A teaching quality evaluation method

Abstract

The application discloses a teaching quality evaluation method, and relates to the technical field of teaching quality evaluation, and comprises the following steps: S1, fixed information of the current class is acquired, wherein the fixed information at least includes a subject category, a current class sequence, a time length from the current day to the nearest rest day, and a time length from the current day to the nearest examination day; a determination factor influence coefficient is calculated through a pre-stored psychological cycle influence model; S2, a teaching video under a classroom scene of the current class is collected, and students in a video frame are detected in real time based on a YOLO series target detection algorithm, the application realizes real-time monitoring of a classroom teaching process, can timely find problems existing in teaching, provides timely feedback for teaching improvement, and integrates psychological cycle factors such as class fatigue, weekend expectation and examination pressure and environmental factors such as a sudden event outside the window into an evaluation model, realizes comprehensive consideration of various objective factors influencing teaching effect, and makes the evaluation result more comprehensive and objective.

Description

Technical Field

[0001] This invention relates to the field of teaching quality assessment technology, and in particular to a teaching quality assessment method. Background Technology

[0002] Teaching quality assessment is a core component of educational management. Scientific and objective assessment results are of significant reference value for teacher improvement, school management decisions, and educational policy formulation. Traditional teaching quality assessment methods primarily rely on student exam scores and student questionnaires. These methods have the following problems: a lag – assessments based on student exam scores typically only yield results weeks or even months after the course ends, failing to provide real-time monitoring and timely feedback of the classroom teaching process; and subjectivity – student questionnaires are easily influenced by factors such as teacher-student relationships, student emotional states, and the appropriateness of questionnaire design, leading to significant subjective bias in the assessment results.

[0003] With the rapid development of computer vision technology, classroom teaching quality assessment methods based on object detection algorithms have gradually become a research hotspot. Existing technologies include methods that collect teaching videos in classroom settings and use YOLO series object detection algorithms to perform real-time detection and status analysis of students in the video frames. Specifically, these methods determine students' attention levels during classroom learning by detecting visual features such as head posture, eye state, and facial orientation, and then statistically analyze the proportion of students attentive in class, using this as a core indicator for evaluating the quality of classroom teaching. While these visual object detection-based teaching quality assessment methods address the issues of lag and subjectivity inherent in traditional methods, they still suffer from the following technical limitations: The assessment dimensions are too simplistic, and students' classroom focus is not constant but significantly influenced by various physiological and psychological cyclical factors. For example, students' physiological arousal levels vary across different periods each day. Students are generally at a low point in the first period of the morning and the first period after lunch break, making it difficult for them to concentrate; in the last period, students are at their peak of fatigue, naturally leading to a decline in focus. Furthermore, the proximity to the weekend affects students' psychological expectations. On Friday afternoons, students generally experience the "weekend expectation effect," making them easily distracted; while on Monday mornings, they suffer from "weekend syndrome," making it difficult for them to quickly get back into a learning mindset. Disruptive Factors from Unexpected Events: Classroom teaching takes place in an open environment, inevitably subject to disruptions from unexpected external events. For example, sudden snowfall, rain, thunderstorms, strong winds, and other weather changes can attract students' attention; sudden high-decibel noises outside the window (such as construction noise, vehicle horns, cheers from extracurricular activities, etc.) can interrupt students' thought processes. These unexpected events can significantly affect students' concentration, leading to an abnormal decline in student focus within a specific time period. Existing assessment methods fail to identify and eliminate the influence of these environmental distractions, simply attributing the decline in student focus caused by environmental factors to the teacher's teaching ability, resulting in biased assessment results. The influence of special teaching scenarios is not considered: In special teaching scenarios such as open classes, observation classes, and demonstration classes, the classroom teaching ecology is significantly altered due to the presence of observers (other teachers, curriculum researchers, school leaders, etc.). On the one hand, teachers may demonstrate more thorough preparation and a more energetic state than in regular teaching; on the other hand, students may exhibit higher discipline and concentration under the pressure of strangers observing their classes. This "open class effect" leads to significant differences between teaching performance in special scenarios and regular teaching. Simply comparing the evaluation results of such special scenarios with regular teaching will result in distorted evaluation results.

[0004] Therefore, a teaching quality assessment method is proposed to address the above problems. Summary of the Invention

[0005] The purpose of this invention is to solve the problems in the prior art by proposing a teaching quality assessment method.

[0006] A teaching quality assessment method includes the following steps: S1. Obtain the fixed information for this lesson, which includes at least the subject category, the order of lessons on the day, the time remaining until the nearest rest day, and the time remaining until the nearest exam day; based on the order of lessons on the day, the time remaining until the nearest rest day, and the time remaining until the nearest exam day, calculate and determine the factor influence coefficient using a pre-stored psychological cycle influence model. ; S2. Collect the teaching video of this lesson's classroom setting. Using the YOLO series of object detection algorithms, perform real-time detection of students in the video frames. At preset time intervals, extract the proportion of students not paying attention in the current frame relative to the total number of students in the class. Integrate or average all proportions across all time dimensions of this lesson to obtain the attentiveness coefficient for this lesson. ; S3. Collect video footage of the environment outside the classroom window during this lesson. Based on video frame difference and abnormal event recognition algorithms, extract sudden events outside the window. These sudden events include at least sudden snowfall, sudden rain, sudden lightning, sudden strong winds, and sudden high-decibel sounds. According to a pre-stored sudden event impact weight table, assign a corresponding basic impact weight to each identified sudden event outside the window, and calculate the sudden event impact coefficient. ; S4. Perform teacher identity detection on the teaching video collected in step S2 to identify whether there are other adult listeners besides the main lecturer in the video frame, and calculate the teaching status influence coefficient. ; S5. Calculate the overall teaching quality score for this lesson using the following formula. :

[0007] in, This is a compensation function for the impact coefficient of sudden events. To determine the compensation function for the factor influence coefficient, This is a compensation function for the influence coefficient of teaching status; S6. Based on the comprehensive score of the average teaching quality of history in this subject Overall score of teaching quality for this lesson Compare and output the teaching quality assessment results.

[0008] Preferably, in step S1, the factor influence coefficient is calculated and determined using a pre-stored psychological cycle influence model. Specifically:

[0009] in, The fatigue function is the number of cycles. This is the weekend expectation function, which peaks on Friday and reaches a sub-peak on Monday. This is a function of exam stress, which is negatively correlated with the number of days until the exam. , , Weighting coefficients.

[0010] Preferably, in step S2, the attentiveness coefficient for this lesson is obtained. Specifically, it includes: S21. Extract one frame of the teaching video image every 10 seconds; S22. Perform image enhancement and illumination normalization preprocessing on the frame image; S23. Input the trained YOLO object detection model and output the head detection box and state classification for each student; S24. Number of students categorized as not paying attention in class. And the total number of students detectable in the current frame. ; S25. Calculate the percentage of students who were not paying attention in the current frame:

[0011] S26. Calculate the average percentage of students not paying attention across all sampled frames. :

[0012] in, This represents the number of sampled frames.

[0013] Preferably, the criteria for determining whether a student is not paying attention includes at least one of the following: The student's head tilt angle exceeds the preset angle threshold; Students were seen looking down at their phones. The student was seen lying on the desk. Students are eating snacks.

[0014] Preferably, in step S3, the impact coefficient of the emergency is calculated. Specifically, it includes the following: S31. Real-time video of the outside environment is collected by cameras deployed outside the classroom windows; S32. Input the video of the outside environment into the trained emergency classification neural network to identify the types of emergencies. S33. Based on the identified emergency event type, query the pre-stored emergency event impact weight table to obtain the basic impact weight. S34. Calculate the impact coefficient of emergencies. :

[0015] in, For the first The basic impact weight of a sudden event This represents the total number of emergencies detected during this lesson.

[0016] Preferably, in step S4, the influence coefficient of teaching status is calculated. Specifically, it includes the following: S41. Perform face detection on the teaching video frames and extract all face regions; S42. Compare the extracted face region with the pre-stored face feature database of the main lecturer of this class; S43. If at least one face has a similarity to the lecturer's facial features below a set threshold, and this face continues to appear in the back row of the classroom for more than a second preset time, then the person corresponding to this face is determined to be another adult student attending the class, and the teaching status influence coefficient is output. =0.5, otherwise output =0.

[0017] Preferably, the compensation function for the impact coefficient of the sudden event in step S5 is... for: .

[0018] Preferably, the compensation function for determining the factor influence coefficient in step S5 is... for: .

[0019] Preferably, the compensation function for the teaching state influence coefficient in step S5 is... for: .

[0020] Preferably, in step S6, the average comprehensive score of the history teaching quality of the subject is used as a basis. Overall score of teaching quality for this lesson The comparison and output of teaching quality assessment results include: when Below If the score exceeds 10, the teaching quality assessment result is judged as "low"; when exist If the score is within 5 points above or below the passing mark, the teaching quality assessment result is determined to be "moderate". when Higher than If the score exceeds 10, the teaching quality assessment result is judged as "excellent".

[0021] Compared with existing technologies, the advantages of this invention are: 1. This invention uses video-based student behavior recognition technology to achieve real-time monitoring of the classroom teaching process, enabling timely detection of problems in teaching and providing timely feedback for teaching improvement. It also incorporates psychological cyclical factors such as class fatigue, weekend anticipation, and exam pressure, as well as environmental factors such as unexpected events outside the window into the evaluation model, achieving a comprehensive consideration of various objective factors affecting teaching effectiveness, making the evaluation results more comprehensive and objective.

[0022] 2. This invention achieves dynamic calibration of evaluation results by comparing them with the historical average teaching quality comprehensive score, thereby improving the horizontal comparability of teaching quality among different subjects and teachers. Furthermore, by using teacher identity detection technology, it identifies whether there are observers present and corrects the evaluation results in the context of open classes, effectively eliminating the interference of special teaching scenarios on the evaluation results. Attached Figure Description

[0023] Figure 1 This is a flowchart of the teaching quality assessment method in this invention. Detailed Implementation

[0024] To facilitate understanding of this application and to make the aforementioned objectives, features, and advantages of this application more apparent, a detailed description of specific embodiments of this application is provided below in conjunction with the accompanying drawings. Numerous specific details are set forth in the following description to provide a thorough understanding of this application, and preferred embodiments are shown in the accompanying drawings. However, this application can be implemented in many different forms and is not limited to the embodiments described herein. Rather, these embodiments are provided to provide a more thorough and complete understanding of the disclosure of this application. This application can be implemented in many other ways different from those described herein, and those skilled in the art can make similar modifications without departing from the spirit of this application; therefore, this application is not limited to the specific embodiments disclosed below. Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of indicated technical features. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one of that feature. In the description of this application, "a plurality of" means at least two, such as two, three, etc., unless otherwise explicitly specified. In the description of this application, "several" means at least one, such as one, two, etc., unless otherwise explicitly specified. It should be noted that when an element is referred to as being "fixed to" another element, it can be directly attached to the other element or there may be an intervening element. When an element is referred to as being "connected to" another element, it can be directly connected to the other element or there may be an intervening element. The terms "vertical," "horizontal," "left," "right," and similar expressions used herein are for illustrative purposes only and do not represent the only possible implementations. Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs. The terminology used herein is only for describing particular implementations and is not intended to limit the scope of this application. The term "and / or" as used herein includes any and all combinations of one or more of the associated listed items.

[0025] Reference Figure 1 As shown, a teaching quality assessment method includes the following steps: S1. Obtain the fixed information for this lesson, including at least the subject category, the order of lessons on the day, the time remaining until the nearest rest day, and the time remaining until the nearest exam day. Based on the order of lessons on the day, the time remaining until the nearest rest day, and the time remaining until the nearest exam day, calculate and determine the factor influence coefficient using a pre-stored psychological cycle influence model. .

[0026] In step S1, the factor influence coefficient is calculated and determined using a pre-stored psychological cycle influence model. Specifically:

[0027] in, As a fatigue function for each period, according to research in educational psychology, students are at the lowest point of physiological arousal during the first period of the day and the first period after lunch break, and at the peak of fatigue during the last period. The weekend expectation function peaks on Friday and reaches a sub-peak on Monday, reflecting the influence of students' psychological expectations for rest days on their attention. As a function of exam stress, it is negatively correlated with the number of days until the exam, meaning that the closer to the exam, the greater the exam stress and the more focused the students are. , , These are the weighting coefficients.

[0028] S2. Collect the teaching video of this lesson's classroom setting. Using the YOLO series of object detection algorithms, perform real-time detection of students in the video frames. At preset time intervals, extract the proportion of students not paying attention in the current frame relative to the total number of students in the class. Integrate or average all proportions across all time dimensions of this lesson to obtain the attentiveness coefficient for this lesson. ; In step S2, the attentiveness coefficient for this lesson is obtained. Specifically, it includes: S21. Extract one frame of the teaching video image every 10 seconds; S22. Perform image enhancement and illumination normalization preprocessing on the frame image; S23. Input the trained YOLO object detection model and output the head detection box and state classification for each student; S24. Number of students categorized as not paying attention in class. And the total number of students detectable in the current frame. ; S25. Calculate the percentage of students who were not paying attention in the current frame:

[0029] S26. Calculate the average percentage of students not paying attention across all sampled frames. :

[0030] in, This represents the number of sampled frames.

[0031] The criteria for determining whether a student is not paying attention in class include at least one of the following: The student's head tilt angle exceeds the preset angle threshold (e.g., 45 degrees). Students were seen looking down at their phones. The student was seen lying on the desk. Students are eating snacks.

[0032] S3. Collect video footage of the environment outside the classroom window during this lesson. Based on video frame difference and abnormal event recognition algorithms, extract sudden events outside the window. Sudden events outside the window include at least sudden snowfall, sudden rain, sudden lightning, sudden strong winds, and sudden high-decibel sounds. According to the pre-stored sudden event impact weight table, assign a corresponding basic impact weight to each identified sudden event outside the window, and calculate the sudden event impact coefficient. ; In step S3, the impact coefficient of the emergency event is calculated. Specifically, it includes the following: S31. Real-time video of the outside environment is collected by cameras deployed outside the classroom windows; S32. Input the video of the outside environment into the trained emergency classification neural network to identify the type of emergency; S33. Based on the identified emergency event type, query the pre-stored emergency event impact weight table to obtain the basic impact weight; S34. Calculate the impact coefficient of emergencies. :

[0033] in, For the first The basic impact weight of a sudden event This represents the total number of emergencies detected during this lesson.

[0034] The pre-stored impact weight table for unforeseen events is as follows: Sudden snowfall: Weighting value 0.1-0.2; Sudden rainfall: weight value 0.1-0.2; Lightning: Weighting value 0.4-0.6; Gale: Weighting value 0.3-0.5; Sudden high-decibel sound: weighted value 0.5-0.8; The weight values ​​for each item are dynamically adjusted based on the climate characteristics of the school's location and historical teaching data.

[0035] S4. Perform teacher identity detection on the teaching videos collected in step S2 to identify whether there are other adult listeners besides the main lecturer in the video frames, and calculate the teaching status influence coefficient. ; In step S4, the influence coefficient of teaching status is calculated. Specifically, it includes the following: S41. Perform face detection on the teaching video frames and extract all face regions; S42. Compare the extracted face region with the pre-stored face feature database of the main lecturer of this class; S43. If at least one face has a similarity to the lecturer's facial features below a set threshold, and this face continues to appear in the back row of the classroom for more than a second preset time, then the person corresponding to this face is determined to be another adult student attending the class, and the teaching status influence coefficient is output. =0.5, otherwise output =0.

[0036] S5. Calculate the overall teaching quality score for this lesson using the following formula. :

[0037] in, This is a compensation function for the impact coefficient of sudden events. To determine the compensation function for the factor influence coefficient, This is a compensation function for the influence coefficient of teaching status; Compensation function for the impact coefficient of sudden events for:

[0038] The greater the impact of the unexpected event, the higher the compensation value, reflecting the teacher's ability to maintain classroom attention when faced with external interference. This means that, given the same level of student attentiveness, teachers who handle strong interference will receive higher scores.

[0039] Compensation function for determining the influence coefficient of factors for:

[0040] The greater the impact of certain factors (fatigue during class, weekend anticipation, exam pressure), the higher the compensation value, reflecting the teacher's ability to maintain teaching effectiveness during students' physiological and psychological low periods.

[0041] Compensation function for the influence coefficient of teaching status for:

[0042] In the case of open classes, both teachers and students may behave unnaturally due to the presence of observers, which may lead to a downward adjustment of the overall teaching quality score.

[0043] S6. Based on the comprehensive score of the average teaching quality of history in this subject Overall score of teaching quality for this lesson Compare and output the teaching quality assessment results.

[0044] In step S6, the overall score is determined based on the average teaching quality of history in the subject. Overall score of teaching quality for this lesson The comparison and output of teaching quality assessment results include: when Below If the score exceeds 10, the teaching quality assessment result is judged as "low"; when exist If the score is within 5 points above or below the passing mark, the teaching quality assessment result is determined to be "moderate". when Higher than If the score exceeds 10, the teaching quality assessment result is judged as "excellent".

[0045] Overall score of average teaching quality in history The average score is obtained by extracting the comprehensive teaching quality scores of all subjects, similar periods, similar dates, and non-public classes for the same grade in the past at least one semester, and after removing outliers.

[0046] This invention enables real-time monitoring of the classroom teaching process, allowing for timely detection of problems and providing prompt feedback for teaching improvement. It also incorporates psychological factors such as class fatigue, weekend anticipation, and exam pressure, as well as environmental factors such as unexpected events outside the window into the evaluation model. This comprehensive consideration of various objective factors affecting teaching effectiveness makes the evaluation results more comprehensive and objective.

[0047] As is known from common technical knowledge, this invention can be implemented through other embodiments that do not depart from its spirit or essential characteristics. Therefore, the disclosed embodiments described above are merely illustrative in all respects and are not the only ones. All modifications within the scope of this invention or its equivalents are included in this invention.

Claims

1. A method for evaluating teaching quality, characterized in that: Includes the following steps: S1. Obtain the fixed information for this lesson, which includes at least the subject category, the lesson order for the day, the time remaining until the nearest rest day, and the time remaining until the nearest exam day; based on the lesson order, the time remaining until the nearest rest day, and the time remaining until the nearest exam day, calculate and determine the factor influence coefficient using a pre-stored psychological cycle influence model. ; S2. Collect the teaching video of this lesson's classroom setting. Using the YOLO series of object detection algorithms, perform real-time detection of students in the video frames. At preset time intervals, extract the proportion of students not paying attention in the current frame relative to the total number of students in the class. Integrate or average all proportions across all time dimensions of this lesson to obtain the attentiveness coefficient for this lesson. ; S3. Collect environmental video outside the classroom window during this lesson, and extract sudden events outside the window based on video frame difference and abnormal event recognition algorithm. The sudden events outside the window include at least sudden snowfall, sudden rain, sudden lightning, sudden strong wind, and sudden high-decibel sound. Based on the pre-stored emergency impact weight table, a corresponding basic impact weight is assigned to each identified emergency outside the window, and the emergency impact coefficient is calculated. ; S4. Perform teacher identity detection on the teaching video collected in step S2 to identify whether there are other adult listeners besides the main lecturer in the video frame, and calculate the teaching status influence coefficient. ; S5. Calculate the overall teaching quality score for this lesson using the following formula. : in, This is a compensation function for the impact coefficient of sudden events. To determine the compensation function for the factor influence coefficient, This is a compensation function for the influence coefficient of teaching status; S6. Based on the comprehensive score of the average teaching quality of history in this subject Overall score of teaching quality for this lesson Compare and output the teaching quality assessment results.

2. The teaching quality assessment method according to claim 1, characterized in that: In step S1, the factor influence coefficient is calculated and determined using a pre-stored psychological cycle influence model. Specifically: in, The fatigue function is the number of cycles. This is the weekend expectation function, which peaks on Friday and reaches a sub-peak on Monday. This is a function of exam stress, which is negatively correlated with the number of days until the exam. , , These are the weighting coefficients.

3. The teaching quality assessment method according to claim 1, characterized in that: In step S2, the attentiveness coefficient for this lesson is obtained. Specifically, it includes: S21. Extract one frame of the teaching video image every 10 seconds; S22. Perform image enhancement and illumination normalization preprocessing on the frame image; S23. Input the trained YOLO object detection model and output the head detection box and state classification for each student; S24. Number of students categorized as not paying attention in class. And the total number of students detectable in the current frame. ; S25. Calculate the percentage of students who were not paying attention in the current frame: S26. Calculate the average percentage of students not paying attention across all sampled frames. : in, This represents the number of sampled frames.

4. The teaching quality assessment method according to claim 3, characterized in that: The criteria for determining whether a student is not paying attention in class include at least one of the following: The student's head tilt angle exceeds the preset angle threshold; Students were seen looking down at their phones. The student was seen lying on the desk. Students are eating snacks.

5. The teaching quality assessment method according to claim 1, characterized in that: In step S3, the impact coefficient of the emergency event is calculated. Specifically, it includes the following: S31. Real-time video of the outside environment is collected by cameras deployed outside the classroom windows; S32. Input the video of the outside environment into the trained emergency classification neural network to identify the type of emergency; S33. Based on the identified emergency event type, query the pre-stored emergency event impact weight table to obtain the basic impact weight; S34. Calculate the impact coefficient of emergencies. : in, For the first The basic impact weight of a sudden event This represents the total number of emergencies detected during this lesson.

6. The teaching quality assessment method according to claim 1, characterized in that: In step S4, the influence coefficient of teaching status is calculated. Specifically, it includes the following: S41. Perform face detection on the teaching video frames and extract all face regions; S42. Compare the extracted face region with the pre-stored face feature database of the main lecturer of this class; S43. If at least one face has a similarity to the lecturer's facial features below a set threshold, and this face continues to appear in the back row of the classroom for more than a second preset time, then the person corresponding to this face is determined to be another adult student attending the class, and the teaching status influence coefficient is output. =0.5, otherwise output =0.

7. The teaching quality assessment method according to claim 1, characterized in that: The compensation function for the impact coefficient of the sudden event in step S5 for: 。 8. The teaching quality assessment method according to claim 1, characterized in that: The compensation function for determining the factor influence coefficient in step S5 for: 。 9. The teaching quality assessment method according to claim 1, characterized in that: The compensation function for the teaching status influence coefficient in step S5 for: 。 10. A teaching quality assessment method according to claim 1, characterized in that: In step S6, the average teaching quality score of the subject's history is used as a basis for assessment. Overall score of teaching quality for this lesson The comparison and output of teaching quality assessment results include: when Below If the score exceeds 10, the teaching quality assessment result is judged as "low"; when exist If the score is within 5 points above or below the passing mark, the teaching quality assessment result is determined to be "moderate". when Higher than If the score exceeds 10, the teaching quality assessment result is judged as "excellent".

Images