Learning content planning method and system based on AIGC

Abstract

The invention discloses an AI GC-based learning content planning method and system, and relates to the technical field of artificial intelligence generation, and the method comprises the steps: monitoring data preprocessing, analyzing and evaluating the distraction condition of a student, prompting concentration and carrying out circular detection if the student distraction occurs, obtaining learning test data after concentration, analyzing and evaluating an unqualified condition, and continuing to carry out detection if the student is qualified. If not, recording the appearance, generating a wrong question set, and prompting to complete review of the wrong question set, so as to improve the learning effect of students. According to the method, the image data is calculated through an edge detection algorithm and an eye movement tracking technology to obtain a learning abnormal factor, the image data and the sound data are calculated through an image processing technology, the edge detection algorithm and ASCI coding to obtain a cross-over factor, and the learning abnormal factor and the cross-over factor are comprehensively analyzed to judge a distraction evaluation value.

Description

Technical Field

[0001] The present invention relates to the technical field of artificial intelligence generation, and specifically to a learning content planning method and system based on AIGC. Background Art

[0002] With the acceleration of informatization, the traditional teaching mode is changing towards the direction of personalization and interaction. As an important branch of the AI field, AIGC is a new content generation method that uses generative artificial intelligence generation technology to automatically create content such as text, images, and videos, subverting the traditional teaching content generation method.

[0003] Student classification technology is an important foundation for personalized learning and precise teaching in the field of education. It aims to divide students into different groups according to their characteristics and performances, so as to provide targeted educational services for different types of students.

[0004] For example, a method and system for realizing personalized customization of generative content using AIGC technology disclosed in the invention patent with the publication number of CN118051944B includes the following steps: a data collection module, a data processing module, a content generation module, a feedback module, and a system management module. By using a feature extraction algorithm to transform the data type and data encoding, the usage efficiency of data is improved; by using an iterative algorithm to train the recognition accuracy of the AIGC model, the long-term stability during system use is improved; and by using electronic encryption and a firewall to protect the user's data information, the protection of user privacy is improved.

[0005] For example, a teaching system based on AIGC classroom behavior feedback disclosed in the invention patent with the publication number of CN117934228A includes the following steps: realizing the teaching of a virtual private teacher through artificial intelligence technology; dividing the teaching interface into two parts, one part is a virtual teacher, which can be a cartoon character, a social celebrity, etc., and the other part is a visual teaching interface to display teaching content and assist users in better understanding the teaching content; and the system is divided into a massive open online course (MOOC) teaching part and a 1-on-1 question-and-answer part, and each part is further divided into a student end, an administrator end, and a database, which play different roles respectively; a feedback mechanism is added to collect audience feedback information such as eye movement information and facial information to analyze their concentration and unmastered knowledge points; and a speaking end judgment mechanism is proposed to judge whether the audience ends the conversation and feedback the corresponding ending words to the audience.

[0006] Although AIGC technology has broad application scenarios in the field of education, it still faces many challenges in actual teaching. For example, due to reasons such as inaccurate monitoring technology, immature algorithms, and many environmental variables, it cannot timely prompt students to stay focused when they are distracted during learning, resulting in a poor learning state and low test accuracy. Summary of the Invention

[0007] Technical problems to be solved

[0008] Aiming at the deficiencies of the prior art, the present invention provides a learning content planning method and system based on AIGC, which solves the problem of low accuracy caused by the distraction test of students.

[0009] Technical solutions

[0010] To achieve the above objectives, the present invention is realized through the following technical solutions: A learning content planning method and system based on AIGC, including the following specific steps: Step 1: Obtain the monitoring data of students, preprocess the monitoring data of students, analyze and evaluate the preprocessed monitoring data of students to obtain a distraction evaluation value; Step 2: Judge the distraction evaluation value; Step 3: If it is judged that the student is distracted, prompt the student to stay focused, return to Step 1 to continue obtaining the monitoring data of the student until it is judged that the student is focused; If it is judged that the student is focused, then execute Step 4; Step 4: Obtain learning test data, preprocess the learning test data, analyze and evaluate the preprocessed learning test data to obtain a non-compliance evaluation value; Step 5: Judge the non-compliance evaluation value; Step 6: When it is judged to be qualified, continue to detect the next class for the student; When it is judged to be unqualified, record the appearance of the student, generate a wrong question set, prompt the student to stay focused in the next class, and complete the review of the wrong question set.

[0011] Further, in Step 1, the monitoring data includes image data and sound data. The image data is processed with time stamps and grayscale conversion, and the sound data is processed with filtering; The image data after time stamp and grayscale conversion and the sound data after filtering are comprehensively analyzed to obtain a learning anomaly factor and a whispering factor; The learning anomaly factor and the whispering factor are normalized; The normalized learning anomaly factor and whispering factor are comprehensively analyzed to obtain a distraction evaluation value; In the formula, ZS represents the distraction evaluation value, XY represents the learning anomaly factor, JJ represents the whispering factor, and a represents the distraction adjustment factor.

[0012] Further, the specific method for obtaining the learning anomaly factor is as follows: The image data includes the number of pixel points, fixation point information, and head pose information; the edge detection algorithm is used to calculate the student's hair contour and face contour from the image data; the sum of the number of pixel points of the student's hair contour and face contour is calculated to obtain the student's hair contour area and face contour area; the quotient of the student's hair contour area and face contour area is calculated to obtain the contour area ratio; a head-down threshold is set; the contour area ratio is compared with the head-down threshold in real time; if the contour area ratio is greater than or equal to the head-down threshold, the difference between the contour area ratio and the head-down threshold is calculated to obtain the face anomaly factor; if the contour area ratio is less than the head-down threshold, the detection continues; the eye movement tracking technology is used to analyze the image data to obtain the anomaly duration; a fixation threshold is set; the anomaly duration is compared with the fixation threshold in real time; if the anomaly duration is greater than or equal to the fixation threshold, the difference between the anomaly duration and the fixation threshold is calculated to obtain the fixation anomaly factor; if the anomaly duration is less than the fixation threshold, the detection continues; the face anomaly factor and the fixation anomaly factor are comprehensively analyzed to obtain the learning anomaly factor.

[0013] Further, the specific method for obtaining the whispering factor is as follows: The head deflection angle of the student is obtained by analyzing and calculating the image data through image processing technology; a deflection angle threshold is set; the head deflection angle is compared with the deflection angle threshold in real time; if the head deflection angle is greater than or equal to the deflection angle threshold, the difference between the head deflection angle and the deflection angle threshold is calculated to obtain the deflection distance factor; if the head deflection angle is less than the deflection angle threshold, the detection continues; the edge detection algorithm is used to calculate the mouth opening and closing frequency from the image data; a frequency threshold is set; the mouth opening and closing frequency is compared with the frequency threshold in real time; if the mouth opening and closing frequency is greater than or equal to the frequency threshold, the difference between the mouth opening and closing frequency and the frequency threshold is calculated to obtain the opening and closing frequency factor; if the mouth opening and closing frequency is less than the frequency threshold, the detection continues; the voice recognition technology is used for the voice data to obtain the voice content text; the ASCII code database is obtained by converting the voice content text through ASCII encoding; a preset ASCII dictionary is set; the ASCII code database is matched with the preset ASCII dictionary to obtain the content similarity factor; the deflection distance factor, the opening and closing frequency factor, and the content similarity factor are comprehensively analyzed to obtain the whispering factor.

[0014] Further, in step two; the distraction evaluation value is compared with the distraction threshold in real time; if the distraction evaluation value is greater than or equal to the distraction threshold, it is determined that the student is distracted; if the distraction evaluation value is less than the distraction threshold, it is determined that the student is focused.

[0015] Further, in step three, if it is determined that the student is distracted, a prompt sound is provided, and the process returns to step one to continue obtaining the student's monitoring data until the distraction evaluation value is less than the distraction threshold, at which point it is determined that the student is focused; if it is determined that the student is focused, step four is executed.

[0016] Further, in step four, outlier deletion processing is performed on the learning test data; the learning test data after outlier deletion processing is analyzed to obtain test score factors; the test score factors are normalized; and the normalized test score factors are analyzed to obtain a non-compliance evaluation value.

[0017] Further, the specific method for obtaining the test score factors is as follows: clustering analysis is performed on the learning test data after outlier deletion processing to obtain each student category; through the AIGC technology, test content of corresponding difficulty is generated for each student category and tested to obtain test scores; a score standard threshold is set; the test scores and the score standard threshold are compared in real time; if the test score is less than the score standard threshold, the difference between the score standard threshold and the test score is calculated to obtain the test score factor; if the test score is greater than or equal to the test threshold, the detection continues.

[0018] Further, in step five; the non-compliance evaluation value and the non-compliance threshold are compared in real time; if the non-compliance evaluation value is greater than or equal to the non-compliance threshold, it is determined as non-compliant; if the non-compliance evaluation value is less than the non-compliance threshold, it is determined as compliant.

[0019] Further, a learning content planning system based on AIGC, for the above-mentioned learning content planning method based on AIGC, is characterized in that it includes: a distraction evaluation module: used to obtain the student's monitoring data and sound data, preprocess the student's monitoring data and sound data, analyze and evaluate the preprocessed student's monitoring data and sound data to obtain a distraction evaluation value; a distraction judgment processing module: used to judge the distraction evaluation value, process the distraction judgment result to obtain a focus result, and send the focus result to the test evaluation module; a test evaluation module: used to receive the focus result, obtain the learning test data, preprocess the learning test data, analyze and evaluate the preprocessed learning test data to obtain a non-compliance evaluation value; a test judgment processing module: used to judge the non-compliance evaluation value and process the non-compliance judgment result to obtain a compliance result.

[0020] Beneficial effects

[0021] Compared with the prior art, the embodiments of the present invention have at least the following advantages or beneficial effects:

[0022] 1. Calculate the learning anomaly factor by performing calculations on image data through edge detection algorithms and eye tracking technologies. Calculate the whispering factor by performing calculations on image data and sound data through image processing technologies, edge detection algorithms, and ASCII encoding. Comprehensively analyze the learning anomaly factor and the whispering factor to determine the distraction evaluation value.

[0023] 2. Preprocess and perform cluster analysis on learning test data to obtain various student categories. Generate test content of corresponding difficulty for each student category through AIGC technology. Test the test content to obtain the test score factor, and analyze the test score factor to determine the unqualified evaluation value.

[0024] 3. By comparing the distraction evaluation value with the distraction threshold, if the distraction evaluation value is greater than or equal to the distraction threshold, it is determined that the student is distracted, and a prompt sound is provided to remind the student to stay focused, solving the problem of not being able to promptly remind the student to stay focused when distracted during learning.

[0025] 4. By comparing the unqualified evaluation value with the unqualified threshold, if the unqualified evaluation value is greater than or equal to the unqualified threshold, it is determined as unqualified, record the face of the student, generate a wrong question set, and prompt the student to stay focused in the next class and complete the review of the wrong question set, solving the problem of low test accuracy.

[0026] Of course, any product implementing the present invention does not necessarily need to achieve all the above-mentioned advantages simultaneously. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] Figure 1 This is a flowchart of a method for planning learning content based on AIGC according to the present invention.

[0028] Figure 2 This is a structural diagram of a system for planning learning content based on AIGC according to the present invention. DETAILED DESCRIPTION OF THE INVENTION

[0029] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0030] It should be noted that in this article, relational terms such as first and second are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the term "including", "comprising" or any other variant thereof is intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements not only includes those elements, but also includes other elements not explicitly listed, or also includes elements inherent in such process, method, article or device.

[0031] As Figure 1 shown, an embodiment of the present invention provides a learning content planning method based on AIGC, including the following specific steps:

[0032] Step 1: Obtain the monitoring data of students, preprocess the monitoring data of students, analyze and evaluate the preprocessed monitoring data of students, and obtain a distraction evaluation value;

[0033] Collect the monitoring data of students through intelligent monitoring devices, and the monitoring data of students includes image data and sound data;

[0034] Perform timestamp processing on the image data, which helps to associate and synchronize the image data in chronological order;

[0035] Perform filtering processing on the sound data, which helps to improve the sound quality, optimize the sound transmission and storage, and promote the sound analysis and processing;

[0036] Comprehensively analyze the image data after timestamp and grayscale processing and the sound data after filtering processing to obtain a learning anomaly factor and a whispering factor;

[0037] Perform normalization processing on the learning anomaly factor and the whispering factor;

[0038] Use the maximum-minimum normalization method to scale the learning anomaly factor and the whispering factor to between [0, 1], maintaining a consistent numerical range, which helps with data visualization and analysis;

[0039] Comprehensively analyze the learning anomaly factor and the whispering factor after normalization processing to obtain a distraction evaluation value;

[0040]

[0041] In the formula, ZS represents the distraction evaluation value, XY represents the learning anomaly factor, JJ represents the whispering factor, a represents the distraction adjustment factor, and its value range is between 0 and 1, which helps to adjust the output of the distraction evaluation value and maintain stability during exponential and logarithmic operations.

[0042] The specific method for obtaining the learning anomaly factor is as follows:

[0043] The image data includes pixel point information, fixation point information, and head pose information;

[0044] The pixel point information includes the number of pixel points and the area of a single pixel point;

[0045] The fixation point information includes the pupil center, corneal reflection point, and line-of-sight direction vector;

[0046] The head pose information includes the eye center point and the head contour center point;

[0047] The student's hair contour and student's face contour are obtained by calculating the image data through an edge detection algorithm;

[0048] Use the Canny edge detection algorithm; grayscale the image data to obtain grayscale image data, which includes the gradient magnitude and the gradient change direction. Calculate the gradient magnitude and the gradient change direction, retain the edges with large gradient magnitudes in the local area, and obtain the student's hair contour and student's face contour through double-threshold detection and edge connection;

[0049] Sum the number of pixel points for both the student's hair contour and the student's face contour to obtain the student's hair contour area and the student's face contour area;

[0050] Traverse the number of pixel points for both the student's hair contour and the student's face contour to obtain the number of pixel points of the student's hair contour and the number of pixel points of the student's face contour. Multiply the number of pixel points of the student's hair contour and the number of pixel points of the student's face contour by the area of a single pixel point to obtain the student's hair contour area and the student's face contour area;

[0051] Perform a quotient calculation on both the student's hair contour area and the student's face contour area to obtain the contour area ratio;

[0052] Compare the contour area ratio with the head-down threshold in real time;

[0053] If the contour area ratio is greater than or equal to the head-down threshold, then perform a difference calculation on the contour area ratio and the head-down threshold to obtain the facial anomaly factor;

[0054] If the contour area ratio is less than the head-down threshold, then continue the detection;

[0055] Analyze the image data through an eye movement tracking technology to obtain the anomaly duration;

[0056] Set an infrared light source;

[0057] Use the pupil center corneal reflection method; irradiate the smooth surface of the eyeball cornea with the light emitted by the infrared light source to cause reflection on the eyeball cornea, capture the pupil center position and the corneal reflection point position through the intelligent monitoring device, regard the vector pointing from the pupil center position to the corneal reflection point position as the line-of-sight direction vector, define the learning line-of-sight direction vector range, record the duration when the line-of-sight direction vector is not within the learning line-of-sight direction vector range, and obtain the abnormal duration;

[0058] Compare the abnormal duration and the fixation threshold in real time;

[0059] If the abnormal duration is greater than or equal to the fixation threshold, then perform a subtraction calculation on the abnormal duration and the fixation threshold to obtain the fixation abnormality factor;

[0060] If the abnormal duration is less than the fixation threshold, then continue the detection;

[0061] Conduct a comprehensive analysis of the facial abnormality factor and the fixation abnormality factor to obtain the learning abnormality factor;

[0062]

[0063] In the formula, XY represents the learning abnormality factor, MY represents the facial abnormality factor, ZY represents the fixation abnormality factor, f represents the influence weight of the facial abnormality factor and the fixation abnormality factor, g represents the learning abnormality adjustment factor, and its value range is between 0 and 1, which helps to adjust the output of the learning abnormality factor and maintain stability during exponential and logarithmic operations;

[0064] Obtain the weight factor and the adjustment factor through regression analysis of the pixel point information and the fixation point information;

[0065] Analyze and calculate the image data through image processing technology to obtain the head deflection angle of the student;

[0066] Use the coordinate transformation method to grayscale the image data to obtain the grayscale image data, use the Canny edge detection algorithm on the grayscale image data to obtain the two-dimensional coordinates of the head contour center point and the two-dimensional coordinates of the eye center point, take the optical center of the intelligent monitoring device as the origin, convert the two-dimensional coordinates to three-dimensional coordinates to obtain the three-dimensional coordinates of the head contour center point and the three-dimensional coordinates of the eye center point, and calculate the angle between the three-dimensional coordinates of the head contour center point and the three-dimensional coordinates of the eye center point to the origin to obtain the head deflection angle;

[0067] Compare the head deflection angle and the deflection angle threshold in real time;

[0068] If the head deflection angle is greater than or equal to the deflection angle threshold, then perform a subtraction calculation on the head deflection angle and the deflection angle threshold to obtain the deflection distance factor;

[0069] If the head deflection angle is less than the deflection angle threshold, then continue the detection;

[0070] The mouth opening and closing frequency is calculated from the image data through an edge detection algorithm;

[0071] The image data is grayscaled to obtain grayscale image data. The Canny edge detection algorithm is used on the grayscale image data to obtain the mouth region, which includes the upper lip contour and the lower lip contour. The mouth region is intercepted, and the distance between the upper lip contour and the lower lip contour is calculated to obtain the upper and lower lip distance;

[0072] When the upper and lower lip distance is greater than zero, it is regarded as the mouth being open;

[0073] When the upper and lower lip distance is equal to zero, it is regarded as the mouth being closed;

[0074] Taking the mouth being open, the mouth being closed, and the mouth being open as one cycle;

[0075] The number of cycles is calculated frame by frame within a certain period of time to obtain the mouth opening and closing frequency;

[0076] The mouth opening and closing frequency is compared with the frequency threshold in real time;

[0077] If the mouth opening and closing frequency is greater than or equal to the frequency threshold, then the difference between the mouth opening and closing frequency and the frequency threshold is calculated to obtain the opening and closing frequency factor;

[0078] If the mouth opening and closing frequency is less than the frequency threshold, then continue to detect;

[0079] Speech recognition technology is used on the voice data to obtain the speech content text;

[0080] The voice data is transformed from the time domain to the frequency domain, and after logarithmic compression and discrete transformation, Mel frequency cepstral coefficients are obtained. The Mel frequency cepstral coefficients can effectively represent the pitch, timbre, and other characteristics of the speech signal;

[0081] Utilizing the linear correlation of the voice data, a prediction algorithm is used to approximately represent the voice data to obtain linear predictive coding, which can reflect the formant characteristics of the speech signal;

[0082] The Mel frequency cepstral coefficients and the linear predictive coding are combined into feature vectors. Each feature vector corresponds to a speech frame. A speech frame refers to a series of short-time segments obtained by dividing continuous voice data. Using a deep neural network, a large amount of voice data is used for training. The extracted feature vectors are input, and the phoneme or syllable corresponding to each speech frame is output. A syllable is a natural unit in speech, usually consisting of one phoneme or several phonemes, and a phoneme is the smallest unit in speech;

[0083] The output phoneme or syllable is processed by a neural network algorithm to accurately convert the phoneme or syllable into meaningful words and sentences to obtain the speech content text.

[0084] The text of the voice content is converted through ASCII encoding to obtain an ASCII code database;

[0085] The text of the voice content is split one by one in order from beginning to end. For each character obtained by splitting, the corresponding ASCII code value in the ASCII encoding table is searched. The ASCII code value of each character is converted into a string form, and these ASCII code value strings are concatenated in the order of the characters in the original text to obtain an ASCII code database;

[0086] Set a preset ASCII dictionary;

[0087] Set the monitored learning keywords. Arrange the ASCII code values corresponding to each character of the monitored learning keywords in order and store the keywords and their corresponding ASCII code sequences in a dictionary to obtain a preset ASCII dictionary;

[0088] Match the ASCII code database with the preset ASCII dictionary to obtain a content similarity factor;

[0089] Set a counter M, the initial value of the counter M is 0, and count the number of different keywords;

[0090] Set a counter N, the initial value of the counter N is 0, and count the number of times of the same keyword;

[0091] Loop through the ASCII string of the text and check character by character. If M + N is greater than or equal to 3, obtain a content similarity factor;

[0092] Comprehensively analyze the deflection distance factor, opening and closing frequency factor, and content similarity factor to obtain a whispering factor.

[0093]

[0094] In the formula, JJ represents the whispering factor, PJ represents the deflection distance factor, KP represents the opening and closing frequency factor, NX represents the content similarity factor, h represents the influence weight of the deflection distance factor, opening and closing frequency factor, and content similarity factor, and the value range is between 0 and 1. i represents the whispering adjustment factor, which helps to adjust the output of the whispering factor and maintains stability during exponential and logarithmic operations;

[0095] Through the linear regression algorithm, input the deflection distance factor, opening and closing frequency factor, content similarity factor, and whispering factor, establish a neural network model, find the best linear relationship, and output the influence weight of the deflection distance factor, opening and closing frequency factor, content similarity factor, and the whispering adjustment factor;

[0096] Step 2: Judge the distraction evaluation value;

[0097] Compare the distraction evaluation value with the distraction threshold in real time;

[0098] If the distraction evaluation value is greater than or equal to the distraction threshold, it is judged that the student is distracted;

[0099] If the distraction evaluation value is less than the distraction threshold, it is judged that the student is focused.

[0100] Step 3: When it is judged that the student is distracted, prompt the student to stay focused, return to Step 1 to continue obtaining the student's monitoring data until it is judged that the student is focused;

[0101] If it is judged that the student is focused, execute Step 4.

[0102] If it is judged that the student is distracted, provide a prompt sound, return to Step 1 to continue obtaining the student's monitoring data and sound data until the distraction evaluation value is less than the distraction threshold and it is judged that the student is focused;

[0103] If it is judged that the student is focused, execute Step 4.

[0104] Step 4: Obtain learning test data, preprocess the learning test data, analyze and evaluate the preprocessed learning test data to obtain a non-compliance evaluation value;

[0105] In Step 4, perform outlier deletion processing on the learning test data;

[0106] Analyze the learning test data after outlier deletion processing to obtain test score factors;

[0107] The specific method for obtaining the test score factors is as follows:

[0108] Perform clustering analysis on the learning test data after outlier deletion processing to obtain each student category;

[0109] Use the K-means clustering algorithm; perform standardization processing on the learning test data after outlier deletion processing to obtain standard test data. The standard test data contains each student's test data point. Set three student categories: excellent, passing, and to be improved. Randomly select 3 student test data points as the initial center points, and assign each student's test data point to the category represented by the center point with the closest Euclidean distance. Recalculate the mean of all student test data points in each category and use this mean as the new center point. Preset the maximum number of iterations, and repeat the process of assigning data points and updating the center points until the maximum number of iterations is reached. Finally, each student's test data point is assigned to a category, and each category corresponds to a center point, obtaining 3 student categories, representing excellent, passing, and to be improved respectively;

[0110] Use AIGC technology to generate test content of corresponding difficulty for each student category, conduct tests, and obtain test scores;

[0111] The test content parameters include student category, knowledge points, and question type distribution. Using AIGC technology, test content of corresponding difficulty is automatically generated according to the test content parameters, improving the efficiency of test content generation, covering a wide range of knowledge points and question types, generating test content of appropriate difficulty for each student type, keeping the test content novel, and stimulating students' thinking and innovation abilities;

[0112] Compare the test scores with the score standard threshold;

[0113] If the test score is less than the score standard threshold, then calculate the difference between the test score and the score standard threshold to obtain the test score factor;

[0114] If the test score is greater than or equal to the test threshold, continue the detection;

[0115] Normalize the test score factor;

[0116] Scale the test score factor to the range of [0, 1] using the maximum-minimum normalization method, maintaining a consistent numerical range, which is helpful for data visualization and analysis;

[0117] Analyze the normalized test score factor to obtain the unqualified evaluation value.

[0118]

[0119] In the formula, BH represents the unqualified evaluation value, and CC represents the test score factor;

[0120] Step Five: Judge the unqualified evaluation value;

[0121] In Step Five;

[0122] Compare the unqualified evaluation value with the unqualified threshold in real time,

[0123] If the unqualified evaluation value is greater than or equal to the unqualified threshold, then judge as unqualified;

[0124] If the unqualified evaluation value is less than the unqualified threshold, then judge as qualified.

[0125] Step Six: When it is judged as qualified, continue to detect the student in the next class;

[0126] When it is judged as unqualified, record the appearance of the student, generate a wrong question set, and prompt the student to stay focused in the next class and complete the review of the wrong question set;

[0127] Set up a student personal information database and a student facial image database;

[0128] When the judgment is unqualified, the intelligent monitoring device records the image data of the student, uses the feature extraction algorithm to obtain the eye area and the nose area, calculates the length and width of the eyes, the angle between the line connecting the inner and outer eye corners and the horizontal line, the height of the nose bridge, and the width of the nose bridge, combines them into a feature vector, matches it with the feature vector in the stored student facial image database, uses the cosine similarity, sets the cosine similarity threshold, and if it exceeds the cosine similarity threshold, it is regarded as a successful match. Record the unqualified evaluation value in the student personal information database. When the student goes to the next class, prompt the student to stay focused;

[0129] Obtain the wrong question set according to the test content, use AIGC technology to analyze the important knowledge points in the wrong question set, obtain the review key points, save the review key points in the student personal information database, and prompt the student to complete the study of the review key points in the next class.

[0130] As Figure 2 shown: A learning content planning system based on AIGC, including:

[0131] The distraction assessment module: used to obtain the monitoring data and voice data of the student, preprocess the monitoring data and voice data of the student, analyze and evaluate the preprocessed monitoring data and voice data of the student, and obtain the distraction assessment value;

[0132] The distraction judgment processing module: used to judge the distraction assessment value, process the distraction judgment result to obtain the focus result, and send the focus result to the test assessment module;

[0133] The test assessment module: used to receive the focus result, obtain the learning test data, preprocess the learning test data, analyze and evaluate the preprocessed learning test data, and obtain the unqualified evaluation value;

[0134] The test judgment processing module: used to judge the unqualified evaluation value, process the unqualified judgment result, and obtain the qualified result.

[0135] The preferred embodiments of the present invention disclosed above are only used to help illustrate the present invention. The preferred embodiments do not describe all the details in detail, nor do they limit the invention to the specific embodiments described. Obviously, many modifications and variations can be made according to the content of this specification. These embodiments are selected and specifically described in this specification to better explain the principles and practical applications of the present invention, so that those skilled in the relevant technical fields can understand and utilize the present invention well. The present invention is only limited by the claims and their full scope and equivalents.

Claims

1. A learning content planning method based on AIGC, characterized in that: It includes the following specific steps: Step 1: Obtain the monitoring data of students, preprocess the monitoring data of students, analyze and evaluate the preprocessed monitoring data of students, and obtain a distraction evaluation value; Step 2: Judge the distraction evaluation value; Step 3: If distraction is judged, prompt the student to stay focused, return to Step 1 to continue obtaining the monitoring data of the student until focus is judged; If focus is judged, then execute Step 4; Step 4: Obtain learning test data, preprocess the learning test data, analyze and evaluate the preprocessed learning test data, and obtain a non-compliance evaluation value; Step 5: Judge the non-compliance evaluation value; Step 6: When compliance is judged, continue to detect the student in the next class; When non-compliance is judged, record the appearance of the student, generate a wrong-question set, and prompt the student to stay focused in the next class and complete the review of the wrong-question set.

2. The method for planning learning content based on AIGC according to claim 1, characterized in that: In Step 1, the monitoring data includes image data and sound data. Timestamp and grayscale processing are performed on the image data, and filtering processing is performed on the sound data; Comprehensively analyze the image data after timestamp and grayscale processing and the sound data after filtering processing to obtain a learning anomaly factor and a whispering factor; Perform normalization processing on the learning anomaly factor and the whispering factor; Comprehensively analyze the normalized learning anomaly factor and whispering factor to obtain a distraction evaluation value; In the formula, ZS represents the distraction evaluation value, XY represents the learning anomaly factor, JJ represents the whispering factor, and a represents the distraction adjustment factor.

3. The method for planning learning content based on AIGC according to claim 2, wherein: The specific method for obtaining the learning anomaly factor is as follows: The image data includes the number of pixel points, fixation point information, and head pose information; Calculate the student's hair contour and face contour from the image data through an edge detection algorithm; Sum the number of pixel points of the student's hair contour and face contour respectively to obtain the student's hair contour area and face contour area; Perform a quotient calculation on the student's hair contour area and face contour area to obtain a contour area ratio; Compare the contour area ratio with the head-down threshold in real time; If the contour area ratio is greater than or equal to the head-down threshold, then perform a difference calculation on the contour area ratio and the head-down threshold to obtain a facial anomaly factor; If the contour area ratio is less than the head-down threshold, continue to detect; Analyze the image data through an eye movement tracking technology to obtain an abnormal duration; Compare the abnormal duration with the fixation threshold in real time; If the abnormal duration is greater than or equal to the fixation threshold, then perform a difference calculation on the abnormal duration and the fixation threshold to obtain a fixation anomaly factor; If the abnormal duration is less than the fixation threshold, continue to detect; Comprehensively analyze the facial anomaly factor and the fixation anomaly factor to obtain a learning anomaly factor.

4. The method for planning learning content based on AIGC according to claim 2, wherein: The specific method for obtaining the whispering factor is as follows: Analyze and calculate the student's head deflection angle from the image data through an image processing technology; Compare the head deflection angle with the deflection angle threshold in real time; When the deflection angle of the head is greater than or equal to the deflection angle threshold, the difference between the deflection angle of the head and the deflection angle threshold is calculated to obtain the deflection distance factor; When the deflection angle of the head is less than the deflection angle threshold, continue to detect; The mouth opening and closing frequency is calculated from the image data through an edge detection algorithm; The mouth opening and closing frequency is compared with the frequency threshold in real time; When the mouth opening and closing frequency is greater than or equal to the frequency threshold, the difference between the mouth opening and closing frequency and the frequency threshold is calculated to obtain the opening and closing frequency factor; When the mouth opening and closing frequency is less than the frequency threshold, continue to detect; Speech recognition technology is used for the voice data to obtain the speech content text; The speech content text is converted through ASCII encoding to obtain the ASCII code database; A preset ASCII dictionary is set; The ASCII code database is matched with the preset ASCII dictionary to obtain the content similarity factor; The deflection distance factor, the opening and closing frequency factor, and the content similarity factor are comprehensively analyzed to obtain the whispering factor.

5. The method for planning learning content based on AIGC according to claim 1, wherein: In step two, the distraction evaluation value is compared with the distraction threshold in real time; When the distraction evaluation value is greater than or equal to the distraction threshold, it is determined that the student is distracted; When the distraction evaluation value is less than the distraction threshold, it is determined that the student is focused.

6. The method for planning learning content based on AIGC according to claim 1, wherein: In step three, when it is determined that the student is distracted, a prompt sound is provided, and the process returns to step one to continue obtaining the monitoring data of the student until the distraction evaluation value is less than the distraction threshold, then it is determined that the student is focused; When it is determined that the student is focused, step four is executed.

7. The method for planning learning content based on AIGC according to claim 1, wherein: In step four, outlier deletion processing is performed on the learning test data; The learning test data after outlier deletion processing is analyzed to obtain the test score factor; Normalization processing is performed on the test score factor; The normalized test score factor is analyzed to obtain the unqualified evaluation value.

8. The method for planning learning content based on AIGC according to claim 7, wherein: The specific way to obtain the test score factor is: Cluster analysis is performed on the learning test data after outlier deletion processing to obtain each student category; Through AIGC technology, test content of corresponding difficulty is generated for each student category and tested to obtain the test score; The test score is compared with the score standard threshold in real time; When the test score is less than the score standard threshold, the difference between the score standard threshold and the test score is calculated to obtain the test score factor; When the test score is greater than or equal to the test threshold, continue to detect.

9. The method for planning learning content based on AIGC according to claim 1, wherein: In step five, the unqualified evaluation value is compared with the unqualified threshold in real time; When the unqualified evaluation value is greater than or equal to the unqualified threshold, it is determined as unqualified; When the unqualified evaluation value is less than the unqualified threshold, it is determined as qualified.

10. A learning content planning system based on AIGC, for a learning content planning method based on AIGC according to any one of claims 1-9, characterized in that, Including: Mind Wandering Evaluation Module: It is used to obtain the monitoring data and voice data of students, preprocess the monitoring data and voice data of students, analyze and evaluate the preprocessed monitoring data and voice data of students, and obtain the mind wandering evaluation value; Mind Wandering Judgment Processing Module: It is used to judge the mind wandering evaluation value, process the mind wandering judgment result to obtain the concentration result, and send the concentration result to the test evaluation module; Test Evaluation Module: It is used to receive the concentration result, obtain the learning test data, preprocess the learning test data, analyze and evaluate the preprocessed learning test data, and obtain the unqualified evaluation value; Test Judgment Processing Module: It is used to judge the unqualified evaluation value and process the unqualified judgment result to obtain the qualified result.

Images