Digital intelligent teaching recording and broadcasting system and method

Abstract

The invention discloses a digital intelligent teaching recording and broadcasting system and method, and relates to the technical field of data encryption, and the method comprises the steps: collecting teaching videos and audios of a teacher through a high-definition camera and an audio device, carrying out the preprocessing, obtaining the course content, recognizing the key information of the course content through a natural language processing NLP and computer vision technology, and carrying out the recording and broadcasting. The method comprises the following steps: segmenting a video into single knowledge point fragments, performing classified storage, forming a learning resource library, collecting basic information and learning habit data of students, constructing a user portrait, matching learning resources, extracting the learning habit data in the user portrait, performing feature extraction, and obtaining a comprehensive feature vector; constructing a learning state evaluation model by using a convolutional neural network to comprehensively analyze the comprehensive feature vector to obtain a learning state comprehensive score, formulating a playing strategy according to the learning state comprehensive score and matched learning resources, regularly testing knowledge points of students, evaluating the mastering degree of the students, and updating a user portrait according to a result.

Description

Technical Field

[0001] The present invention relates to the field of educational information technology, and in particular to a digital intelligent teaching recording and broadcasting system and method. Background Art

[0002] With the rapid development of information technology, digital education has gradually become an important part of the education field. In recent years, with the increase in Internet penetration and the enhancement of mobile device performance, digital intelligent teaching systems have been widely used and promoted. Traditional teaching methods often rely on face-to-face classroom communication. However, this method is limited by time and space and cannot meet the needs of modern education. In order to overcome these limitations, the emergence of digital intelligent teaching recording and broadcasting systems has made distance education possible, allowing high-quality educational resources to transcend geographical boundaries and be shared. In addition, with the advancement of big data technology and artificial intelligence algorithms, how to effectively integrate these advanced technologies to optimize the teaching process has become a research hotspot. There are already some preliminary attempts on the market, which can solve the problem of uneven distribution of educational resources to a certain extent.

[0003] However, in actual applications, it is found that the existing digital intelligent teaching recording and broadcasting systems still have many problems. During the course recording process, due to the lack of effective pre-processing methods, the quality of the audio and video materials obtained is uneven, which affects the subsequent teaching. In terms of content organization, the traditional method usually simply arranges the videos in order of chapters, fails to deeply explore the internal connections between each knowledge point, and is not conducive to students forming a structured knowledge system. In terms of personalized recommendations, although some platforms try to push relevant content based on user preferences, the lack of in-depth understanding of individual learning habits leads to poor recommendation effects. In the evaluation of learning effects, existing platforms mostly rely on subjective scoring and simple objective test questions, which cannot fully and accurately reflect the students' real learning status. Therefore, how to improve the intelligence level of the digital intelligent teaching recording and broadcasting system so that it can better serve personalized learning needs is one of the current problems to be solved. Summary of the invention

[0004] In view of the above existing problems, the present invention is proposed.

[0005] Therefore, the present invention provides a digital intelligent teaching recording and broadcasting system and method to solve the problem that in the course recording process, due to the lack of effective pre-processing means, the quality of the audio and video materials obtained is uneven, which affects the subsequent teaching.

[0006] In order to solve the above technical problems, the present invention provides the following technical solutions:

[0007] In a first aspect, an embodiment of the present invention provides a digital intelligent teaching recording and broadcasting method, which comprises:

[0008] The teacher's lecture video and audio are collected through high-definition video and audio equipment, and pre-processed to obtain the course content;

[0009] Use natural language processing (NLP) and computer vision technology to identify key information in the course content, divide the video into individual knowledge point segments, classify and store them, and form a learning resource library;

[0010] Collect students' basic information and learning habits data, build personalized user portraits, and select matching learning resources based on the user portraits;

[0011] Extract learning habit data from user portraits, perform feature extraction, and obtain comprehensive feature vectors;

[0012] Use convolutional neural network (CNN) to build a learning state evaluation model to conduct a comprehensive analysis of the comprehensive feature vector, obtain a comprehensive learning state score, set a threshold, and formulate a playback strategy based on the comprehensive learning state score combined with matching learning resources;

[0013] Regularly test students on knowledge points to assess their level of mastery, and update user portraits based on the assessment results.

[0014] As a preferred solution of the digital intelligent teaching recording and broadcasting method of the present invention, wherein: the teacher's teaching video and audio are collected by high-definition video and audio equipment, and pre-processed, and the specific steps are:

[0015] Install high-definition cameras and microphones in classrooms to record lecture videos and audio;

[0016] Adjust the brightness, contrast and color of the collected teaching video materials;

[0017] Wavelet transform is used to reduce noise on the collected lecture audio;

[0018] The processed video and audio data are stored in the cloud database.

[0019] As a preferred solution of the digital intelligent teaching recording and broadcasting method of the present invention, wherein: the key information of the course content is identified by using natural language processing NLP and computer vision technology, the video is divided into single knowledge point segments, classified and stored, and a learning resource library is formed. The specific steps are:

[0020] Use computer vision technology to analyze lecture videos, identify the visual information of blackboard content and PPT presentations, and extract text information from them;

[0021] Apply natural language processing technology to transcribe lecture audio and convert speech into text form;

[0022] Perform word segmentation and part-of-speech tagging on the extracted text, use natural language processing (NLP) technology to extract keywords and recognize named entities, identify key terms, concepts, and sentences, determine which parts represent independent knowledge points, combine the switching in visual information with the logical pauses and marks in speech transcription, determine the segmentation points, and use FFmpeg to segment the video into small clips containing independent knowledge points;

[0023] Each segmented knowledge point fragment is labeled with the name of the knowledge point, the subject area to which it belongs, and the level of difficulty, and is stored in a classified manner.

[0024] As a preferred solution of the digital intelligent teaching recording and broadcasting method of the present invention, wherein: the basic information and learning habit data of students are collected, personalized user portraits are constructed, and matching learning resources are selected based on the user portraits. The specific steps are as follows:

[0025] Obtain students' name, age, gender, grade, and major information through the registration form;

[0026] Record students’ study time preferences, study frequency, study duration, video viewing time, number of pauses, number of replays, skipped parts, participation in interactive questions, and feedback information;

[0027] Clean, process and analyze the collected data, use association rule mining tools to identify students’ learning patterns and preferences, extract key features, and generate labels for students;

[0028] Integrate all label information to form a multi-dimensional user portrait that integrates students' basic attributes, learning habits, preferences, and ability evaluation. The expression is:

[0029] U=(a,g,l,p,w,f,t,h,p c , r c ,s n , q, b, d);

[0030] Among them, a represents the student's age, g represents gender, l represents grade, p represents major, w represents the time of watching videos, f represents learning frequency, t represents learning time, h represents learning time preference, and p c Indicates the number of pauses, r c Indicates the number of playbacks, s n represents the number of skipped parts, q represents whether to participate in the interactive questions, b represents the average score of feedback information, and d represents the learning time;

[0031] According to the information in the user portrait, a comprehensive evaluation function is constructed to evaluate the matching degree E between learning resources and students. The expression is:

[0032]

[0033] Among them, N represents the number of majors, and p j represents the jth major;

[0034] The value range of E is (-∞, ∞). Set a threshold T1. When E < T1, it means that the matching degree between the learning resource and the student is low and it cannot be used by the student;

[0035] When E ≥ T1, it means that the matching degree between the learning resource and the student is high and it can be used by the student.

[0036] As a preferred scheme of the digital intelligent teaching recording and broadcasting method described in the present invention, wherein: extracting the learning habit data in the user portrait, performing feature extraction, and obtaining a comprehensive feature vector, the specific steps are as follows:

[0037] Collect data on learning time preference, learning frequency, learning duration, time of watching videos, number of pauses, number of replays, and number of skipped parts from the user portrait, and perform standardization processing using Z-score;

[0038] Construct a comprehensive feature vector X, which includes the standardized features, and the expression is:

[0039]

[0040] Among them, T represents the learning time preference, and Z T represents the standardized value of the learning time preference, F represents the learning frequency, and Z F represents the standardized value of the learning frequency, D represents the learning duration, and Z D represents the standardized value of the learning duration, W represents the time of watching videos, and Z W represents the standardized value of the time of watching videos, P represents the number of pauses, and Z P represents the standardized value of the number of pauses, R represents the number of replays, and Z R represents the standardized value of the number of replays, and S represents the number of skipped parts.

[0041] As a preferred scheme of the digital intelligent teaching recording and broadcasting method described in the present invention, wherein: using the convolutional neural network CNN to construct a learning state evaluation model to comprehensively analyze the comprehensive feature vector, obtaining a comprehensive learning state score, setting a threshold, and formulating a playback strategy according to the comprehensive learning state score combined with the matched learning resources, the specific steps are as follows:

[0042] Use the convolutional neural network CNN to construct a learning state evaluation model, input the comprehensive feature vector X, and output the comprehensive learning state score S, and the expression is:

[0043] S = M(X) = σ(W2·σ(W1·X + b1) + b2);

[0044] Among them, W1 represents the weight matrix from the input layer to the hidden layer, W2 represents the weight matrix from the hidden layer to the output layer, b1 represents the bias term of the hidden layer, b2 represents the bias term of the output layer, and σ represents the activation function ReLU;

[0045] The value range of S is [-1, 1]. Define the threshold T2. When -1 ≤ S ≤ T2, it indicates that the student's state is negative, and the video playback speed is reduced. Low-difficulty interactive questions and small tests are inserted in combination with the matching learning resources;

[0046] When T2 < S ≤ 1, it indicates that the student's state is positive, and the advanced video of the matching learning resource is played.

[0047] As a preferred solution of the digital intelligent teaching recording and broadcasting method described in the present invention, wherein: regularly conduct knowledge point tests on students, evaluate their mastery level, and update the user portrait according to the evaluation results. The specific steps are as follows:

[0048] Regularly conduct knowledge point tests on students. Set the threshold T3. When the test score C ≤ T3, it indicates that the mastery level is insufficient, and the current user portrait and the current learning resources are maintained;

[0049] When the test score C > T3, it indicates that the mastery level of the current learning resources is relatively high, and the user portrait is updated, and advanced learning resources are provided. The update expression is:

[0050]

[0051] Among them, U' represents the updated user portrait, U represents the user portrait before update, C represents the test score of the student, C avg represents the average value of the test scores, σ C represents the standard deviation of the test scores, w p represents the weight of the pause times, w r represents the weight of the playback times, w t represents the weight of the learning time, w t represents the weight of the learning time.

[0052] In the second aspect, the present invention provides a digital intelligent teaching recording and broadcasting system, including,

[0053] Multimedia acquisition module: Collect the teacher's teaching video and audio through high-definition cameras and audio devices, and perform preprocessing to obtain clear course content;

[0054] Content segmentation module: Use natural language processing NLP and computer vision technology to identify the key information of the course content, segment the video into individual knowledge point segments, classify and store them to form a learning resource library;

[0055] User portrait building module: collects students’ basic information and learning habits data, builds user portraits, and provides customized learning resources for students;

[0056] Feature extraction module: extracts learning habit data from user portraits, performs feature extraction, and obtains a comprehensive feature vector that can reflect students’ learning behavior patterns;

[0057] Learning status evaluation module: Use convolutional neural network (CNN) to build a learning status evaluation model, conduct a comprehensive analysis of the comprehensive feature vector, obtain a comprehensive learning status score, and adjust the playback strategy based on the score;

[0058] Learning effect evaluation module: Regularly test students’ knowledge points to assess their mastery, and update user portraits based on the test results to optimize future learning resource recommendations.

[0059] In a third aspect, an embodiment of the present invention provides a computer device, comprising a memory and a processor, wherein the memory stores a computer program, wherein: when the computer program is executed by the processor, any step of the digital intelligent teaching recording and broadcasting method described in the first aspect of the present invention is implemented.

[0060] In a fourth aspect, an embodiment of the present invention provides a computer-readable storage medium having a computer program stored thereon, wherein: when the computer program is executed by a processor, any step of the digital intelligent teaching recording and broadcasting method described in the first aspect of the present invention is implemented.

[0061] The beneficial effects of the present invention are: not only the quality and technical level of teaching recording and broadcasting are improved, but also the personalization and intelligence of learning are greatly enhanced. Through high-quality content collection and processing, accurate knowledge point division, personalized learning resource recommendation and dynamic learning status evaluation, the invention significantly improves the effect and experience of online learning, improves the utilization rate of educational resources, and also provides new ideas and technical support for the development of educational technology. BRIEF DESCRIPTION OF THE DRAWINGS

[0062] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without paying creative work.

[0063] Figure 1 This is a flow chart of the digital intelligent teaching recording and broadcasting method in Example 1.

[0064] Figure 2 This is a flow chart of the digital intelligent teaching recording and broadcasting system in Example 1. DETAILED DESCRIPTION

[0065] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are described in detail below in conjunction with the accompanying drawings.

[0066] In the following description, many specific details are set forth to facilitate a full understanding of the present invention, but the present invention may also be implemented in other ways different from those described herein, and those skilled in the art may make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0067] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The term "in one embodiment" that appears in different places in this specification does not necessarily refer to the same embodiment, nor does it refer to a separate or selective embodiment that is mutually exclusive with other embodiments.

[0068] Example 1, reference Figure 1 and Figure 2 , which is the first embodiment of the present invention, and provides a digital intelligent teaching recording and broadcasting method, comprising the following steps:

[0069] S1 collects the teacher's lecture video and audio through high-definition video and audio equipment and performs pre-processing;

[0070] Install high-definition cameras and microphones in classrooms to record lecture videos and audio;

[0071] Traditional recording and broadcasting systems have problems such as fuzzy video and audio noise due to equipment limitations or improper processing, which affects the learning experience. The present invention significantly improves the quality of video and audio through high-definition equipment and advanced processing technology, making learning materials clearer and easier to understand, thereby improving learning efficiency;

[0072] Adjust the brightness, contrast and color of the collected teaching video materials;

[0073] Wavelet transform is used to reduce noise on the collected lecture audio;

[0074] By installing high-definition cameras and microphones in the classroom to record lecture videos and audios, the brightness, contrast and color of the video materials are adjusted, and the audio is denoised using wavelet transform to ensure the quality of the final data;

[0075] The processed video and audio data are stored in the cloud database.

[0076] S2 uses natural language processing (NLP) and computer vision technology to identify key information in the course content, divide the video into individual knowledge point segments, and store them in categories to form a learning resource library;

[0077] Use computer vision technology to analyze lecture videos, identify the visual information of blackboard content and PPT presentations, and extract text information from them;

[0078] Apply natural language processing technology to transcribe lecture audio and convert speech into text form;

[0079] Perform word segmentation and part-of-speech tagging on the extracted text, use natural language processing (NLP) technology to extract keywords and recognize named entities, identify key terms, concepts, and sentences, determine which parts represent independent knowledge points, combine the switching in visual information with the logical pauses and marks in speech transcription, determine the segmentation points, and use FFmpeg to segment the video into small clips containing independent knowledge points;

[0080] Traditional teaching videos lack effective knowledge point segmentation methods, which makes it difficult for students to quickly find the required content. The present invention uses automation technology to achieve intelligent segmentation and classified storage of video content, which is convenient for students to retrieve knowledge points on demand, and improves the flexibility and efficiency of learning. It uses computer vision technology to analyze teaching videos, extract blackboard content and PPT presentation information, and transcribe audio into text in combination with natural language processing technology to further identify key information. Combined with visual information and voice logic pauses, the video is segmented into single knowledge point segments using FFmpeg;

[0081] Each segmented knowledge point fragment is labeled with the name of the knowledge point, the subject area to which it belongs, and the degree of difficulty, and is stored in a classified manner;

[0082] When explaining a complex mathematical theorem, natural language processing technology can be used to identify the various components of the theorem, and combined with the blackboard switching in the visual information, the theorem can be divided into several knowledge point segments such as definition, derivation process, conclusion, etc., to facilitate students' understanding and review.

[0083] S3 collects students’ basic information and learning habits data, builds personalized user portraits, and selects matching learning resources based on the user portraits;

[0084] Obtain students' name, age, gender, grade, and major information through the registration form;

[0085] Record students’ study time preferences, study frequency, study duration, video viewing time, number of pauses, number of replays, skipped parts, participation in interactive questions, and feedback information;

[0086] Clean, process, and analyze the collected data, use association rule mining tools to identify students' learning patterns and preferences, extract key features, and generate labels for students;

[0087] Integrate all label information to form a multi-dimensional user profile that integrates students' basic attributes, learning habits, preferences, and ability assessments. The expression is:

[0088] U = (a, g, l, p, w, f, t, h, p c , r c , s n , q, b, d);

[0089] Among them, a represents the student's age, g represents gender, l represents grade, p represents major, w represents the time of watching videos, f represents learning frequency, t represents learning time, h represents learning time preference,, p c represents the number of pauses, r c represents the number of replays, s n represents the number of skipped parts, q represents whether to participate in interactive questions, the value range is {0, 1}, b represents the average score of feedback information, and d represents learning duration;

[0090] The resources provided by traditional online learning platforms are often one-size-fits-all and do not fully consider individual differences. Through the construction of user profiles, the present invention realizes personalized recommendation of learning resources, enabling each student to obtain the most suitable learning materials for themselves and improving the pertinence and effectiveness of learning;

[0091] According to the information in the user profile, construct a comprehensive evaluation function to evaluate the matching degree E between learning resources and students. The expression is:

[0092]

[0093] Among them, N represents the number of majors, p j represents the jth major;

[0094] The value range of E is (-∞, ∞). Set a threshold T1. When E < T1, it means that the matching degree between learning resources and students is low and cannot be used by students;

[0095] When E ≥ T1, it means that the matching degree between learning resources and students is high and can be used by students.

[0096] S4 Extract the learning habit data in the user profile, perform feature extraction, and obtain a comprehensive feature vector;

[0097] Collect data on learning time preference, learning frequency, learning duration, time of watching videos, number of pauses, number of replays, and number of skipped parts from the user profile, and perform standardization processing using Z-score;

[0098] Through Z-score standardization, different features can be compared at the same scale, which improves the consistency and reliability of feature vectors;

[0099] Construct a comprehensive feature vector X, which contains the standardized features, and the expression is:

[0100]

[0101] Among them, T represents learning time preference, Z T represents the standardized value of learning time preference, F represents learning frequency, and Z F represents the standardized value of learning frequency, D represents the learning duration, and Z D represents the standardized value of learning time, W represents the time of watching videos, and Z W represents the normalized value of the time of watching the video, P represents the number of pauses, and Z P represents the normalized value of the number of pauses, R represents the number of replays, and Z R represents the normalized value of the number of playbacks, and S represents the number of skipped parts;

[0102] Learning habit data are collected from user portraits and standardized using Z-score to construct a comprehensive feature vector for subsequent learning status evaluation. The learning habit data of different students may have different dimensions and scales, and direct comparison will lead to deviations.

[0103] S5 uses convolutional neural network (CNN) to build a learning state evaluation model to conduct a comprehensive analysis of the comprehensive feature vector, obtain a comprehensive learning state score, set a threshold, and formulate a playback strategy based on the comprehensive learning state score combined with matching learning resources;

[0104] The convolutional neural network CNN is used to build a learning state evaluation model. The comprehensive feature vector X is input and the comprehensive score S of the learning state is output. The expression is:

[0105] S=M(X)=σ(W2·σ(W1·X+b1)+b2);

[0106] Among them, W1 represents the weight matrix from the input layer to the hidden layer, W2 represents the weight matrix from the hidden layer to the output layer, b1 represents the bias term of the hidden layer, b2 represents the bias term of the output layer, and σ represents the activation function ReLU;

[0107] The value range of S is [-1,1], and the threshold T2 is defined. When -1≤S≤T2, it means that the student is in a negative state. The video playback speed is reduced, and low-difficulty interactive questions and small tests are inserted in combination with matching learning resources;

[0108] When T2 < S ≤ 1, it indicates that the student's state is positive, and the advanced video of the matching learning resource is played;

[0109] Traditional online learning platforms lack a real-time monitoring mechanism for students' learning status and cannot dynamically adjust teaching strategies. Through the introduction of a CNN model, this invention realizes the dynamic evaluation of students' learning status, can adjust the video playback speed and insert interactive questions according to the actual performance of students, and enhances the interactivity and adaptability of learning.

[0110] S6 Regularly conduct knowledge point tests on students, evaluate their mastery level, and update the user profile according to the evaluation results;

[0111] Regularly conduct knowledge point tests on students, set a threshold T3. When the test score C ≤ T3, it indicates insufficient mastery, keep the current user profile, and maintain the current learning resources;

[0112] When the test score C > T3, it indicates a relatively high mastery of the current learning resources, update the user profile, and provide advanced learning resources. The update expression is:

[0113]

[0114] Among them, U' represents the updated user profile, U represents the user profile before update, C represents the student's test score, C avg represents the average value of the test scores, σ C represents the standard deviation of the test scores, w p represents the weight of the pause times, w r represents the weight of the replay times, w t represents the weight of the learning time, w t represents the weight of the learning time;

[0115] By regularly testing to evaluate students' mastery of knowledge points and updating the user profile according to the test results, the learning situation of students changes over time, and a fixed user profile cannot reflect the latest learning status. Through regular testing and updating the user profile, the system can timely adjust the learning resource recommendation and better adapt to the learning progress of students.

[0116] This embodiment also provides a digital intelligent teaching recording and broadcasting system, including:

[0117] Multimedia acquisition module: Collect the teacher's teaching video and audio through high-definition cameras and audio devices, and perform preprocessing to obtain clear course content;

[0118] Content segmentation module: Use natural language processing NLP and computer vision technologies to identify the key information of the course content, segment the video into individual knowledge point segments, classify and store them to form a learning resource library;

[0119] User portrait building module: collects students’ basic information and learning habits data, builds user portraits, and provides customized learning resources for students;

[0120] Feature extraction module: extracts learning habit data from user portraits, performs feature extraction, and obtains a comprehensive feature vector that can reflect students’ learning behavior patterns;

[0121] Learning status evaluation module: Use convolutional neural network (CNN) to build a learning status evaluation model, conduct a comprehensive analysis of the comprehensive feature vector, obtain a comprehensive learning status score, and adjust the playback strategy based on the score;

[0122] Learning effect evaluation module: Regularly test students’ knowledge points to assess their mastery, and update user portraits based on the test results to optimize future learning resource recommendations.

[0123] This embodiment also provides a computer device, which is suitable for the digital intelligent teaching recording and broadcasting method, including: a memory and a processor; the memory is used to store computer executable instructions, and the processor is used to execute computer executable instructions to implement the digital intelligent teaching recording and broadcasting method proposed in the above embodiment.

[0124] The computer device may be a terminal, and the computer device includes a processor, a memory, a communication interface, a display screen and an input device connected via a system bus. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The communication interface of the computer device is used to communicate with an external terminal in a wired or wireless manner, and the wireless manner can be achieved through WIFI, an operator network, NFC (near field communication) or other technologies. The display screen of the computer device may be a liquid crystal display screen or an electronic ink display screen, and the input device of the computer device may be a touch layer covering the display screen, or a key, trackball or touchpad provided on the housing of the computer device, or an external keyboard, touchpad or mouse, etc.

[0125] This embodiment also provides a storage medium on which a computer program is stored. When the program is executed by a processor, the method for recording and broadcasting digital intelligent teaching proposed in the above embodiment is implemented; the storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (Static Random Access Memory, referred to as SRAM), electrically erasable programmable read-only memory (Electrically Erasable Programmable Read-Only Memory, referred to as EEPROM), erasable programmable read-only memory (Erasable Programmable Read Only Memory, referred to as EPROM), programmable read-only memory (Programmable Red-Only Memory, referred to as PROM), read-only memory (Read-Only Memory, referred to as ROM), magnetic storage, flash memory, magnetic disk or optical disk.

[0126] In summary, the present invention not only improves the quality and technical level of teaching recording and broadcasting, but also greatly enhances the personalization and intelligence of learning. Through high-quality content collection and processing, accurate knowledge point division, personalized learning resource recommendation and dynamic learning status evaluation, the invention significantly improves the effect and experience of online learning, improves the utilization rate of educational resources, and also provides new ideas and technical support for the development of educational technology.

Claims

1. A digital intelligent teaching recording and broadcasting method, characterized in that: include, The teacher's lecture video and audio are collected through high-definition video and audio equipment, and pre-processed to obtain the course content; Use natural language processing (NLP) and computer vision technology to identify key information in the course content, divide the video into individual knowledge point segments, classify and store them, and form a learning resource library; Collect students' basic information and learning habits data, build personalized user portraits, and select matching learning resources based on the user portraits; Extract learning habit data from user portraits, perform feature extraction, and obtain comprehensive feature vectors; Use convolutional neural network (CNN) to build a learning state evaluation model to conduct a comprehensive analysis of the comprehensive feature vector, obtain a comprehensive learning state score, set a threshold, and formulate a playback strategy based on the comprehensive learning state score combined with matching learning resources; Regularly test students on knowledge points to assess their level of mastery, and update user portraits based on the assessment results.

2. The digital intelligent teaching recording and broadcasting method according to claim 1, characterized in that: The teacher's teaching video and audio are collected by high-definition video and audio equipment, and pre-processed. The specific steps are as follows: Install high-definition cameras and microphones in classrooms to record lecture videos and audio; Adjust the brightness, contrast and color of the collected teaching video materials; Wavelet transform is used to reduce noise on the collected lecture audio; The processed video and audio data are stored in the cloud database.

3. The digital intelligent teaching recording and broadcasting method according to claim 2, characterized in that: The specific steps of using natural language processing (NLP) and computer vision technology to identify key information in the course content, dividing the video into individual knowledge point segments, and classifying and storing them to form a learning resource library are as follows: Use computer vision technology to analyze lecture videos, identify the visual information of blackboard content and PPT presentations, and extract text information from them; Apply natural language processing technology to transcribe lecture audio and convert speech into text form; Perform word segmentation and part-of-speech tagging on the extracted text, use natural language processing (NLP) technology to extract keywords and recognize named entities, identify key terms, concepts, and sentences, determine which parts represent independent knowledge points, combine the switching in visual information with the logical pauses and marks in speech transcription, determine the segmentation points, and use FFmpeg to segment the video into small clips containing independent knowledge points; Each segmented knowledge point fragment is labeled with the name of the knowledge point, the subject area to which it belongs, and the level of difficulty, and is stored in a classified manner.

4. The digital intelligent teaching recording and broadcasting method according to claim 3, characterized in that: The specific steps of collecting students' basic information and learning habit data, building personalized user portraits, and selecting matching learning resources based on the user portraits are as follows: Obtain students' name, age, gender, grade, and major information through the registration form; Record students’ study time preferences, study frequency, study duration, video viewing time, number of pauses, number of replays, skipped parts, participation in interactive questions, and feedback information; Clean, process and analyze the collected data, use association rule mining tools to identify students’ learning patterns and preferences, extract key features, and generate labels for students; Integrate all label information to form a multi-dimensional user portrait that integrates students' basic attributes, learning habits, preferences, and ability evaluation. The expression is: U=(a,g,l,p,w,f,t,h,p c ,r c ,s n ,q,b,d); Among them, a represents the student's age, g represents gender, l represents grade, p represents major, w represents the time of watching videos, f represents learning frequency, t represents learning time, h represents learning time preference, and p c Indicates the number of pauses, r c Indicates the number of playbacks, s n represents the number of skipped parts, q represents whether to participate in the interactive questions, b represents the average score of feedback information, and d represents the learning time; According to the information in the user profile, construct a comprehensive evaluation function to evaluate the matching degree E between the learning resource and the student. The expression is as follows: Where N represents the number of majors, p j represents the jth major; The value range of E is (-∞, ∞). Set a threshold T1. When E < T1, it means that the matching degree between the learning resource and the student is low, and the learning resource cannot be used by the student; When E ≥ T1, it means that the matching degree between the learning resource and the student is high, and the learning resource can be used by the student.

5. The digital intelligent teaching recording and broadcasting method according to claim 4, characterized in that: The steps of extracting the learning habit data in the user profile, performing feature extraction, and obtaining the comprehensive feature vector are as follows: Collect data on learning time preference, learning frequency, learning duration, video viewing time, pause times, replay times, and the number of skipped parts from the user profile, and perform standardization processing using Z-score; Construct a comprehensive feature vector X, which contains the standardized features. The expression is as follows: Among them, T represents learning time preference, Z T represents the standardized value of learning time preference, F represents learning frequency, and Z F represents the standardized value of learning frequency, D represents the learning duration, and Z D represents the standardized value of learning time, W represents the time of watching videos, and Z W represents the normalized value of the time of watching the video, P represents the number of pauses, and Z P represents the normalized value of the number of pauses, R represents the number of replays, and Z R It represents the normalized value of the number of replays, and S represents the number of skipped parts.

6. The digital intelligent teaching recording and broadcasting method according to claim 5, characterized in that: Use the convolutional neural network CNN to construct a learning state evaluation model to comprehensively analyze the comprehensive feature vector, obtain the comprehensive learning state score, set a threshold, and formulate a playback strategy based on the comprehensive learning state score and the matched learning resources. The specific steps are as follows: Use the convolutional neural network CNN to construct a learning state evaluation model. Input the comprehensive feature vector X and output the comprehensive learning state score S. The expression is as follows: S = M(X) = σ(W2·σ(W1·X + b1) + b2); Among them, W1 represents the weight matrix from the input layer to the hidden layer, W2 represents the weight matrix from the hidden layer to the output layer, b1 represents the bias term of the hidden layer, b2 represents the bias term of the output layer, and σ represents the activation function ReLU; The value range of S is [-1, 1]. Define a threshold T2. When -1 ≤ S ≤ T2, it means that the student's state is negative. Reduce the video playback speed, and insert low-difficulty interactive questions and small tests in combination with the matched learning resources; When T2 < S ≤ 1, it means that the student's state is positive, and play the advanced video of the matched learning resource.

7. The digital intelligent teaching recording and broadcasting method according to claim 6, characterized in that: The steps of regularly testing students on knowledge points, evaluating their mastery level, and updating the user profile according to the evaluation results are as follows: Regularly test students on knowledge points. Set a threshold T3. When the test score C ≤ T3, it means that the mastery level is insufficient, keep the current user profile, and keep the current learning resources; When the test score C > T3, it means that the mastery level of the current learning resource is relatively high. Update the user profile, provide advanced learning resources, and the update expression is as follows: Among them, U' represents the updated user profile, U represents the user profile before the update, C represents the student's test score, and C avg represents the average of the test scores, σ C represents the standard deviation of the test scores, w p represents the weight of the number of pauses, w r Represents the weight of the number of replays, w t represents the weight of learning time, w t Represents the weight of learning time.

8. A digital intelligent teaching recording and broadcasting system, based on the digital intelligent teaching recording and broadcasting method according to any one of claims 1 to 7, characterized in that: Including, Multimedia acquisition module: Collect the teacher's teaching video and audio through high-definition cameras and audio devices, and perform preprocessing to obtain clear course content; Content segmentation module: Use natural language processing NLP and computer vision technology to identify the key information of the course content, segment the video into individual knowledge point segments, classify and store them to form a learning resource library; User profile construction module: Collect students' basic information and learning habit data, construct user profiles, and provide customized learning resources for students; Feature extraction module: Extract learning habit data from the user profile, perform feature extraction, and obtain a comprehensive feature vector that can reflect the student's learning behavior pattern; Learning status evaluation module: Use convolutional neural network (CNN) to build a learning status evaluation model, conduct a comprehensive analysis of the comprehensive feature vector, obtain a comprehensive learning status score, and adjust the playback strategy based on the score; Learning effect evaluation module: Regularly test students’ knowledge points to assess their mastery, and update user portraits based on the test results to optimize future learning resource recommendations.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the digital intelligent teaching recording and broadcasting method described in any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the digital intelligent teaching recording and broadcasting method described in any one of claims 1 to 7 are implemented.

Images