Learning course user intelligent supervision method and system based on image recognition

Abstract

The invention discloses a learning course user intelligent supervision method and system based on image recognition, and relates to the technical field of learning state supervision, and the method comprises the steps: collecting user academic record level data, user learning image data and user learning range course data; based on the user learning image data and the micro-expression type image data, micro-expression type analysis processing during user learning is carried out, and user micro-expression type change feature data is generated; according to the learning course user intelligent supervision method and system based on image recognition, the learning efficiency of the user is judged by collecting the micro-expression data of the user and analyzing the micro-expression change characteristics of the user in the learning process, the learning efficiency of each user student in the learning course can be known conveniently, and when the learning efficiency of the user is reduced, the learning efficiency of the user is improved. The method for adjusting the learning state is recommended to the user to ensure that the user learns in a good learning state, and a more suitable course is recommended to the user when the efficiency of learning the course by the user is extremely low.

Description

Technical Field

[0001] The present invention relates to the technical field of learning state supervision, and specifically relates to an intelligent supervision method and system for learning course users based on image recognition. Background Art

[0002] In order to ensure that users can learn effectively, intelligent means are now used to monitor, evaluate, and optimize learners' online learning behaviors and states during the learning process, mainly applied to online education platforms. The prior art with the publication number CN114612271A discloses an intelligent supervision system, method, and storage medium for online courses, belonging to the technical field of online supervision, including a photo pre-storage device, a first image detection device, a permission judgment device, a human body detection device, an automatic pause device, a second image detection device, and an automatic resume device. By controlling the orderly startup of each device, it can effectively ensure that the user learning the online course is the person himself, and can also prevent the user learning the online course from hanging up. When the user learning the online course leaves, the learning course will be automatically paused. After the user comes back, it will continue to judge whether it is the user himself learning the online course. After face recognition is completed, it will judge whether the user learning the online course is annotating the online course direction to determine whether to enter the learning state. After all requirements are met, the learning course will be automatically resumed; moreover, the orderly startup of each device can greatly reduce the energy consumption of the entire system and avoid some devices from being in an ineffective working state all the time.

[0003] However, only judging whether the user is learning by whether the line of sight is fixed still has problems such as being unable to judge whether the user is distracted or whether the user is focused on learning, and also unable to adjust the user's learning rhythm according to the user's learning state to ensure the user's learning efficiency. Summary of the Invention

[0004] The purpose of the present invention is to provide an intelligent supervision method and system for learning course users based on image recognition to solve the above deficiencies in the prior art.

[0005] To achieve the above purpose, the present invention provides the following technical solution: An intelligent supervision method for learning course users based on image recognition, including the following steps: S1. Collect user learning achievement level data, user learning image data, and user learning scope course data; S2. Based on the user learning image data and micro-expression type image data, perform micro-expression type analysis and processing during user learning to generate user micro-expression type change feature data; S3. Based on the user micro-expression type change feature data of the recently set duration and the micro-expression type change feature data of the abnormal type, perform user abnormal learning state type analysis and processing to generate user abnormal learning state type analysis data; S4. Analyze the change characteristics of the data based on the types of abnormal learning states of the user, and the change characteristic data of the high- and low-efficiency learning abnormal learning state types, and conduct high- and low-efficiency learning analysis of the user; S5. If the user is in low-efficiency learning, based on the user's learning performance level data, the user's learning scope course data, and the non-low-efficiency learning user scope course performance level data, search for non-low-efficiency learning courses with the same learning scope and user learning performance level, and generate user recommended course data; S6. If the user is in non-low-efficiency learning, based on the change characteristics of the data analyzed for the types of abnormal learning states of the user and the characteristic data of the adjustment plan for the types of abnormal learning states of the user, conduct analysis of the adjustment plan matching the abnormal learning states of the user, and generate adjustment plan analysis data; S7. Construct management data and execute user learning management operations based on the management data.

[0006] Furthermore, the S1 includes the following steps: S11. Collect user learning performance level data and generate a set of user learning performance level data , , indicating the i-th user learning performance level data, where is the maximum number of users; S12. Generate a set of user learning image data by collecting the monitoring video during the user's learning , , indicating the w-th user learning image data of the i-th user, where is the maximum number of user learning image data for the i-th user; S13. Collect user learning scope course data and generate a set of user learning scope course data , , indicating the p-th course in the o-th learning scope, where is the maximum number of course types in the o-th learning scope.

[0007] Furthermore, the S2 includes the following steps: S21. Collect micro-expression type image data and generate a set of micro-expression type image data , , indicating the u-th micro-expression type image data, where is the maximum number of micro-expression types; S22. After extracting the features of the set of user learning image data B and the set of micro-expression type image data D, search for the user image data in the set of user image data B based on the nearest neighbor search algorithm Matched micro-expression type image data to generate a set of user micro-expression type change feature data ; Among them, feature extraction can be carried out by using a pre-trained deep convolutional neural network (such as ResNet, VGG, EfficientNet) to extract high-dimensional feature vectors of images. For example, all images in B and D are input into the model, and the features of the last fully connected layer or pooling layer are extracted as representations.

[0008] Furthermore, S3 includes the following steps: S31. Use a sliding window with a set length to slide on the set of user micro-expression type change feature data to generate the current user micro-expression type change feature data ; S32. Collect abnormal type micro-expression type change feature data to generate a set of abnormal type micro-expression type change feature data , , denotes the micro-expression type change feature data of the w-th abnormal type, denotes the maximum number of abnormal types; S33. Based on the ant colony search algorithm, search for the abnormal type micro-expression type change feature data matching the current user micro-expression type change feature data to generate user abnormal learning state type analysis data , specifically including the following steps: S331. Initialize the algorithm parameters, pheromone, the number of abnormal type search ants N, and the maximum number of iterations T. The algorithm parameters include the pheromone evaporation coefficient (usually 0 < < 1), the importance degree of pheromone , the importance degree of heuristic factor , the pheromone enhancement constant ; S332. Let each abnormal type search ant start from the starting point, select the next node to reach according to the probability selection rule formula until the complete path is completed, and calculate the complete path distance , and the probability selection rule formula is as follows: , where, denotes the probability that the abnormal type search ant k selects to move from node i to node j, denotes the pheromone concentration of the path (i, j) at the t-th iteration, is the heuristic factor, and u is the point not visited by the abnormal type search ant k; S333. Update the pheromone concentration according to the proportional evaporation of the pheromones of all edges. The formula is as follows: ; S334. Simulate that the ant searching for the abnormal type releases pheromones on the path (i, j) to update the pheromones. The formula is as follows: ; Among them, ; S335. If the maximum number of iterations T is reached, output the complete path distance The characteristic data of the change in the micro-expression type of the abnormal type corresponding to the shortest complete path , and generate the analysis data of the user's abnormal learning state type , otherwise return to step S332.

[0009] Furthermore, the said S4 includes the following steps: S41. Collect the characteristic data of the change in the abnormal learning state type of efficient learning and the characteristic data of the change in the abnormal learning state type of inefficient learning, and generate a set of characteristic data of the change in the abnormal learning state type of high and low efficiency learning , where represents the characteristic data of the change in the abnormal learning state type of efficient learning, represents the characteristic data of the change in the abnormal learning state type of inefficient learning; S42. Collect the analysis data of the user's abnormal learning state type for the most recent set duration at the set frequency , and generate the analysis data of the user's abnormal learning state type for the current period ; S43. Based on the ant colony search algorithm, use the ant searching for the characteristic data of the change in the abnormal learning state type to search for the characteristic data of the change in the abnormal learning state type of high and low efficiency learning that matches the analysis data of the user's abnormal learning state type for the current period , and generate the judgment data of the high and low efficiency learning state , where if no characteristic data of the change in the abnormal learning state type of high and low efficiency learning that matches the analysis data of the user's abnormal learning state type for the current period is searched, then is 0.

[0010] Furthermore, the said S5 includes the following steps: S51. Collect, combine the judgment data of the high and low efficiency learning state , the user's learning achievement level data and the user's learning scope course data , and generate the learning user scope course achievement level data G = ( , , ); S52. Mark the learning user range course score level data G that is not the low - efficiency learning abnormal learning state type change feature data as the non - low - efficiency learning user range course score level data ; ; S53. Collect and combine the user learning score level data and the user learning range course data to generate the user learning score level range course data H = ( , ); S54. If the high - low - efficiency learning state judgment data is the low - efficiency learning abnormal learning state type change feature data , then search for the non - low - efficiency learning user range course score level data with the same labels i and o as in the user learning score level range course data H based on the double - pointer algorithm to generate the user recommended course data L.

[0011] Further, S6 includes the following steps: S61. Collect the adjustment plan user abnormal learning state type feature data to generate the adjustment plan user abnormal learning state type feature data set , , representing the user abnormal learning state type feature data corresponding to the r - th adjustment plan, representing the maximum number of adjustment plans; S62. If the high - low - efficiency learning state judgment data is not the low - efficiency learning abnormal learning state type change feature data , then search for the adjustment plan user abnormal learning state type feature data matching the current - period user abnormal learning state type analysis data to generate the adjustment plan analysis data .

[0012] Further, S7 includes the following steps: S71. Collect and combine the user recommended course data L and the adjustment plan analysis data to generate the user course learning management data K = (L, ); S72. Based on L and in the user course learning management data K, perform course recommendation and adjustment plan feedback operations for the user.​

[0013] Intelligent supervision system for learning course users based on image recognition, including a monitoring module, a data input module, a memory, a processor, and a management feedback module.

[0014] The monitoring module is used to monitor the state of users during learning and generate user learning image data; The data input module is used to input the collected data, such as user learning achievement level data, user learning image data, user learning scope course data, micro-expression type image data, abnormal type micro-expression type change feature data, high / low efficiency learning abnormal learning state type change feature data, adjustment plan user abnormal learning state type feature data, etc. into the system; The memory is used to store computer programs; The processor is used to run the computer program to execute the intelligent supervision method for learning course users based on image recognition provided by the present invention and generate user course learning management data; The management feedback module is used to feedback recommended courses and adjustment plans to users based on user course learning management data.

[0015] 1. Compared with the prior art, the intelligent supervision method and system for learning course users based on image recognition provided by the present invention can judge the learning efficiency of users by collecting user micro-expression data and analyzing the micro-expression change characteristics of users during the learning process, which is convenient for understanding the learning efficiency of each user of the learning course.

[0016] 2. Compared with the prior art, the intelligent supervision method and system for learning course users based on image recognition provided by the present invention can also recommend methods for adjusting the learning state to users when the learning efficiency of users decreases by monitoring the micro-expression change characteristics of users, so as to ensure that users learn in a good learning state.

[0017] 3. Compared with the prior art, the intelligent supervision method and system for learning course users based on image recognition provided by the present invention can also judge the adaptability of the course to users at each achievement level by understanding the learning efficiency of users at each achievement level of each course, so as to recommend more suitable courses to users when the learning efficiency of users in the course is extremely low. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the following will briefly introduce the drawings required for use in the embodiments. Obviously, the drawings described below are only some embodiments recorded in the present invention. For those of ordinary skill in the art, other drawings can also be obtained based on these drawings.

[0019] Figure 1Method step diagram provided by an embodiment of the present invention; Figure 2 System structure block diagram provided by an embodiment of the present invention. Detailed implementation manners

[0020] In order to enable those skilled in the art to better understand the technical solutions of the present invention, the present invention will be further described in detail below with reference to the accompanying drawings.

[0021] In the following, example embodiments will be described more fully with reference to the accompanying drawings, but the example embodiments may be embodied in different forms and should not be construed as limited to the embodiments set forth herein. On the contrary, these embodiments are provided so that this disclosure will be thorough and complete, and will fully convey the scope of the disclosure to those skilled in the art.

[0022] In the case of no conflict, the various embodiments of the present disclosure and the features in the embodiments may be combined with each other.

[0023] As used herein, the term "and / or" includes any and all combinations of one or more of the associated listed items.

[0024] The terms used herein are only for describing specific embodiments and are not intended to limit the present disclosure. As used herein, the singular forms "a" and "the" are also intended to include the plural forms unless the context clearly indicates otherwise. It will also be understood that when the terms "comprises" and / or "consists of" are used in this specification, it specifies the presence of the stated features, integers, steps, operations, elements and / or components, but does not preclude the presence or addition of one or more other features, integers, steps, operations, elements, components and / or their groups.

[0025] The embodiments described herein may be described with reference to plan views and / or cross-sectional views by means of ideal schematic diagrams of the present disclosure. Therefore, the example illustrations may be modified according to manufacturing techniques and / or tolerances. Therefore, the embodiments are not limited to the embodiments shown in the drawings, but include modifications of configurations formed based on manufacturing processes. Therefore, the regions illustrated in the drawings have schematic properties, and the shapes of the regions shown in the drawings illustrate the specific shapes of the regions of the elements, but are not intended to be restrictive.

[0026] Please refer to Figure 1 , a user intelligent supervision method for learning courses based on image recognition, including the following steps: S1. Collect user learning achievement level data, user learning image data, and user learning scope course data, including the following steps: S11. Collect user learning achievement level data and generate a user learning achievement level data set , , represents the i-th user learning achievement level data, is the maximum number of users; further, the learning achievement level data of each user can be the average value or median value of the test scores of the user's various subjects in the most recent setting times, etc.

[0027] S12. Generate a user learning image data set by collecting the monitoring video during the user's learning , , represents the w-th user learning image data of the i-th user, is the maximum number of user learning image data of the i-th user, and each user learning image data can be a frame in the monitoring video of the user.

[0028] S13. Collect user learning scope course data and generate a user learning scope course data set , , represents the p-th course in the o-th learning scope, represents the maximum number of course types in the o-th learning scope, where the learning scope is the teaching content scope of the course, that is, the teaching content of the courses within each learning scope is the same, but the teaching methods, styles, and difficulties of understanding are different.

[0029] S2. Based on the user learning image data and micro-expression type image data, perform micro-expression type analysis and processing during the user's learning to generate user micro-expression type change feature data, including the following steps: S21. Collect micro-expression type image data and generate a micro-expression type image data set , , represents the u-th micro-expression type image data, represents the maximum number of micro-expression types; S22. After extracting features from the user learning image data set B and the micro-expression type image data set D, search for the micro-expression type image data matching the user image data in the user image data set B based on the nearest neighbor search algorithm to generate a user micro-expression type change feature data set ; among them, feature extraction can extract the high-dimensional feature vectors of the images by using a pre-trained deep convolutional neural network (such as ResNet, VGG, EfficientNet), such as inputting all the images in B and D into the model and extracting the features of the last fully connected layer or pooling layer as the representation.

[0030] S3. Based on the user micro-expression type change feature data and abnormal type micro-expression type change feature data of the most recent set duration, perform analysis and processing on the user's abnormal learning status type to generate user abnormal learning status type analysis data, including the following steps: S31. Use a sliding window with a length of the set length to slide on the user micro-expression type change feature data set to generate the current user micro-expression type change feature data , that is, the current user micro-expression type change feature data is the user micro-expression type change feature data within the sliding window when the right end of the sliding window is aligned with the current time; S32. Collect the abnormal type micro-expression type change feature data to generate an abnormal type micro-expression type change feature data set , , represents the w-th type of abnormal type micro-expression type change feature data, that is represents the micro-expression type change feature data cluster of the w-th type of abnormal type, represents the maximum number of abnormal types; S33. Based on the ant colony search algorithm, search for the abnormal type micro-expression type change feature data matching the current user micro-expression type change feature data to generate user abnormal learning status type analysis data , specifically including the following steps: S331. Initialize the algorithm parameters, pheromone, and the number N of abnormal type search ants, and the maximum number of iterations T. The algorithm parameters include the pheromone evaporation coefficient (usually 0 < < 1), the importance degree of pheromone , the importance degree of heuristic factor , and the pheromone enhancement constant ; S332. Let each abnormal type search ant start from the starting point, select the next node to reach according to the probability selection rule formula until the complete path is completed, and calculate the complete path distance , and the probability selection rule formula is as follows: , where represents the probability that the abnormal type search ant k selects to move from node i to node j, represents the pheromone concentration of the path (i,j) at the t-th iteration, is the heuristic factor, and u is the point not visited by the abnormal type search ant k; S333. Update the pheromone concentration according to the proportional evaporation of the pheromone of all edges. The formula is as follows: ; S334. Simulate that the ant searching for the abnormal type releases pheromone on the path (i, j) to update the pheromone. The formula is as follows: ; Among them, ; S335. If the maximum number of iterations T is reached, output the distance of the complete path The change characteristic data of the abnormal micro-expression types corresponding to the shortest complete path , and generate the analysis data of the user's abnormal learning state type , otherwise return to step S332.

[0031] S4. Based on the change characteristics of the analysis data of the user's abnormal learning state type, and the change characteristic data of the abnormal learning state types of high and low-efficiency learning, conduct user high and low-efficiency learning analysis, including the following steps: S41. Collect the change characteristic data of the abnormal learning state types of high-efficiency learning and the change characteristic data of the abnormal learning state types of low-efficiency learning, and generate a set of change characteristic data of the abnormal learning state types of high and low-efficiency learning , among which represents the change characteristic data of the abnormal learning state types of high-efficiency learning, represents the change characteristic data of the abnormal learning state types of low-efficiency learning; S42. Collect the analysis data of the user's abnormal learning state type for the recently set duration at the set frequency , and generate the analysis data of the user's abnormal learning state type for the current period ; S43. Based on the ant colony search algorithm, use the ant searching for the change characteristics of the abnormal learning state type to search for the change characteristic data of the high and low-efficiency learning abnormal learning state types that match the analysis data of the user's abnormal learning state type for the current period to generate the high and low-efficiency learning state judgment data . The principle of the steps is the same as that of S331 - S335, and will not be elaborated here. Among them, if no change characteristic data of the high and low-efficiency learning abnormal learning state types that match the analysis data of the user's abnormal learning state type for the current period is searched, then is 0.

[0032] S5. If the user is in inefficient learning, based on the user's learning performance level data, the user's learning scope course data, and the non-inefficient learning user scope course performance level data, search for non-inefficient learning courses with the same learning scope and user learning performance level, and generate user recommended course data, including the following steps: S51. Collect and combine the high / low efficient learning status judgment data , the user's learning performance level data and the user's learning scope course data to generate the learning user scope course performance level data G = ( , , ); S52. Mark the learning user scope course performance level data G that is not the change characteristic data of the inefficient learning abnormal learning status type as the non-inefficient learning user scope course performance level data ; ; ; S53. Collect and combine the user's learning performance level data and the user's learning scope course data to generate the user's learning performance level range course data H = ( , ); S54. If the high / low efficient learning status judgment data is the change characteristic data of the inefficient learning abnormal learning status type , then based on the double-pointer algorithm, search for the non-inefficient learning user scope course performance level data with the same labels i and o as those in the user's learning performance level range course data H to generate the user recommended course data L.

[0033] S6. If the user is in non-inefficient learning, based on the change characteristics of the user's abnormal learning status type analysis data and the adjustment plan user abnormal learning status type characteristic data, conduct an analysis of the adjustment plan for user abnormal learning status matching to generate the adjustment plan analysis data, including the following steps: S61. Collect the adjustment plan user abnormal learning status type characteristic data to generate the adjustment plan user abnormal learning status type characteristic data set , , indicating the user abnormal learning status type characteristic data corresponding to the r-th adjustment plan, indicating the maximum number of adjustment plans; S62. If the high / low efficient learning status judgment data is not the change characteristic data of the inefficient learning abnormal learning status type , then based on the ant colony search algorithm, use the adjustment plan to search for the ant search and the analysis data of the user's abnormal learning state type in the current period The characteristic data of the adjustment plan user abnormal learning state type that matches , generate the analysis data of the adjustment plan , the step principle is the same as S331 - S335, and will not be elaborated here.

[0034] S7. Construct management data, and perform user learning management operations based on the management data, including the following steps: S71. Collect and combine the recommended course data L for the user and the analysis data of the adjustment plan to generate the user course learning management data K = (L, ); S72. Based on L and in the user course learning management data K, perform course recommendation and adjustment plan feedback operations for the user. When recommending courses to the user, if there is data on the course performance level within the range of non - inefficient learning users in the recommended course data L for the user among is the characteristic data of the change in the abnormal learning state type of efficient learning , then give priority to recommending to the user which is the characteristic data of the change in the abnormal learning state type of efficient learning of the user learning range course data within .

[0035] Please refer to Figure 2 , the present invention also provides an intelligent supervision system for learning course users based on image recognition, including a monitoring module, a data input module, a storage, a processor, and a management feedback module.

[0036] Among them, the monitoring module is used to monitor the state of the user during learning and generate user learning image data; The data input module is used to input the collected data, such as user learning performance level data, user learning image data, user learning range course data, micro - expression type image data, abnormal type micro - expression type change characteristic data, high - low efficient learning abnormal learning state type change characteristic data, adjustment plan user abnormal learning state type characteristic data, etc. into the system; The storage is used to store computer programs; The processor is used to run the computer program to execute the intelligent supervision method for learning course users based on image recognition provided by the present invention and generate user course learning management data; The management feedback module is used to feedback the recommended courses and adjustment plans to the user based on the user course learning management data.

[0037] Only some exemplary embodiments of the present invention have been described above by way of illustration. Without doubt, for those of ordinary skill in the art, the described embodiments can be modified in various different ways without departing from the spirit and scope of the present invention. Therefore, the above drawings and description are illustrative in nature and should not be construed as limiting the scope of protection of the claims of the present invention.

Claims

1. An intelligent supervision method for users of learning courses based on image recognition, characterized in that: It includes the following steps: S1. Collect user learning achievement level data, user learning image data, and user learning scope course data; S2. Based on the user learning image data and micro-expression type image data, perform micro-expression type analysis processing during user learning to generate user micro-expression type change feature data; S3. Based on the user micro-expression type change feature data of the recently set duration and the micro-expression type change feature data of the abnormal type, perform user abnormal learning state type analysis processing to generate user abnormal learning state type analysis data; S4. Based on the change characteristics of the user abnormal learning state type analysis data and the change characteristics of the high and low efficiency learning abnormal learning state type data, perform user high and low efficiency learning analysis; S5. If the user is inefficient in learning, then based on the user learning achievement level data, the user learning scope course data, and the non-inefficient learning user scope course achievement level data, search for non-inefficient learning courses with the same learning scope and user learning achievement level to generate user recommended course data; S6. If the user is non-inefficient in learning, then based on the change characteristics of the user abnormal learning state type analysis data and the adjustment plan user abnormal learning state type feature data, perform adjustment plan analysis for matching the user abnormal learning state to generate adjustment plan analysis data; S7. Construct management data and execute user learning management operations according to the management data.

2. The intelligent supervision method for learning course users based on image recognition according to claim 1, characterized in that: The S1 includes the following steps: S11. Collect user learning achievement level data and generate a set of user learning achievement level data , , represents the learning achievement level data of the i-th user, is the maximum number of users; S12. Generate a user learning image data set by collecting the monitoring video during the user's learning , , indicating the w-th user learning image data of the i-th user, being the maximum number of user learning image data of the i-th user; S13. Collect the course data within the user's learning scope and generate a set of course data within the user's learning scope , , represents the p-th course in the o-th learning scope represents the maximum number of course types in the o-th learning scope 3. The intelligent supervision method for learning course users based on image recognition according to claim 2, wherein: The S2 includes the following steps: S21. Collect micro-expression type image data and generate a set of micro-expression type image data , , represents the image data of the u-th type of micro-expression, represents the maximum number of micro-expression types; S22. After extracting features from the user learning image data set B and the micro-expression type image data set D, based on the nearest neighbor search algorithm, search for the micro-expression type image data matching the user image data in the user image data set B , and generate a user micro-expression type change feature data set .

4. The intelligent supervision method for learning course users based on image recognition according to claim 3, characterized in that: The S3 includes the following steps: S31. Slide a sliding window with a length equal to the set length over the user micro-expression type change feature data set to generate the current user micro-expression type change feature data ; S32. Collect the characteristic data of the micro-expression type changes of the abnormal types, and generate a set of characteristic data of the micro-expression type changes of the abnormal types , , represents the characteristic data of the micro-expression type changes of the w-th type of abnormal type, represents the maximum number of abnormal types; S33. Based on the ant colony search algorithm, search for the abnormal type micro-expression type change feature data matching the current user's micro-expression type change feature data , and generate user abnormal learning state type analysis data , specifically including the following steps: S331. Initialize algorithm parameters, pheromone, the number N of ants searching for abnormal types, and the maximum number of iterations T; S332. Let each ant for searching abnormal types start from the starting point, select the next node to reach according to the probability selection rule formula until a complete path is completed, and calculate the distance of the complete path. ; S333. Update the pheromone concentration according to the proportional evaporation of the pheromone of all edges; S334. Simulate the ants searching for abnormal types to release pheromone on the path (i,j) to update the pheromone; S335: If the maximum number of iterations T is reached, output the complete path distance Abnormal type micro-expression type change feature data corresponding to the shortest complete path , generate user abnormal learning status type analysis data , otherwise return to step S332.

5. The intelligent supervision method for learning course users based on image recognition according to claim 4, characterized in that: The S4 includes the following steps: S41. Collect the characteristic data of the change in the type of abnormal learning state for efficient learning and the characteristic data of the change in the type of abnormal learning state for inefficient learning, and generate a set of characteristic data of the change in the type of abnormal learning state for high and low efficiency learning , where represents the characteristic data of the change in the type of abnormal learning state for efficient learning, represents the characteristic data of the change in the type of abnormal learning state for inefficient learning; S42. Collect the analysis data of the user's abnormal learning status type for the most recent set duration according to the set frequency, and generate the analysis data of the user's abnormal learning status type for the current period. ; S43. Based on the ant colony search algorithm, use the change feature of the abnormal learning state type to search for ants and analyze the data of the user's abnormal learning state type in the current period Match the high- and low-efficiency learning abnormal learning state type change feature data to generate high- and low-efficiency learning state judgment data .

6. The intelligent supervision method for learning course users based on image recognition according to claim 5, characterized in that: The S5 includes the following steps: S51. Collect and combine the high- and low-efficiency learning state judgment data , the user's learning achievement level data and the user's learning scope course data to generate the learning user scope course achievement level data G = ( , , ); S52. Mark the course score level data G of the learning user range that is not the change characteristic data of the low-efficiency learning abnormal learning state type as the non-low-efficiency learning user range course score level data. ; S53. Collect and combine the user's academic performance level data and the user's learning scope course data to generate the user's academic performance level scope course data H = ( , ); S54. If the high-low efficiency learning state judgment data is inefficient learning abnormal learning state type change feature data , then search for non-inefficient learning user range course score level data with the same labels i and o in the course data H of the user's learning achievement level range based on the double pointer algorithm , and generate user recommended course data L.

7. The intelligent supervision method for learning course users based on image recognition according to claim 6, characterized in that: The S6 includes the following steps: S61. Collect the characteristic data of the abnormal learning state types of the adjustment plan users, and generate a set of characteristic data of the abnormal learning state types of the adjustment plan users , , represents the characteristic data of the abnormal learning state type corresponding to the r-th adjustment plan, represents the maximum number of adjustment plans; S62. If the high / low-efficiency learning state judgment data is not the change characteristic data of the low-efficiency learning abnormal learning state type , then search for the user abnormal learning state type analysis data in the current period matching the adjustment plan user abnormal learning state type characteristic data , and generate the adjustment plan analysis data .

8. The intelligent supervision method for learning course users based on image recognition according to claim 7, characterized in that: The S7 includes the following steps: S71. Collect and combine the user recommended course data L and the adjusted plan analysis data to generate user course learning management data K = (L, ) ; S72. Based on L in the user course learning management data K and perform course recommendation and feedback on the adjustment plan for the user. An intelligent supervision system for learning course users based on image recognition, which is used to execute the intelligent supervision method for learning course users based on image recognition according to any one of claims 1-8, is characterized in that: it includes a monitoring module, a data input module, a storage, a processor, and a management feedback module.

Images