division, and there may be additional manners of dividing the actual implementation, such as multiple units or components may be combined or integrated into another system, or some features may be omitted, or not performed. Alternatively, the coupling or direct coupling or communication connection shown or discussed with each other may be through some service interface, device or unit indirect coupling or communication connection, electrical or otherwise.

The units described as separate units may or may not be physically separate, and units shown as units may or may not be physical units, may be located in one place, or may be distributed on a plurality of network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

In addition, each functional unit in the embodiments of the present application may be integrated in one processing unit, or each unit may exist alone physically, or two or more units may be integrated in one unit. The integrated units may be implemented in hardware or in software functional units.

The integrated units, if implemented in the form of software functional units and sold or used as stand-alone products, may be stored in a computer readable memory. Based on such understanding, the technical solution of the present application may be embodied essentially or partly in the form of a software product or all or part of the technical solution, which is stored in a memory, and includes several instructions for causing an electronic device (which may be a personal computer, a server, a network device, or the like) to perform all or part of the steps of the method according to the embodiments of the present application. And the aforementioned memory includes: a U-disk, a Read-Only Memory (ROM), a random access Memory (Random Access Memory, RAM), a removable hard disk, a magnetic disk, or an optical disk, or other various media capable of storing program codes.

Those of ordinary skill in the art will appreciate that all or a portion of the steps in the various methods of the above embodiments may be performed by hardware associated with a program that is stored in a computer readable memory, which may include: flash disk, read-Only Memory (ROM), random-access Memory (Random Access Memory, RAM), magnetic or optical disk, and the like.

The foregoing is merely exemplary embodiments of the present disclosure and is not intended to limit the scope of the present disclosure. That is, equivalent changes and modifications are contemplated by the teachings of this disclosure, which fall within the scope of the present disclosure. Other embodiments of the disclosure will be apparent to those skilled in the art from consideration of the specification and practice of the disclosure herein. This application is intended to cover any variations, uses, or adaptations of the disclosure following, in general, the principles of the disclosure and including such departures from the present disclosure as come within known or customary practice within the art to which the disclosure pertains. It is intended that the specification and examples be considered as exemplary only, with a scope and spirit of the disclosure being indicated by the claims.

Claims

1. A classroom advance assessment method, the method comprising:

2. The method of claim 1, wherein after determining the advanced result for the target class based on the first probability vector and the second probability vector, further comprising:

3. The method of claim 1, wherein the obtaining the number of occurrences of each term belonging to a different entity category in the classroom case text, generating a first probability vector based on each of the number of occurrences, comprises:

4. The method of claim 1, wherein the inputting the text of the classroom case into a language processing model, before outputting the probability vector corresponding to each word in the text of the classroom case, further comprises:

5. The method of claim 1, wherein the obtaining the number of devices belonging to different device categories in the recorded video, generating a second probability vector based on each of the numbers, comprises:

And determining the maximum probability vector in the probability vectors, determining the number corresponding to each equipment category based on the equipment category corresponding to the maximum probability, and generating a second probability vector based on the number.

6. The method of claim 1, wherein the inputting the recorded video into an image device detection model, before outputting the maximum probability that each device in the recorded video belongs to a different device class, further comprises:

7. The method of claim 1, wherein the evaluating the advancement of the target class based on the first probability vector and the second probability vector comprises:

8. A class advance assessment apparatus, the apparatus comprising:

9. A computer storage medium storing a plurality of instructions adapted to be loaded by a processor and to perform the method steps of any of claims 1-7.

10. An electronic device, comprising: a processor and a memory; wherein the memory stores a computer program adapted to be loaded by the processor and to perform the method steps of any of claims 1-7.