Teaching skill evaluation process video acquisition and evaluation method and evaluation system

Abstract

The invention discloses a teaching skill evaluation process video acquisition and evaluation method. The method comprises the following steps: S1, uploading lesson preparation data; s2, inputting a tentative video; s3, analyzing the teaching content; s4, analyzing language expression; s5, analyzing the classroom atmosphere; s6, analyzing a teaching mode; s7, analyzing blackboard writing; and S8, generating an analysis report. The method can promote the rapid improvement and growth of teaching skills of normal teachers, and improves the overall teaching level. The invention also discloses a teaching skill evaluation system.

Description

Technical Field

[0001] The present invention relates to digital and intelligent teaching skill evaluation technology, and in particular to a teaching skill evaluation process video acquisition and evaluation method and evaluation system. Background Art

[0002] The evaluation of a student's teaching skills is a crucial component of teacher education. For a long time, this evaluation relied primarily on manual methods. Typically, after observing a student's trial lecture and prepared materials, a teaching panel would assess the student's teaching process based on a set of criteria and subjective judgment. The overall evaluation process was then combined to determine the student's teaching skill score. Sometimes, the entire evaluation process was recorded using a camera. This method, which previously dominated the evaluation system for student teaching skills, relied on human perception and experience to assess teaching behavior.

[0003] Under the traditional evaluation model, every teaching trial by a teacher training student must be observed and scored on-site by a professional teaching evaluator. This means that for a large number of teacher training students, the evaluation process consumes a considerable amount of time and manpower. For example, organizing a teaching skills assessment for teacher training students may involve dozens or even thousands of students undergoing teaching trial evaluations, and multiple teaching evaluators will be required to conduct the evaluations on-site. Based on an average evaluation time of 30 minutes per person, this can take hundreds of hours, undoubtedly a tremendous workload.

[0004] On the other hand, manual evaluation is subject to significant subjectivity. Different evaluators, due to differences in educational background, teaching experience, personal preferences, and other factors, may have different evaluation results for the same normal school student's teaching performance. This lacks consistency and objectivity in the evaluation criteria, making it difficult to accurately and comprehensively assess normal school students' teaching skills. Furthermore, manual evaluation methods lack sufficient adaptability and flexibility in the face of ever-changing educational environments and teaching needs. With the continuous development of educational technology and the continuous updating of teaching concepts, the evaluation of normal school students' teaching skills also needs to be continuously adjusted and improved.

[0005] Furthermore, organizing professional evaluations of teacher trainees' teaching skills requires significant human, material, and time investment. This results in relatively limited opportunities for teacher trainees to receive professional evaluations and feedback on their own teaching skills. This inability to receive timely, accurate, and targeted improvement suggestions and guidance significantly hinders the rapid improvement and growth of their teaching skills. This situation leaves teacher trainees without timely and effective guidance on the path to developing their teaching skills, making it difficult for them to quickly identify and address their shortcomings, thus hindering their overall teaching performance. Summary of the Invention

[0006] The purpose of this invention is to solve the problems existing in the traditional teaching skills evaluation method of normal school students, and to propose a teaching skills evaluation process video acquisition and evaluation method and a teaching evaluation system. This method and system can promote the rapid improvement and growth of normal school students' teaching skills and improve the overall teaching level.

[0007] The technical solution for achieving the purpose of the present invention is: A method for acquiring and evaluating a teaching skill evaluation process video includes the following steps: S1. Upload lesson plan materials: Normal school students should fill in lesson plan information and submit lesson plans based on the content of their trial lectures. This information should include the student's subject category and major, teaching target audience, trial lecture theme and content, student situation analysis, teaching objectives, key points, difficulties, and teaching methods. S2. Input trial teaching videos: Input the teaching trial teaching videos of normal school students into the teaching skills assessment system; S3 analysis of teaching content: analysis of teaching content of the lesson preparation materials uploaded in step S1 and the trial video recorded in step S2; S4. Analyze language expression: Analyze the language expression of normal school students during the trial teaching; S5. Analyze classroom atmosphere: Analyze how the teacher trainees control the classroom atmosphere during the trial teaching process; S6. Analyze teaching methods: Analyze the appropriateness of the teacher training students' behavior, gaze, and clothing during the trial teaching; S7. Analyze blackboard writing: Analyze the blackboard writing of normal school students during their trial lectures; S8. Generate an analysis report: After completing the analysis in steps S3 to S7 in parallel, integrate and summarize the results of each analysis dimension. Generate a detailed analysis report based on the scores of the rationality of the teaching objectives, the relevance of the teaching content, the rationality of the content presentation, the various indicators of language expression, the control of the classroom atmosphere, and the appropriateness of the teaching methods. The analysis report will show the total score, evaluation criteria, and evaluation scores of each aspect, and will also propose specific improvement measures and suggestions for specific problems. For places where the teaching objectives are set unreasonably, the problems will be explained in detail and clear modification directions will be given; for problems in language expression, it will specifically point out which sounds are prone to errors and provide corresponding pronunciation practice methods; finally, the generated analysis report will be presented to the teacher trainees in a clear and intuitive form so that they can fully understand their performance in the trial teaching process and make targeted improvements and enhancements.

[0008] In step S2, the teaching trial video of the normal school students is recorded into the teaching skills assessment system in the following two ways: S21: Upload the completed trial teaching video: The content of the uploaded teaching trial teaching video corresponds to the lesson plan. At the same time, to ensure the accuracy of the evaluation results, the video must be of good clarity. The normal student must be within the camera's range throughout the trial teaching process, and the podium must also be within the camera's range. S22: Reserve a classroom to record a trial lecture video: The teacher trainee reserves the time and classroom for recording the trial lecture video. The reserved classroom should be equipped with a camera that supports the real-time streaming protocol. Use FFmpeg to control the camera to start recording, pause recording, continue recording, and end recording. The teacher trainee completes the recording at the reserved time and in the classroom. The requirements for recording the trial lecture video are the same as step S21.

[0009] The teaching content is analyzed in step S3 through the following four aspects: S31: Teaching Objectives: Use OCR text recognition technology to extract the text content of the lesson plan and provide it to the language model. The language model then uses the teaching objective information in the lesson plan materials to determine the rationality of the teaching objective setting and lesson plan design, and then provides a scoring evaluation and improvement suggestions for the lesson plan design. S32: Relevance: Use the audio and video processing library FFmpeg to extract audio from the trial lecture video. The extracted audio is named "student ID plus entry time plus .wav" and saved. Use the language recognition model Faster-Whisper to convert the extracted audio into speech text. The file is named "student ID plus entry time plus .txt" and saved. This allows subsequent analysis items to directly read this data without repeating the extraction or conversion operation. Use the language model to summarize and analyze the fit between the speech text of the trial lecture video and the text content of the lesson plan, score and evaluate the fit, and ask questions to the language model, allowing it to provide improvement suggestions for the relevance of the teaching content of the trial lecture video based on the fit analysis and scoring. S33: Content Presentation: Use the image content recognition library Pix2Text to analyze the blackboard or computer screen area in the trial lecture video, calculate the number and proportion of text, titles, images, tables, and formulas that appear during the trial lecture, and then score and evaluate the video based on the video length, the trial lecture topic and content, and the subject category and major of the normal school student. Based on the analysis results, ask the language big model questions to obtain improvement suggestions for the language big model on the teaching content presentation of the trial lecture video; S34: Interdisciplinary: Let the language model summarize the trial lecture video voice text, and ask the language model whether the normal students introduced the practical application of knowledge points in this subject area and in other subject areas during the trial lecture. Score the results of the analysis, and obtain improvement suggestions from the language model on the interdisciplinary aspects of the teaching content during the trial lecture.

[0010] The analysis of language expression in step S4 includes the following seven aspects: S41: Mandarin: Use speech recognition technology to identify pronunciation in the trial lecture video in real time. Compare the recognition results with the standard Mandarin pronunciation of the trial lecture video audio text. Count the number and type of pronunciation errors, and score and evaluate them accordingly. Simultaneously, let the language model analyze the causes of pronunciation errors and provide suggestions for improving pronunciation. S42: Oral Expression: Through audio analysis of trial lecture videos, determine whether the normal school students' oral expression is fluent and coherent, and whether there are any pauses or repetitions. Count the number of these occurrences and score them accordingly. The language model analyzes the factors that lead to poor oral expression and provides specific improvement suggestions. S43: Speaking Speed: Audio processing technology is used to extract speaking speed information from trial lecture videos. The average and fluctuation range of speaking speed are calculated, compared with the reasonable speaking speed range, and scores are given based on the differences. At the same time, the language model analyzes the impact of excessively fast or slow speaking speed on teaching effectiveness based on the teaching object, trial lecture topic, and content, and provides suggestions for adjusting speaking speed. S44: Intonation: Analyze the ups and downs, emphasis, and urgency of the intonation in the trial teaching video to determine whether the intonation meets the teaching context and the needs of the teaching object. Count the number of times it does not meet the requirements and provide a score. The language model analyzes the problems caused by inappropriate intonation and provides suggestions for optimizing the intonation. S45: Positive Language: Analyze the audio text in the trial lecture video, count the frequency of positive language usage, compare it with the words in the ideal positive language library, and score and evaluate it. The language model analyzes the impact of positive language on the classroom atmosphere and student enthusiasm, and proposes techniques and methods for using positive language. S46: Negative Language: Similarly, the audio text in the trial lecture video is analyzed to identify the occurrence of negative language, count the number of occurrences, and score the occurrences. The language model analyzes the negative impact of negative language on students' psychology and classroom atmosphere, and provides strategies to avoid the use of negative language. S47: Ideology: Analyze whether there is ideological language in the audio text of the trial lecture video, make qualitative judgments and score evaluations, let the language model analyze the consequences of ideological issues, and propose requirements and methods for adhering to correct ideology in teaching.

[0011] The analysis of classroom atmosphere in step S5 includes the following three aspects: S51: Teaching Expressions: Through frame-by-frame analysis of trial lecture videos, using the facial expression recognition library DeepFace and the image processing library OpenCV, we identify the facial expressions of normal school students during teaching. We then count the frequency and duration of different expressions and score them based on their compatibility with the teaching content. At the same time, we use a large language model to analyze the emotional information conveyed by different teaching expressions and their impact on students' emotions and learning motivation. Furthermore, we propose suggestions for optimizing teaching expressions based on their frequency and duration. S52: Classroom Sound: Through audio spectrum analysis of trial lecture videos, classroom noise is carefully detected, and the characteristics of the noise frequency range, sound intensity, and duration of occurrence are analyzed. If the noise is concentrated in the low frequency band, the sound intensity is low, and the duration of occurrence is low, then the impact of the noise on the overall classroom atmosphere is small and the noise is within the acceptable range. If the noise is concentrated in the high frequency band, is loud, or lasts for a long time, it will affect the classroom atmosphere. Score the noise according to the proportion of noise, and let the language model analyze the impact of classroom sound on students' attention and learning experience, so as to provide suggestions for improving classroom sound. S53: Classroom Interaction: Analyze the interaction between teacher trainees and students in the trial teaching videos from the audio text of the trial teaching videos, including the frequency of questions asked, students’ answers, and presentations. Count the positive level of the interaction and give a score evaluation. Let the language model analyze the impact of different types of classroom interaction on student participation and learning outcomes, and then put forward suggestions to promote classroom interaction.

[0012] The analysis of the teaching method in step S6 is specifically performed from the following three aspects: S61: Body Movement: By analyzing students' body movements during the teaching process through frame-by-frame analysis of trial teaching videos, the frequency and duration of different body movements are counted to determine whether the body movements are natural, appropriate, and helpful for teaching expression. The body movements are scored and evaluated based on their compatibility with the teaching content. The language model also analyzes the emotions and information conveyed by different body movements, as well as their impact on students' learning status, and provides suggestions for optimizing body movements. S62: Viewpoint Distribution: Analyze the teacher trainee's gaze focus during the trial lecture video frame by frame, that is, the duration and frequency of the teacher trainee's gaze in different areas of the classroom. This determines whether the teacher trainee can make effective eye contact with all students, avoiding staring at a single point for too long or frequently scanning and ignoring some students. The viewpoint distribution is statistically analyzed to determine whether it is even and reasonable, and a score is assigned based on the degree of evenness and rationality. The language model is then used to analyze the differences in student engagement and learning outcomes caused by improper viewpoint distribution, and to provide improvement suggestions. S63: Appearance: Use clothing recognition technology to analyze the appearance of normal school students in trial teaching videos, determine whether the clothing meets the requirements of the teaching environment, count the number of problems with appearance, and give scores and evaluations. Let the language model analyze the impact of appearance on the classroom, and then propose requirements and methods for maintaining a good appearance.

[0013] The analysis of the blackboard writing in step S7 is carried out from the following two aspects: S71: Handwriting neatness: Frames are taken from the trial lecture video at intervals of 30 seconds to one minute to analyze each font used on the blackboard. This includes whether the font size is appropriate, the strokes are clear, and the font shape is standardized. The percentage of fonts that do not meet the requirements is counted and scored based on the percentage of fonts. The language model also analyzes the impact of inappropriate fonts on students' reading and comprehension of the blackboard content, and provides suggestions for optimizing the font. S72: Layout aesthetics: Similarly, take frames from the trial lecture video at intervals of 30 seconds to one minute to analyze the blackboard layout, select the text and chart areas, calculate the average tilt of each frame, and score and evaluate based on the average tilt. Then, let the language model analyze the impact of improper layout on students' learning process, and then put forward suggestions for optimizing the layout.

[0014] A teaching skills assessment system is used to assess the teaching skills of normal school students through the above-mentioned teaching skills assessment process video acquisition and assessment method. The system includes three modules: application for assessment, analysis of teaching skills, and generation of analysis reports. The application assessment module is divided into a video assessment submission module and a classroom reservation assessment module. The video assessment submission module is used to upload lesson preparation materials and completed trial teaching videos to provide data for subsequent teaching skills analysis; the classroom reservation assessment module is used for normal school students to upload lesson preparation materials, make classroom reservations, and complete the recording of trial teaching videos in the reserved classroom, providing more realistic scene data for teaching skills assessment. The teaching skills analysis module consists of five sub-modules: analyzing teaching content, analyzing language expression, analyzing classroom atmosphere, analyzing teaching methods, and analyzing blackboard writing. The teaching skills analysis module will conduct a detailed analysis of the incoming lesson preparation materials and trial teaching videos from the dimensions of these five sub-modules; The generated analysis report module integrates the analysis results of each dimension in the teaching skills module and presents them in a graphical visual form. That is, the generated analysis report module will display in detail the evaluation criteria, specific scores, corresponding grades, professional comments and targeted improvement suggestions for a total of 19 items under the five dimensions of the teaching skills module, and display key teaching data such as speaking speed, intonation, classroom language, and teaching expressions through intuitive statistical charts, so that normal school students can clearly and intuitively understand their own teaching skills.

[0015] The key points of this technical solution are: (1) There are two ways to record the trial lecture video: one is to directly upload the completed trial lecture video, and the other is to reserve a classroom for recording. The camera equipment installed in the classroom has the support function of Real-Time Streaming Protocol (RTSP), and FFmpeg is used to control the camera equipment to complete the recording process smoothly.

[0016] (2) Use a variety of advanced technologies to conduct multimodal analysis of the trial teaching videos, such as analyzing the teaching content presentation by taking frames from the video, analyzing the Mandarin language of the video audio, and analyzing the positive and negative terms by converting the audio into text, etc., to ensure the accurate analysis of the five dimensions and nineteen items during the normal students' trial teaching process.

[0017] (3) The use of language big models, which play a key role in summarizing, judging, analyzing and making suggestions. From teaching objectives to classroom atmosphere, all aspects rely on the intelligent processing and improvement suggestions of language big models.

[0018] The beneficial effects of this technical solution are: In terms of teaching efficiency, this technology significantly improves the speed and efficiency of evaluation compared to traditional manual evaluation methods. Traditionally, conducting a teaching skills assessment for teacher trainees would require significant time and manpower. However, this technology's automated intelligent analysis allows the evaluation of numerous teacher trainees' trial lectures to be completed in a fraction of the time.

[0019] In terms of evaluation accuracy, this technical solution overcomes the subjective differences in manual evaluation by leveraging multiple advanced technologies and precise analysis of large language models. Through multimodal analysis of trial lecture videos, such as analyzing the presentation of teaching content through frame-by-frame analysis and Mandarin analysis of audio and video, it comprehensively and accurately captures various information about the teaching process of normal school students, thus avoiding the subjective differences in manual evaluation caused by evaluators.

[0020] This technical solution also plays a positive role in promoting the growth and development of teacher trainees themselves. The generated detailed analysis report provides comprehensive and specific feedback to teacher trainees, allowing them to clearly understand their strengths and weaknesses in the teaching process. For example, the report not only points out the unreasonable aspects of the lesson plan design, but also explains the problems in detail and provides clear optimization suggestions. For problems in language expression, such as a large number of pronunciation errors, it will specifically point out which sounds are prone to errors and provide corresponding pronunciation practice suggestions. This enables teacher trainees to improve and enhance their teaching skills in a targeted manner, quickly identify their own shortcomings and correct them, thereby effectively promoting the rapid improvement and growth of their teaching skills.

[0021] In terms of overall improvement in education and teaching, this technical solution has injected new vitality into the development of teacher education. By evaluating and providing feedback on the teaching skills of a large number of teacher trainees, it can encourage teacher education to focus more on cultivating their practical teaching abilities and improve overall teaching standards. For example, based on the analysis results of this technical solution, teacher training courses can adjust teaching content and methods in a targeted manner, strengthening training in teaching content presentation, language expression, and classroom atmosphere control, thereby cultivating more outstanding teachers who meet the needs of modern education.

[0022] In summary, the teacher training students' teaching skills assessment system and method of this technical solution have achieved excellent results in teaching efficiency, evaluation accuracy, teacher training students' growth, and overall improvement of education and teaching, making important contributions to the development of teacher education. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] Figure 1 It is a module diagram of the system in the embodiment; Figure 2 Schematic diagram of the process in the embodiment; Figure 3 This is an example diagram of the shooting range in the embodiment. DETAILED DESCRIPTION

[0024] The present invention will be further described below with reference to the accompanying drawings and embodiments, but the present invention is not limited thereto. Example

[0025] Reference Figure 2 A teaching skill evaluation process video acquisition and evaluation method includes the following steps: S1. Upload lesson plan materials: Normal school students should fill in lesson plan information and submit lesson plans based on the content of their trial lectures. This information should include the student's subject category and major, teaching target audience, trial lecture theme and content, student situation analysis, teaching objectives, key points, difficulties, and teaching methods. S2. Input the trial teaching video: Input the normal school students' teaching trial teaching video into the teaching skills assessment system in the following two ways: S21: Upload the completed trial teaching video: The content of the uploaded teaching trial teaching video should correspond to the lesson plan. At the same time, to ensure the accuracy of the evaluation results, the video should be clear. The normal students should be within the camera range during the trial teaching process, and the podium should also be within the camera range. Figure 3 As shown; S22: Book a classroom to record a trial lecture video: The teacher trainee books a time and classroom for recording a trial lecture video. The reserved classroom should be equipped with a camera that supports the Real-Time Streaming Protocol (RTSP). FFmpeg is used to control the camera to start, pause, continue, and end recording. The teacher trainee completes the recording in the reserved classroom at the reserved time. The requirements for recording the trial lecture video are the same as those in step S21. S3. Analyze teaching content: Analyze the teaching content of the lesson preparation materials uploaded in step S1 and the trial teaching video recorded in step S2, and analyze the following four aspects: S31: Teaching Objectives: Use OCR text recognition technology to extract the text content of the lesson plan and provide it to the language model. The language model then uses the teaching objective information in the lesson plan materials to determine the rationality of the teaching objective setting and lesson plan design, and then provides a scoring evaluation and improvement suggestions for the lesson plan design. S32: Relevance: Use the audio and video processing library FFmpeg to extract audio from the trial lecture video. The extracted audio is named "student ID plus entry time plus .wav" and saved. Use the language recognition model Faster-Whisper to convert the extracted audio into speech text. The file is named "student ID plus entry time plus .txt" and saved. This allows subsequent analysis items to directly read this data without repeating the extraction or conversion operation. Use the language model to summarize and analyze the fit between the speech text of the trial lecture video and the text content of the lesson plan, score and evaluate the fit, and ask questions to the language model, allowing it to provide improvement suggestions for the relevance of the teaching content of the trial lecture video based on the fit analysis and scoring. S33: Content Presentation: Use the image content recognition library Pix2Text to analyze the blackboard or computer screen area in the trial lecture video, calculate the number and proportion of text, titles, images, tables, and formulas that appear during the trial lecture, and then score and evaluate the video based on the video length, the trial lecture topic and content, and the subject category and major of the normal school student. Based on the analysis results, ask the language big model questions to obtain improvement suggestions for the language big model on the teaching content presentation of the trial lecture video; S34: Interdisciplinary: Have the language model summarize the audio and video transcripts of the trial lectures and ask the model questions about whether the teacher trainees introduced the practical application of the knowledge points in their own subject area and in other subject areas during the trial lectures. The model will be scored based on the analysis results and receive suggestions for improvement on the interdisciplinary aspects of the teaching content during the trial lectures from the language model. S4. Language expression analysis: Analyze the language expression of normal school students during the trial teaching, including the following seven aspects: S41: Mandarin: Use speech recognition technology to identify pronunciation in the trial lecture video in real time. Compare the recognition results with the standard Mandarin pronunciation of the trial lecture video audio text. Count the number and types of pronunciation errors, such as confusion between flat and retroflex consonants and inability to distinguish between front and back nasal sounds. Score and evaluate the errors accordingly. Simultaneously, use the language model to analyze the causes of pronunciation errors and provide suggestions for improving pronunciation. S42: Oral Expression: Through audio analysis of trial lecture videos, determine whether the normal school students' oral expression is fluent and coherent, and whether there are any pauses or repetitions. Count the number of these occurrences and score them accordingly. The language model also analyzes factors that lead to poor oral expression, such as unclear thinking logic and nervousness, and provides specific improvement suggestions. S43: Speaking Speed: Audio processing technology is used to extract speaking speed information from trial lecture videos. The average and fluctuation range of speaking speed are calculated, compared with the reasonable speaking speed range, and scores are given based on the differences. At the same time, the language model analyzes the impact of excessively fast or slow speaking speed on teaching effectiveness based on the teaching object, trial lecture topic, and content, and provides suggestions for adjusting speaking speed. S44: Intonation: Analyze the ups and downs, emphasis, and urgency of the intonation in the trial teaching video to determine whether the intonation meets the teaching context and the needs of the teaching object. Count the number of times it does not meet the requirements and provide a score. The language model analyzes the problems caused by inappropriate intonation, such as affecting students' attention and inappropriate expression of emotions, and provides suggestions for optimizing intonation. S45: Positive Language: Analyze the audio text in the trial lecture video, count the frequency of positive language usage, compare it with the ideal positive language library, such as words such as "awesome," "very good," and "amazing," and provide a score evaluation. The language model analyzes the impact of positive language on the classroom atmosphere and student motivation, and proposes techniques and methods for using positive language. S46: Negative Language: Similarly, the audio text in the trial lecture video is analyzed to identify the occurrence of negative language, count the number of occurrences, and score the occurrences. The language model analyzes the negative impact of negative language on students' psychology and classroom atmosphere, and provides strategies to avoid the use of negative language. S47: Ideology: Analyze the audio text of the trial lecture video to see if there is any ideological language, such as whether it guides incorrect values and whether it conforms to the mainstream ideology of the country and society. Make qualitative judgments and score evaluations, and let the language model analyze the consequences of ideological problems and propose requirements and methods for adhering to correct ideology in teaching. S5. Analyze the classroom atmosphere: Analyze the teacher trainees' control of the classroom atmosphere during the trial teaching, including the following three aspects: S51: Teaching Expressions: Through frame-by-frame analysis of trial lecture videos, the facial expression recognition library DeepFace and the image processing library OpenCV are used to determine the facial expressions of normal school students during the teaching process, such as happiness, seriousness, surprise, sadness, anger, fear, etc. The frequency and duration of different expressions are counted, and scores are given based on the adaptability of the expressions to the teaching content. At the same time, the language model is used to analyze the emotional information conveyed by different teaching expressions and their impact on students' emotions and learning enthusiasm. Then, based on the frequency and duration of expressions, suggestions for optimizing teaching expressions are made. For example, when explaining key content, appropriate smiles can be added to enhance affinity, and when guiding students to think, seriousness can be maintained to attract attention. S52: Classroom Sound: Through audio spectrum analysis of trial lecture videos, classroom noise is carefully detected, and the characteristics of the noise frequency range, sound intensity, and duration of occurrence are analyzed. If the noise is concentrated in the low frequency band, the sound intensity is low, and the duration of occurrence is low, then the impact of the noise on the overall classroom atmosphere is small and the noise is within the acceptable range. If the noise is concentrated in the high frequency band, is loud, or lasts for a long time, it will affect the classroom atmosphere. Score the noise according to the proportion of noise, and let the language model analyze the impact of classroom sound on students' attention and learning experience, so as to provide suggestions for improving classroom sound. S53: Classroom Interaction: Analyze the interaction between teacher trainees and students in the trial lecture videos from the audio transcripts, including the frequency of questions asked, students’ responses, and presentations. Count the level of positivity in these interactions and provide a scoring evaluation. The language model analyzes the impact of different types of classroom interaction on student engagement and learning outcomes, and then proposes suggestions for promoting classroom interaction. S6. Analyze teaching methods: Analyze the appropriateness of the teacher training students' behavior, gaze, and clothing during the trial teaching process, specifically focusing on the following three aspects: S61: Body Movement: Frames from trial lecture videos are analyzed to analyze students' body movements during the teaching process, such as sitting, walking, standing, writing on the blackboard, and gesturing. The frequency and duration of different body movements are counted to determine whether the body movements are natural, appropriate, and helpful for teaching expression. The body movements are scored and evaluated based on their compatibility with the teaching content. The language model also analyzes the emotions and information conveyed by different body movements, as well as their impact on students' learning status, and provides suggestions for optimizing body movements. S62: Viewpoint Distribution: Analyze the teacher trainee's gaze focus during the trial lecture video frame by frame, that is, the duration and frequency of the teacher trainee's gaze in different areas of the classroom. This determines whether the teacher trainee can make effective eye contact with all students, avoiding staring at a single point for too long or frequently scanning and ignoring some students. The viewpoint distribution is statistically analyzed to determine whether it is even and reasonable, and a score is assigned based on the degree of evenness and rationality. The language model is then used to analyze the differences in student engagement and learning outcomes caused by improper viewpoint distribution, and to provide improvement suggestions. S63: Appearance: Use clothing recognition technology to analyze the appearance of normal school students in trial teaching videos to determine whether their clothing meets the requirements of teaching environments, such as not wearing slippers or shorts. Count the number of appearance problems and score them. Use the language model to analyze the impact of appearance on the classroom and propose requirements and methods for maintaining a good appearance. S7. Blackboard Writing Analysis: Analyze the blackboard writing of normal school students during their trial lectures from the following two aspects: S71: Handwriting neatness: Frames are taken from the trial lecture video at intervals of 30 seconds to one minute to analyze each font used on the blackboard. This includes whether the font size is appropriate, the strokes are clear, and the font shape is standardized. The percentage of fonts that do not meet the requirements is counted and scored based on the percentage of fonts. The language model also analyzes the impact of inappropriate fonts on students' reading and comprehension of the blackboard content, and provides suggestions for optimizing the font. S72: Layout aesthetics: Similarly, frames were taken from the trial lecture video at intervals of 30 seconds to one minute to analyze the layout of the blackboard. The text and chart areas were framed, and the average tilt of each frame was calculated. Scoring was performed based on the average tilt. The language model was then used to analyze the impact of inappropriate layout on students' learning process and to provide suggestions for optimizing the layout. S8. Generate an analysis report: After completing the analysis in steps S3 to S7 in parallel, integrate and summarize the results of each analysis dimension. Generate a detailed analysis report based on the scores of the rationality of the teaching objectives, the relevance of the teaching content, the rationality of the content presentation, the various indicators of language expression, the control of the classroom atmosphere, and the appropriateness of the teaching methods. The analysis report will show the total score, evaluation criteria, and evaluation scores of each aspect, and will also propose specific improvement measures and suggestions for specific problems. For places where the teaching objectives are set unreasonably, the problems will be explained in detail and clear modification directions will be given. For problems in language expression, such as a large number of pronunciation errors, it will be specifically pointed out which sounds are prone to errors, and corresponding pronunciation practice methods will be provided. Finally, the generated analysis report will be presented to the teacher trainees in a clear and intuitive form so that they can fully understand their performance in the trial teaching process and make targeted improvements and enhancements.

[0026] like Figure 2 As shown, a teaching skills assessment system is used to assess the teaching skills of normal school students through the above-mentioned teaching skills assessment process video acquisition and assessment method. The system includes three modules: application assessment, teaching skills analysis, and analysis report generation. The application assessment module is divided into a video assessment submission module and a classroom reservation assessment module. The video assessment submission module is used to upload lesson preparation materials and completed trial teaching videos to provide data for subsequent teaching skills analysis; the classroom reservation assessment module is used for normal school students to upload lesson preparation materials, make classroom reservations, and complete the recording of trial teaching videos in the reserved classroom, providing more realistic scene data for teaching skills assessment. The teaching skills analysis module consists of five sub-modules: analyzing teaching content, analyzing language expression, analyzing classroom atmosphere, analyzing teaching methods, and analyzing blackboard writing. The teaching skills analysis module will conduct a detailed analysis of the incoming lesson preparation materials and trial teaching videos from the dimensions of these five sub-modules; The generated analysis report module integrates the analysis results of each dimension in the teaching skills module and presents them in a graphical visual form. That is, the generated analysis report module will display in detail the evaluation criteria, specific scores, corresponding grades, professional comments and targeted improvement suggestions for a total of 19 items under the five dimensions of the teaching skills module, and display key teaching data such as speaking speed, intonation, classroom language, and teaching expressions through intuitive statistical charts, so that normal school students can clearly and intuitively understand their own teaching skills.

Claims

1. A method for acquiring and evaluating a teaching skill evaluation process video, characterized in that: The steps include: S1. Upload lesson plan materials: Normal school students should fill in lesson plan information and submit lesson plans based on the content of their trial lectures. This information should include the student's subject category and major, teaching target audience, trial lecture theme and content, student situation analysis, teaching objectives, key points, difficulties, and teaching methods. S2. Input trial teaching videos: Input the teaching trial teaching videos of normal school students into the teaching skills assessment system; S3 analysis of teaching content: analysis of teaching content of the lesson preparation materials uploaded in step S1 and the trial video recorded in step S2; S4. Analyze language expression: Analyze the language expression of normal school students during the trial teaching; S5. Analyze classroom atmosphere: Analyze how the teacher trainees control the classroom atmosphere during the trial teaching process; S6. Analyze teaching methods: Analyze the appropriateness of the teacher training students' behavior, gaze, and clothing during the trial teaching; S7. Analyze blackboard writing: Analyze the blackboard writing of normal school students during their trial lectures; S8. Generate an analysis report: After completing the analysis in steps S3 to S7 in parallel, integrate and summarize the results of each analysis dimension. Generate a detailed analysis report based on the scores of the rationality of the teaching objectives, the relevance of the teaching content, the rationality of the content presentation, the various indicators of language expression, the control of the classroom atmosphere, and the appropriateness of the teaching methods. The analysis report will show the total score, evaluation criteria, and evaluation scores of each aspect, and will also propose specific improvement measures and suggestions for specific problems. For places where the teaching objectives are set unreasonably, the problem will be explained in detail and a clear direction for modification will be given; for problems in language expression, it will specifically point out which sounds are prone to errors and provide corresponding pronunciation practice methods; finally, the generated analysis report will be presented to the teacher trainees in a clear and intuitive form so that they can fully understand their performance in the trial teaching process and make targeted improvements and enhancements.

2. The teaching skill evaluation process video acquisition and evaluation method according to claim 1 is characterized in that: In step S2, the teaching trial video of the normal school students is recorded into the teaching skills assessment system in the following two ways: S21: Upload the completed trial teaching video: The content of the uploaded teaching trial teaching video corresponds to the lesson plan. At the same time, to ensure the accuracy of the evaluation results, the video must be of good clarity. The normal student must be within the camera's range throughout the trial teaching process, and the podium must also be within the camera's range. S22: Reserve a classroom to record a trial lecture video: The teacher trainee reserves the time and classroom for recording the trial lecture video. The reserved classroom should be equipped with a camera that supports the real-time streaming protocol. Use FFmpeg to control the camera to start recording, pause recording, continue recording, and end recording. The teacher trainee completes the recording at the reserved time and in the classroom. The requirements for recording the trial lecture video are the same as step S21.

3. The teaching skill evaluation process video acquisition and evaluation method according to claim 1 is characterized in that: The teaching content is analyzed in step S3 through the following four aspects: S31: Teaching Objectives: Use OCR text recognition technology to extract the text content of the lesson plan and provide it to the language model. The language model then uses the teaching objective information in the lesson plan materials to determine the rationality of the teaching objective setting and lesson plan design, and then provides a scoring evaluation and improvement suggestions for the lesson plan design. S32: Relevance: Use the audio and video processing library FFmpeg to extract audio from the trial lecture video. The extracted audio is named "student ID plus entry time plus .wav" and saved. The language recognition model Faster-Whisper converts the extracted audio into speech text, naming it "student ID plus entry time plus .txt" and saving it. This allows subsequent analysis items to directly read this data without repeating the extraction or conversion operation. Use the language model to summarize and analyze the fit between the speech text of the trial lecture video and the text content of the lesson plan, scoring and evaluating the fit. Questions are then posed to the language model, allowing it to provide improvement suggestions for the relevance of the teaching content of the trial lecture video based on the fit analysis and scoring. S33: Content Presentation: Use the image content recognition library Pix2Text to analyze the blackboard or computer screen area in the trial lecture video, calculate the number and proportion of text, titles, images, tables, and formulas that appear during the trial lecture, and then score and evaluate the video based on the video length, the trial lecture topic and content, and the subject category and major of the normal school student. Based on the analysis results, ask the language big model questions to obtain improvement suggestions for the language big model on the teaching content presentation of the trial lecture video; S34: Interdisciplinary: Let the language model summarize the trial lecture video voice text, and ask the language model whether the normal students introduced the practical application of knowledge points in this subject area and in other subject areas during the trial lecture. Score the results of the analysis, and obtain improvement suggestions from the language model on the interdisciplinary aspects of the teaching content during the trial lecture.

4. The teaching skill evaluation process video acquisition and evaluation method according to claim 1 is characterized in that: The analysis of language expression in step S4 includes the following seven aspects: S41: Mandarin: Use speech recognition technology to identify pronunciation in the trial lecture video in real time. Compare the recognition results with the standard Mandarin pronunciation of the trial lecture video audio text. Count the number and type of pronunciation errors, and score and evaluate them accordingly. Simultaneously, let the language model analyze the causes of pronunciation errors and provide suggestions for improving pronunciation. S42: Oral Expression: Through audio analysis of trial lecture videos, determine whether the normal school students' oral expression is fluent and coherent, and whether there are any pauses or repetitions. Count the number of these occurrences and score them accordingly. The language model analyzes the factors that lead to poor oral expression and provides specific improvement suggestions. S43: Speaking Speed: Audio processing technology is used to extract speaking speed information from trial lecture videos. The average and fluctuation range of speaking speed are calculated, compared with the reasonable speaking speed range, and scores are given based on the differences. At the same time, the language model analyzes the impact of excessively fast or slow speaking speed on teaching effectiveness based on the teaching object, trial lecture topic, and content, and provides suggestions for adjusting speaking speed. S44: Intonation: Analyze the ups and downs, emphasis, and urgency of the intonation in the trial teaching video to determine whether the intonation meets the teaching context and the needs of the teaching object. Count the number of times it does not meet the requirements and provide a score. The language model analyzes the problems caused by inappropriate intonation and provides suggestions for optimizing the intonation. S45: Positive Language: Analyze the audio text in the trial lecture video, count the frequency of positive language usage, compare it with the words in the ideal positive language library, and score and evaluate it. The language model analyzes the impact of positive language on the classroom atmosphere and student enthusiasm, and proposes techniques and methods for using positive language. S46: Negative Language: Similarly, the audio text in the trial lecture video is analyzed to identify the occurrence of negative language, count the number of occurrences, and score the occurrences. The language model analyzes the negative impact of negative language on students' psychology and classroom atmosphere, and provides strategies to avoid the use of negative language. S47: Ideology: Analyze whether there is ideological language in the audio text of the trial lecture video, make qualitative judgments and score evaluations, let the language model analyze the consequences of ideological issues, and propose requirements and methods for adhering to correct ideology in teaching.

5. The teaching skill evaluation process video acquisition and evaluation method according to claim 1 is characterized in that: The analysis of classroom atmosphere in step S5 includes the following three aspects: S51: Teaching Expressions: Through frame-by-frame analysis of trial lecture videos, using the facial expression recognition library DeepFace and the image processing library OpenCV, we identify the facial expressions of normal school students during teaching. We then count the frequency and duration of different expressions and score them based on their compatibility with the teaching content. Simultaneously, we use a large language model to analyze the emotional information conveyed by different teaching expressions and their impact on students' emotions and learning motivation. Furthermore, we propose suggestions for optimizing teaching expressions based on their frequency and duration. S52: Classroom Sound: Through audio spectrum analysis of trial lecture videos, classroom noise is carefully detected, and the characteristics of the noise frequency range, sound intensity, and duration of occurrence are analyzed. If the noise is concentrated in the low frequency band, the sound intensity is low, and the duration of occurrence is low, then the impact of the noise on the overall classroom atmosphere is small and the noise is within the acceptable range. If the noise is concentrated in the high frequency band, is loud, or lasts for a long time, it will affect the classroom atmosphere. Score the noise according to the proportion of noise, and let the language model analyze the impact of classroom sound on students' attention and learning experience, so as to provide suggestions for improving classroom sound. S53: Classroom Interaction: Analyze the interaction between teacher trainees and students in the trial teaching videos from the audio text of the trial teaching videos, including the frequency of questions asked, students’ answers, and presentations. Count the positive level of the interaction and give a score evaluation. Let the language model analyze the impact of different types of classroom interaction on student participation and learning outcomes, and then put forward suggestions to promote classroom interaction.

6. The teaching skill evaluation process video acquisition and evaluation method according to claim 1 is characterized in that: The analysis of the teaching method in step S6 is specifically performed from the following three aspects: S61: Body Movement: By analyzing students' body movements during the teaching process through frame-by-frame analysis of trial teaching videos, the frequency and duration of different body movements are counted to determine whether the body movements are natural, appropriate, and helpful for teaching expression. The body movements are scored and evaluated based on their compatibility with the teaching content. The language model also analyzes the emotions and information conveyed by different body movements, as well as their impact on students' learning status, and provides suggestions for optimizing body movements. S62: Viewpoint Distribution: Analyze the teacher trainee's gaze focus during the trial lecture video frame by frame, that is, the duration and frequency of the teacher trainee's gaze in different areas of the classroom. This determines whether the teacher trainee can make effective eye contact with all students, avoiding staring at a single point for too long or frequently scanning and ignoring some students. The viewpoint distribution is then statistically analyzed to determine whether it is even and reasonable, and a score is assigned based on the degree of evenness and rationality. The language model is then used to analyze differences in student engagement and learning outcomes caused by improper viewpoint distribution, and to provide suggestions for improvement. S63: Appearance: Use clothing recognition technology to analyze the appearance of normal school students in trial teaching videos, determine whether the clothing meets the requirements of the teaching environment, count the number of problems with appearance, and give scores and evaluations. Let the language model analyze the impact of appearance on the classroom, and then propose requirements and methods for maintaining a good appearance.

7. The teaching skill evaluation process video acquisition and evaluation method according to claim 1 is characterized in that: The analysis of the blackboard writing in step S7 is carried out from the following two aspects: S71: Handwriting neatness: Frames are taken from the trial lecture video at intervals of 30 seconds to one minute to analyze each font used on the blackboard. This includes whether the font size is appropriate, the strokes are clear, and the font shape is standardized. The percentage of fonts that do not meet the requirements is counted and scored based on the percentage of fonts. The language model also analyzes the impact of inappropriate fonts on students' reading and comprehension of the blackboard content, and provides suggestions for optimizing the font. S72: Layout aesthetics: Similarly, take frames from the trial lecture video at intervals of 30 seconds to one minute to analyze the blackboard layout, select the text and chart areas, calculate the average tilt of each frame, and score and evaluate based on the average tilt. Then, let the language model analyze the impact of improper layout on students' learning process, and then put forward suggestions for optimizing the layout.

8. A teaching skills assessment system, characterized in that: The system is used for the teaching skills evaluation process video acquisition and evaluation method according to claim 1, and the system includes three modules: application evaluation, teaching skills analysis, and analysis report generation. The application assessment module is divided into a video assessment submission module and a classroom reservation assessment module. The video assessment submission module is used to upload lesson preparation materials and completed trial teaching videos to provide data for subsequent teaching skills analysis; the classroom reservation assessment module is used for normal school students to upload lesson preparation materials, make classroom reservations, and complete the recording of trial teaching videos in the reserved classroom, providing more realistic scene data for teaching skills assessment. The teaching skills analysis module consists of five sub-modules: analyzing teaching content, analyzing language expression, analyzing classroom atmosphere, analyzing teaching methods, and analyzing blackboard writing. The teaching skills analysis module will conduct a detailed analysis of the incoming lesson preparation materials and trial teaching videos from the dimensions of these five sub-modules; The generated analysis report module integrates the analysis results of each dimension in the teaching skills module and presents them in a graphical visual form. That is, the generated analysis report module will display in detail the evaluation criteria, specific scores, corresponding grades, professional comments and targeted improvement suggestions for a total of 19 items under the five dimensions of the teaching skills module, and display key teaching data such as speaking speed, intonation, classroom language, and teaching expressions through intuitive statistical charts, so that normal school students can clearly and intuitively understand their own teaching skills.

Images