A 5G-based intelligent inspection and proctoring method, system, and electronic equipment

Abstract

The present invention provides a 5G-based intelligent inspection and proctoring method, system, and electronic device. The method includes: constructing an examination room inspection map, retrieving a personnel information set related to each examinee from a personnel database; constructing a cheating matrix for the corresponding examinee and obtaining cheating features; sorting similar cheating features from largest to smallest in number to obtain a first sorting result; sorting the feature importance of the corresponding examination venues from largest to smallest according to the weighting of different cheating features to obtain a second sorting result; locking key supervision features and non-key supervision features to obtain a first inspection mode; adjusting the corresponding first inspection mode based on the proctoring mode to obtain a second inspection mode; and obtaining a third inspection mode according to the second inspection mode for all examination venues in the examination room inspection map and the location deployment of the examination venues in the examination room inspection map. This reduces the consumption of human proctoring and improves the intelligence of proctoring.

Description

Technical Field

[0001] The present invention relates to the technical field of Internet of Things, and in particular to a 5G-based intelligent examination inspection and supervision method, system and electronic equipment. Background Art

[0002] With the accelerated deployment of fifth-generation mobile communications (5G) and the gradual application of emerging technologies such as the Internet of Things, cloud computing, big data analytics, and artificial intelligence, education is gradually transforming from traditional education to smart education, ushering in a new era of educational informatization. Furthermore, with the development of 5G and digital technologies, the educational examination field is also undergoing a transformation towards informatization, digitization, and intelligence. Compared to traditional online exams, 5G+ smart exams leverage the inherent technical and business advantages of 5G networks by utilizing 5G-enabled hardware terminals, providing school users with a faster, better, and smoother experience.

[0003] In recent years, with the increase in the number of examinees and the rapid development of new technologies, various factors have collectively exerted greater comprehensive pressure on the National Open University's examination management, examination evaluation, examination bases and examination team building, resulting in unreasonable allocation of personnel in the process of invigilation, reducing invigilation efficiency.

[0004] Therefore, the present invention proposes a 5G-based intelligent inspection and supervision method, system and electronic equipment. Summary of the Invention

[0005] The present invention provides a 5G-based intelligent inspection and supervision method, system and electronic equipment, which are used to determine the supervision characteristics of each examination venue by obtaining a personnel information set, and then obtain an inspection method for the examination venue. The inspection method is adjusted by obtaining the historical supervision status of the invigilators, so as to achieve reasonable and strict supervision of different venues, realize intelligent inspection and supervision, reduce the consumption of human supervision and improve the intelligence of supervision.

[0006] The present invention provides a 5G-based intelligent inspection and proctoring method, comprising:

[0007] Step 1: Construct an examination room inspection map and obtain the examiners of each examination venue in the examination room inspection map;

[0008] Step 2: Retrieve the personnel information set related to each examinee from the personnel database;

[0009] Step 3: Based on the personnel information set, construct a cheating matrix corresponding to the examinee and obtain the cheating characteristics of the examinee;

[0010] Step 4: Based on the obtained cheating features, all cheating features of the corresponding examination venue are obtained, and similar cheating features are sorted from the largest to the smallest number to obtain a first sorting result;

[0011] At the same time, according to the pre-set weight allocation of different cheating characteristics, the characteristics of the corresponding examination venues are ranked from large to small in importance to obtain a second ranking result;

[0012] Step 5: Based on the first sorting result and the second sorting result, lock the key supervision features and the non-key supervision features to obtain the first inspection method;

[0013] Step 6: Obtain the historical invigilation status of the invigilators assigned to each examination venue, obtain the invigilation method of the corresponding examination venue, and adjust the corresponding first inspection method based on the invigilation method to obtain an adapted second inspection method;

[0014] Step 7: According to the second inspection method of all the examination venues in the examination venue inspection map and the venue deployment of the examination venues in the examination venue inspection map, a third inspection method is obtained to realize intelligent inspection and supervision.

[0015] Preferably, the personnel information set includes the examination status of the corresponding examinee during the historical examination process, wherein the examination status includes whether the corresponding examinee cheated and the cheating patterns corresponding to each examination subject in each historical examination, wherein the cheating patterns are obtained by extracting and analyzing the local movements of the examinee.

[0016] Preferably, the third inspection method is obtained to realize the process of intelligent inspection and proctoring, further comprising:

[0017] Determining the examination inspection equipment at the examination site based on the third inspection method;

[0018] Controlling the examination inspection equipment to capture videos and conduct behavior analysis on the persons to be monitored during the examination;

[0019] Remote proctoring of examinees at the same examination location based on cameras installed at each examination location;

[0020] When the person to be monitored is a first person who continues to answer questions at the examination location, a first event of interest for the first person is determined based on the corresponding behavior analysis result, and remote proctoring information for the first person is obtained to obtain a second event of interest;

[0021] Based on the first event of concern and the second event of concern, extracting and compressing a first abnormal video segment, and transmitting it to a remote end for storage and a first early warning reminder;

[0022] When the monitored person leaves the examination location to answer questions, the second person is dynamically tracked and monitored to obtain a third event of interest, and a second abnormal video segment is extracted and compressed, and transmitted to a remote terminal for storage and a second warning reminder;

[0023] When the person to be monitored is an invigilator, abnormal invigilation behavior is determined based on the behavior analysis result corresponding to the invigilator and transmitted to the remote end for a third early warning reminder.

[0024] Preferably, based on the personnel information set, a cheating matrix corresponding to the examinee is constructed, and cheating characteristics of the examinee are obtained, including:

[0025] Analyze the personnel information set of the same examinee and construct the cheating matrix Z of the corresponding examinee;

[0026]

[0027] Where Z represents the cheating matrix of the corresponding examinee; A1 represents the behavior sub-matrix of the corresponding examinee in the first examination of subject A obtained according to the analysis results; An1 represents the behavior sub-matrix of the corresponding examinee in the n1th examination of subject A obtained according to the analysis results; B1 represents the behavior sub-matrix of the corresponding examinee in the first examination of subject B obtained according to the analysis results; Bn1 represents the behavior sub-matrix of the corresponding examinee in the n1th examination of subject B obtained according to the analysis results;

[0028] Obtaining behavioral sub-features of each behavioral sub-matrix in the cheating matrix Z, and constructing a feature array for the same test subject;

[0029] Calculate the behavioral reference value for each feature array;

[0030]

[0031] Among them, sim(s i1 ,s j1 ) represents the similarity value between the i1th behavioral sub-feature and the j1th behavioral sub-feature in the corresponding feature array, and i1≠j1; sim(s i1 ,s j1 ) max Indicates that from all sim(s i1 ,s j1 ); m represents the maximum similarity value obtained from all sim(s i1 ,s j1 ) to obtain satisfaction The number of behavioral sub-features; n1 represents the total number of behavioral sub-features in the corresponding feature array, and is consistent with the number of behavioral sub-matrices in the corresponding row;

[0032] When the behavior reference value is greater than or equal to the preset value, all sim(s) corresponding to the same test subject are selected. i1 ,s j1 ) to extract the largest number of consistent sub-behaviors;

[0033] When the behavior reference value is less than the preset value, all sim(s) corresponding to the same test subject are i1 ,s j1 ) to sort by size and get the first sim(s i1 ,s j1 ) corresponds to the existing matching sub-behavior, where [] represents the rounding symbol;

[0034] According to the last sub-behavior corresponding to each row in the cheating matrix and the subject attributes of the examination subject in each row, the cheating characteristics of the examinee for different examination subjects are obtained.

[0035] Preferably, according to the first sorting result and the second sorting result, locking the key supervision features and the non-key supervision features includes:

[0036] Determine the supervision value J for the same cheating feature in the same examination venue;

[0037]

[0038] Where u1 represents the number of occurrences of the same cheating feature determined based on the first sorting result; u2 represents the feature importance of the same cheating feature determined based on the second sorting result; sum(u1) represents the sum of all occurrences; sum(u2) represents the sum of all feature importances, which is 1.

[0039] Establish a locking function S(J,u1,u2) for the same cheating feature based on the supervision value, number of occurrences, and feature importance;

[0040] Determine a lock symbol corresponding to the same cheating feature according to the lock function S(J, u1, u2), wherein the lock symbol is composed of three characters, and the first character matches the supervision level of J, the second character matches the quantity level of u1, and the third character matches the importance level of u2;

[0041] The locking symbol is allocated based on a symbol allocation database to obtain key supervision features and non-key supervision features.

[0042] Preferably, based on the invigilation mode, the corresponding first inspection mode is adjusted to obtain an adapted second inspection mode, including:

[0043] Identify the invigilators present in the same examination venue;

[0044] When the number of invigilators is 0, the strictness of the invigilation mode is determined to be 0, and the first inspection mode is used as the second inspection mode;

[0045] When the number of invigilators is not 0, obtaining the current strictness of the invigilation method and the inspection strictness of the first inspection method;

[0046] When the current strictness is greater than or equal to the preset strictness of the corresponding examination venue, the corresponding first inspection mode is adjusted downward to obtain a second inspection mode;

[0047] When the current strictness is less than the preset strictness of the corresponding examination venue, the corresponding first inspection method is adjusted upward to obtain a second inspection method.

[0048] Preferably, the corresponding first inspection mode is adjusted downward to obtain a second inspection mode, including:

[0049] Obtaining a first degree difference between the current strictness and a preset strictness, and simultaneously obtaining the inspection strictness of the first inspection mode;

[0050] Extracting the set inspection indicators that exceed the standard values ​​of the standard inspection indicators in the first inspection mode;

[0051] Determine a first comparison degree corresponding to all set inspection indicators, and when the first comparison degree is greater than the first degree difference, perform an indicator combination on all set inspection indicators in the first inspection mode, and a second comparison degree corresponding to the indicator combination is less than or equal to the first degree difference;

[0052] When there is one indicator combination, the set inspection indicator in the indicator combination is replaced with the corresponding standard inspection indicator;

[0053] When there are multiple indicator combinations, obtain the cumulative sum of the indicator weights of each indicator combination, and replace the set inspection indicator in the indicator combination corresponding to the minimum cumulative sum with the corresponding standard inspection indicator;

[0054] When the first comparison degree is less than or equal to the first degree difference, the set inspection indicator is replaced with the corresponding standard inspection indicator.

[0055] Preferably, when the number of invigilators is not 0, obtaining the current strictness of the invigilation method and the inspection strictness of the first inspection method includes:

[0056] Extract the historical proctoring status of each proctor at the same examination venue, obtain the historical proctoring coefficient of each proctor, and obtain the current strictness;

[0057] The inspection index corresponding to each cheating feature in the first inspection method is extracted, and the inspection strictness is obtained according to the index coefficient of each inspection index.

[0058] The present invention proposes a 5G-based intelligent inspection and proctoring system, comprising:

[0059] A personnel acquisition module is used to construct an examination room inspection map and obtain the examinees of each examination venue in the examination room inspection map;

[0060] An information determination module is used to retrieve a set of personnel information related to each examinee from a personnel database;

[0061] A feature acquisition module is used to construct a cheating matrix corresponding to the examinee based on the personnel information set and obtain the cheating features of the examinee;

[0062] A sorting module is used to obtain all cheating features of the corresponding examination venue based on the obtained cheating features, and sort the cheating features of the same type from the largest to the smallest number to obtain a first sorting result;

[0063] At the same time, according to the pre-set weight allocation of different cheating characteristics, the characteristics of the corresponding examination venues are ranked from large to small in importance to obtain a second ranking result;

[0064] A feature locking module, configured to lock key supervision features and non-key supervision features according to the first sorting result and the second sorting result, and obtain a first inspection mode;

[0065] The inspection determination module is used to obtain the historical invigilation status of the invigilators assigned to each examination venue, obtain the invigilation method of the corresponding examination venue, and adjust the corresponding first inspection method based on the invigilation method to obtain an adapted second inspection method;

[0066] The intelligent inspection and supervision module is used to obtain a third inspection method according to the second inspection method of all examination places in the examination room inspection map and the location deployment of the examination places in the examination room inspection map, so as to realize intelligent inspection and supervision.

[0067] The present invention proposes an electronic device, comprising: a processor and a memory, wherein the memory is used to store one or more program instructions; and the processor is used to execute the one or more program instructions to perform any one of the methods described.

[0068] Other features and advantages of the present invention will be described in the following description, and in part will become apparent from the description, or will be understood by practicing the present invention. The purposes and other advantages of the present invention can be realized and obtained by the structures particularly pointed out in the written description, claims, and drawings.

[0069] The technical solution of the present invention is further described in detail below through the accompanying drawings and embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0070] The accompanying drawings are used to provide a further understanding of the present invention and constitute a part of the specification. Together with the embodiments of the present invention, they are used to explain the present invention and do not constitute a limitation of the present invention. In the accompanying drawings:

[0071] Figure 1 This is a flow chart of a 5G-based intelligent inspection and proctoring method in an embodiment of the present invention;

[0072] Figure 2 This is a structural diagram of a 5G-based intelligent inspection and supervision system in an embodiment of the present invention. DETAILED DESCRIPTION

[0073] The preferred embodiments of the present invention are described below with reference to the accompanying drawings. It should be understood that the preferred embodiments described herein are only used to illustrate and explain the present invention, and are not used to limit the present invention.

[0074] The present invention provides a 5G-based intelligent inspection and supervision method, such as Figure 1 As shown, including:

[0075] Step 1: Construct an examination room inspection map and obtain the examiners of each examination venue in the examination room inspection map;

[0076] Step 2: Retrieve the personnel information set related to each examinee from the personnel database;

[0077] Step 3: Based on the personnel information set, construct a cheating matrix corresponding to the examinee and obtain the cheating characteristics of the examinee;

[0078] Step 4: Based on the obtained cheating features, all cheating features of the corresponding examination venue are obtained, and similar cheating features are sorted from largest to smallest in number to obtain a first sorting result. According to the pre-set weight allocation for different cheating features, the features of the corresponding examination venues are sorted from largest to smallest in importance to obtain a second sorting result;

[0079] Step 5: Based on the first sorting result and the second sorting result, lock the key supervision features and the non-key supervision features to obtain the first inspection method;

[0080] Step 6: Obtain the historical invigilation status of the invigilators assigned to each examination venue, obtain the invigilation method of the corresponding examination venue, and adjust the corresponding first inspection method based on the invigilation method to obtain an adapted second inspection method;

[0081] Step 7: According to the second inspection method of all the examination venues in the examination venue inspection map and the venue deployment of the examination venues in the examination venue inspection map, a third inspection method is obtained to realize intelligent inspection and supervision.

[0082] Preferably, the personnel information set includes the examination status of the corresponding examinee during the historical examination process, wherein the examination status includes whether the corresponding examinee cheated and the cheating patterns corresponding to each examination subject in each historical examination, wherein the cheating patterns are obtained by extracting and analyzing the local movements of the examinee.

[0083] In this embodiment, the examination room inspection map is based on the examination subjects in different classrooms in different teaching buildings at different time periods and the examinees who refer to the examination subjects, which are planned in advance by the Academic Affairs Office, and the examination venue is the examination classroom. Therefore, the examination room inspection map can be obtained, and then the examinees in the corresponding examination venue can also be obtained. The examination position of each examinee is predetermined, and the examination number and admission ticket will be set in advance. The examinee will find the corresponding examination position to prepare for the examination.

[0084] In this embodiment, the personnel database includes the historical examination information of several examinees. For example, before the college entrance examination, each examinee (student) will go through several monthly examinations, mock examinations, etc., and there will be corresponding historical supervision information for each examination in which the examinee participates. The historical supervision information includes the local movements of the corresponding examinee during the examination, and the local movements include the examinee's facial expressions, body movements, etc.

[0085] In this embodiment, the personnel information set refers to the supervision information obtained for the same examinee in different examination subjects, so as to determine the cheating behavior of the same examinee in different examination subjects. That is, the number of examination subjects is used as the number of row vectors, and the number of examinations for the same examination subject is used as the number of column vectors. In addition, each examination subject is filled with cheating-related information according to the set indicators during each examination, so as to obtain a cheating matrix.

[0086] In this embodiment, the cheating characteristics are the cheating behavior characteristics of the corresponding examinees in different subjects. For example, subject A is a liberal arts subject, and the examinee will continuously lower his head to copy. Subject B is a science subject, and the examinee will rotate left and right to ask questions when the exam is about to end.

[0087] In this embodiment, since the corresponding personnel in different examination rooms are different, it is necessary to comprehensively determine the cheating situation in the corresponding examination venue based on the attributes of the personnel and the examination subjects. For example, the number of occurrences of cheating feature 1 obtained is 4, the number of occurrences of cheating feature 2 is 5, and the number of occurrences of cheating feature 3 is 1. At this time, the first sorting result is: cheating feature 2, cheating feature 1, cheating feature 3.

[0088] In this embodiment, for example, the weight of cheating feature 1 is a1, the weight of cheating feature 2 is a2, and the weight of cheating feature 3 is a3. At this time, the importance of each cheating weight is determined by the set weight size, a1>a2>a3.

[0089] In this embodiment, the first ranking result is a ranking of the number of similar cheating features, and the second ranking result is a ranking of the importance of the features.

[0090] In this embodiment, the key supervision features and non-key supervision features are related to the quantity and feature importance. The greater the quantity, the greater the feature importance, and the more likely it is to be used as a key supervision feature.

[0091] In this embodiment, the first inspection method is obtained based on a feature-inspection database, which includes features of different combinations and inspection methods matching the features of the combinations. The inspection method is mainly for an inspection of the examination venue (classroom).

[0092] For example, there are key supervision features 1 and 2, and non-key supervision feature 3, and inspection method 1 is obtained by matching from the database.

[0093] In this embodiment, the historical invigilation status refers to the invigilation behavior of the corresponding invigilator during the invigilation examination, such as inspection behavior in the classroom, searching for cheat sheets, etc.

[0094] In this embodiment, the invigilation method mainly refers to the number of invigilators and the strictness of the invigilators' invigilation.

[0095] In this embodiment, the first inspection method is to control the inspection equipment for examination to realize intelligent supervision, and the invigilation method is a manual invigilation situation. The second inspection method is mainly obtained by adjusting the first inspection method according to the invigilation method. For example, if the invigilation method is very strict, then the first inspection method can be adjusted in a less strict direction.

[0096] In this embodiment, the second inspection method is mainly aimed at the examination venue, and the third inspection method is mainly aimed at the entire examination room.

[0097] The beneficial effects of the above technical solution are: by obtaining a set of personnel information, the supervision characteristics of each examination venue are determined, and then the inspection method for the examination venue is obtained, and the inspection method is adjusted by obtaining the historical invigilation status of the invigilators, so as to achieve reasonable and strict supervision of different venues, realize intelligent inspection and supervision, reduce the consumption of human invigilation and improve the intelligence of invigilation.

[0098] The present invention provides a 5G-based intelligent inspection and proctoring method, which obtains a third inspection mode and implements the intelligent inspection and proctoring process, further comprising:

[0099] Determining the examination inspection equipment at the examination site based on the third inspection method;

[0100] Controlling the examination inspection equipment to capture videos and conduct behavior analysis on the persons to be monitored during the examination;

[0101] Remote proctoring of examinees at the same examination location based on cameras installed at each examination location;

[0102] When the person to be monitored is a first person who continues to answer questions at the examination location, a first event of interest for the first person is determined based on the corresponding behavior analysis result, and remote proctoring information for the first person is obtained to obtain a second event of interest;

[0103] Based on the first event of concern and the second event of concern, extracting and compressing a first abnormal video segment, and transmitting it to a remote end for storage and a first early warning reminder;

[0104] When the monitored person leaves the examination location to answer questions, the second person is dynamically tracked and monitored to obtain a third event of interest, and a second abnormal video segment is extracted and compressed, and transmitted to a remote terminal for storage and a second warning reminder;

[0105] When the person to be monitored is an invigilator, abnormal invigilation behavior is determined based on the behavior analysis result corresponding to the invigilator and transmitted to the remote end for a third early warning reminder.

[0106] In this embodiment, based on 5G network transmission and high concurrency, joint debugging and inspection management of examination-related equipment are carried out before the exam, and video analysis of examinees / invigilators' entry, exit, suspected cheating, etc. during the exam is carried out. Key videos during the exam are condensed and stored in the cloud, allowing users to only focus on videos of key events, achieving a management closed loop of discovering problems, analyzing problems, and solving problems, creating a safe, fair, and just examination environment for examinees, and comprehensively improving the examination management level, command ability, and service quality.

[0107] 5G's high bandwidth allows for remote video monitoring and inspection using mobile security cameras, built-in computer cameras, or external cameras, ensuring fair and impartial exams. High-definition video cameras placed directly in front of test takers can monitor and analyze any abnormalities in real time, providing timely warnings to invigilators. Remote video streams can be used to analyze abnormal behavior in real time to determine if it indicates cheating. Screenshots and video recordings of the frequency of abnormal behavior can be taken and archived for invigilators to review and collect evidence.

[0108] In this embodiment, the events of interest are mainly related to cheating behaviors of people.

[0109] The beneficial effect of the above technical solution is: by performing different analyses on different personnel during the inspection according to the third inspection method, and making reasonable and effective early warning reminders, the efficiency of invigilation and the fairness of the examination are guaranteed.

[0110] The present invention provides a 5G-based intelligent inspection and proctoring method, which constructs a cheating matrix corresponding to the examinee based on the personnel information set and obtains the cheating characteristics of the examinee, including:

[0111] Analyze the personnel information set of the same examinee and construct the cheating matrix Z of the corresponding examinee;

[0112]

[0113] Where Z represents the cheating matrix of the corresponding examinee; A1 represents the behavior sub-matrix of the corresponding examinee in the first examination of subject A obtained according to the analysis results; An1 represents the behavior sub-matrix of the corresponding examinee in the n1th examination of subject A obtained according to the analysis results; B1 represents the behavior sub-matrix of the corresponding examinee in the first examination of subject B obtained according to the analysis results; Bn1 represents the behavior sub-matrix of the corresponding examinee in the n1th examination of subject B obtained according to the analysis results;

[0114] Obtaining behavioral sub-features of each behavioral sub-matrix in the cheating matrix Z, and constructing a feature array for the same test subject;

[0115] Calculate the behavioral reference value for each feature array;

[0116]

[0117] Among them, sim(s i1 ,s j1 ) represents the similarity value between the i1th behavioral sub-feature and the j1th behavioral sub-feature in the corresponding feature array, and i1≠j1; sim(s i1 ,s j1 )max Indicates that from all sim(s i1 ,s j1 ); m represents the maximum similarity value obtained from all sim(s i1 ,s j1 ) to obtain satisfaction The number of behavioral sub-features; n1 represents the total number of behavioral sub-features in the corresponding feature array, and is consistent with the number of behavioral sub-matrices in the corresponding row;

[0118] When the behavior reference value is greater than or equal to the preset value, all sim(s) corresponding to the same test subject are selected. i1 ,s j1 ) to extract the largest number of consistent sub-behaviors;

[0119] When the behavior reference value is less than the preset value, all sim(s) corresponding to the same test subject are i1 ,s j1 ) to sort by size and get the first sim(s i1 ,s j1 ) corresponds to the existing matching sub-behavior, where [] represents the rounding symbol;

[0120] According to the last sub-behavior corresponding to each row in the cheating matrix and the subject attributes of the examination subject in each row, the cheating characteristics of the examinee for different examination subjects are obtained.

[0121] In this embodiment, the numbers of rows and columns of different behavioral sub-matrices under the same examination subject are equal, and the numbers of rows and columns in the same behavioral sub-matrix may be equal or unequal.

[0122] In this embodiment, the behavior sub-matrix includes expressions, behavior postures, etc. existing at different moments.

[0123] In this embodiment, the feature array is the corresponding row content in the cheating matrix Z.

[0124] In this embodiment, the preset value is pre-set.

[0125] In this embodiment, according to sim(s i1 ,s j1 ) can get: s1 and s2, s1 and s2, ..., s1 and s n1 , s2 and s3, s2 and s4, ..., s2 and s n1 ,...,s n1-1 With s n1 ;

[0126] Then, similar pairs with similarity values ​​higher than the preset value can be screened out, for example, s1 and s2, s3, sn1 Meet similar screening conditions, s n1-1 With s n1 The similarity screening condition is met. At this time, there are 3 similar pairs corresponding to s1, that is, the number of extractions is 3. At this time, the behavior corresponding to s1 is regarded as a consistent sub-behavior.

[0127] In this embodiment, non-overlapping sorting refers to s1 and s2, s1 and s2, ..., s1 and s n1 , s2 and s3, s2 and s4, ..., s2 and s n1 ,...,s n1-1 With s n1 Sort by similarity values.

[0128] In this embodiment, the matching sub-behavior refers to the previous The corresponding behavior sub-feature is regarded as a matching sub-behavior.

[0129] The beneficial effects of the above technical solution are: by analyzing the personnel information set, constructing the cheating matrix, and by constructing the feature array and calculating the behavior reference value, the final determination of the sub-behavior is achieved, providing a basis for the subsequent acquisition of cheating features, and ensuring the efficiency of the inspection and the fairness of the examination.

[0130] The present invention provides a 5G-based intelligent inspection and proctoring method, which locks key supervision features and non-key supervision features according to a first sorting result and a second sorting result, including:

[0131] Determine the supervision value J for the same cheating feature in the same examination venue;

[0132]

[0133] Where u1 represents the number of occurrences of the same cheating feature determined based on the first sorting result; u2 represents the feature importance of the same cheating feature determined based on the second sorting result; sum(u1) represents the sum of all occurrences; sum(u2) represents the sum of all feature importances, which is 1.

[0134] Establish a locking function S(J,u1,u2) for the same cheating feature based on the supervision value, number of occurrences, and feature importance;

[0135] Determine a lock symbol corresponding to the same cheating feature according to the lock function S(J, u1, u2), wherein the lock symbol is composed of three characters, and the first character matches the supervision level of J, the second character matches the quantity level of u1, and the third character matches the importance level of u2;

[0136] The locking symbol is allocated based on a symbol allocation database to obtain key supervision features and non-key supervision features.

[0137] In this embodiment, for example, the supervision level is @1, the quantity level is #2, and the importance level is ￥3. At this time, @, #, and ￥ represent corresponding symbols, and 1, 2, and 3 represent corresponding levels. Therefore, the obtained locking symbol is @1#2￥3.

[0138] In this embodiment, the symbol allocation database includes different symbol combinations and emphasis or non-emphasis setting labels that match the symbol combinations, and allocation is then achieved according to the labels.

[0139] In this embodiment, cheating feature 1 is labeled 01, cheating feature 2 is labeled 01, and cheating feature 3 is labeled 02, wherein label 02 is an assigned key label and label 01 is an assigned non-key label.

[0140] The beneficial effect of the above technical solution is that according to the supervision value, number of occurrences and feature importance of the same cheating feature, the key supervision features and non-key supervision features can be effectively determined, providing a basis for subsequent inspections and further ensuring the fairness of the examination.

[0141] The present invention provides a 5G-based intelligent inspection and proctoring method, which adjusts the corresponding first inspection mode based on the proctoring mode to obtain an adapted second inspection mode, including:

[0142] Identify the invigilators present in the same examination venue;

[0143] When the number of invigilators is 0, the strictness of the invigilation mode is determined to be 0, and the first inspection mode is used as the second inspection mode;

[0144] When the number of invigilators is not 0, obtaining the current strictness of the invigilation method and the inspection strictness of the first inspection method;

[0145] When the current strictness is greater than or equal to the preset strictness of the corresponding examination venue, the corresponding first inspection mode is adjusted downward to obtain a second inspection mode;

[0146] When the current strictness is less than the preset strictness of the corresponding examination venue, the corresponding first inspection method is adjusted upward to obtain a second inspection method.

[0147] In this embodiment, if there are 0 invigilators, the venue needs to be fully intelligently patrolled and supervised. If there are invigilators, manual and intelligent supervision will be combined.

[0148] In this embodiment, the current strictness of the invigilation method depends on the strictness of the corresponding invigilator, which is generally determined based on the number of rounds of the invigilator and the number of conversations with the examinees in the examination room, the number of times cheat sheets are collected, and voice reminders and messages not to cheat.

[0149] In this embodiment, the strictness of the inspection method is related to the cheating behavior of the examinees in the corresponding place. The more reference value the cheating behavior has, the stricter the corresponding inspection is.

[0150] In this embodiment, the preset strictness is pre-set, and the downward adjustment of the strictness means that the strictness in the first inspection mode can be adjusted to be weaker, and the upward adjustment of the strictness means that the strictness in the first inspection mode can be adjusted to be stronger.

[0151] The beneficial effect of the above technical solution is: by analyzing and comparing the strictness of the invigilator's supervision and the strictness of the first inspection method, the first inspection method can be effectively adjusted downward or upward, ensuring the rationality of manual and intelligent inspections and the fairness of the examination.

[0152] The present invention provides a 5G-based intelligent inspection and proctoring method, which performs a degree-down adjustment on a corresponding first inspection mode to obtain a second inspection mode, including:

[0153] Obtaining a first degree difference between the current strictness and a preset strictness, and simultaneously obtaining the inspection strictness of the first inspection mode;

[0154] Extracting the set inspection indicators that exceed the standard values ​​of the standard inspection indicators in the first inspection mode;

[0155] Determine a first comparison degree corresponding to all set inspection indicators, and when the first comparison degree is greater than the first degree difference, perform an indicator combination on all set inspection indicators in the first inspection mode, and a second comparison degree corresponding to the indicator combination is less than or equal to the first degree difference;

[0156] When there is one indicator combination, the set inspection indicator in the indicator combination is replaced with the corresponding standard inspection indicator;

[0157] When there are multiple indicator combinations, obtain the cumulative sum of the indicator weights of each indicator combination, and replace the set inspection indicator in the indicator combination corresponding to the minimum cumulative sum with the corresponding standard inspection indicator;

[0158] When the first comparison degree is less than or equal to the first degree difference, the set inspection indicator is replaced with the corresponding standard inspection indicator.

[0159] In this embodiment, the standard values ​​of the standard inspection indicators are pre-set. For example, in the first inspection method, there are standard inspection indicators 01 and 02 for cheating behavior 1, and a standard inspection indicator 03 for cheating behavior 2. Among them, the value of the standard inspection indicator 01 is b1, the value of the standard inspection indicator 02 is b2, and the value of the standard inspection indicator 03 is b3. At this time, the actual inspection indicator 01 in the first inspection method has a value of c1. At this time, c1 is greater than b1, and the indicator 01 is regarded as the set inspection indicator.

[0160] In this embodiment, the first comparison degree is determined based on all set inspection indicators, which is also the inspection strictness. For example, there is indicator combination 1. At this time, the set inspection indicators in indicator combination 1 are replaced, and the replacement is to replace their corresponding values ​​with standard values.

[0161] In this embodiment, each indicator in the indicator combination has a set weight, so the cumulative sum of the indicator weights can be obtained to obtain the final combination.

[0162] The beneficial effect of the above technical solution is that by determining the indicator combination and performing indicator replacement, the efficiency of the inspection is effectively achieved.

[0163] The present invention provides a 5G-based intelligent inspection and proctoring method, which, when the number of invigilators is not zero, obtains the current strictness of the proctoring method and the inspection strictness of the first inspection method, including:

[0164] Extract the historical proctoring status of each proctor at the same examination venue, obtain the historical proctoring coefficient of each proctor, and obtain the current strictness;

[0165] The inspection index corresponding to each cheating feature in the first inspection method is extracted, and the inspection strictness is obtained according to the index coefficient of each inspection index.

[0166] In this embodiment, the historical invigilation coefficient ranges from 0 to 1. The larger the coefficient, the higher the corresponding strictness. For example, the historical invigilation status of invigilator 1 is a free state throughout the whole process, the corresponding coefficient is 0, and the corresponding current strictness is 0.01.

[0167] In this embodiment, cheating feature 1 corresponds to inspection indicators 01 and 02. At this time, the indicator coefficient of inspection indicator 01 is 0.2, the inspection coefficient of indicator 02 is 0.3, and the corresponding strictness is 0.32.

[0168] The beneficial effect of the above technical solution is: by obtaining the degree of strictness according to the coefficient, an effective basis is provided for the subsequent acquisition of the inspection method, thereby ensuring the efficiency of the inspection.

[0169] The present invention proposes a 5G-based intelligent inspection and supervision system. Figure 2 As shown, including:

[0170] A personnel acquisition module is used to construct an examination room inspection map and obtain the examinees of each examination venue in the examination room inspection map;

[0171] An information determination module is used to retrieve a set of personnel information related to each examinee from a personnel database;

[0172] A feature acquisition module is used to construct a cheating matrix corresponding to the examinee based on the personnel information set and obtain the cheating features of the examinee;

[0173] A sorting module is used to obtain all cheating features of the corresponding examination venue based on the obtained cheating features, and sort the cheating features of the same type from the largest to the smallest number to obtain a first sorting result;

[0174] At the same time, according to the pre-set weight allocation of different cheating characteristics, the characteristics of the corresponding examination venues are ranked from large to small in importance to obtain a second ranking result;

[0175] A feature locking module, configured to lock key supervision features and non-key supervision features according to the first sorting result and the second sorting result, and obtain a first inspection mode;

[0176] The inspection determination module is used to obtain the historical invigilation status of the invigilators assigned to each examination venue, obtain the invigilation method of the corresponding examination venue, and adjust the corresponding first inspection method based on the invigilation method to obtain an adapted second inspection method;

[0177] The intelligent inspection and supervision module is used to obtain a third inspection method according to the second inspection method of all examination places in the examination room inspection map and the location deployment of the examination places in the examination room inspection map, so as to realize intelligent inspection and supervision.

[0178] The present invention proposes an electronic device, comprising: a processor and a memory, wherein the memory is used to store one or more program instructions; and the processor is used to execute the one or more program instructions to perform any one of the methods described.

[0179] The beneficial effects of the above technical solution are: by obtaining a set of personnel information, the supervision characteristics of each examination venue are determined, and then the inspection method for the examination venue is obtained, and the inspection method is adjusted by obtaining the historical invigilation status of the invigilators, so as to achieve reasonable and strict supervision of different venues, realize intelligent inspection and supervision, reduce the consumption of human invigilation and improve the intelligence of invigilation.

[0180] Obviously, those skilled in the art may make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if such changes and modifications fall within the scope of the claims and their equivalents, the present invention is intended to include such changes and modifications.

Claims

1. A 5G-based intelligent inspection and proctoring method, characterized in that: include: Step 1: Construct an examination room inspection map and obtain the examiners of each examination venue in the examination room inspection map; Step 2: Retrieve the personnel information set related to each examinee from the personnel database; Step 3: Based on the personnel information set, construct a cheating matrix corresponding to the examinee and obtain the cheating characteristics of the examinee; Step 4: Based on the obtained cheating features, all cheating features of the corresponding examination venue are obtained, and similar cheating features are sorted from the largest to the smallest number to obtain a first sorting result; At the same time, according to the pre-set weight allocation of different cheating characteristics, the characteristics of the corresponding examination venues are ranked from large to small in importance to obtain a second ranking result; Step 5: Based on the first sorting result and the second sorting result, lock the key supervision features and the non-key supervision features to obtain the first inspection method; Step 6: Obtain the historical invigilation status of the invigilators assigned to each examination venue, obtain the invigilation method of the corresponding examination venue, and adjust the corresponding first inspection method based on the invigilation method to obtain an adapted second inspection method; Step 7: Obtain a third inspection method according to the second inspection method of all examination venues in the examination venue inspection map and the location deployment of the examination venues in the examination venue inspection map, thereby realizing intelligent inspection and proctoring; Wherein, step 3 includes: Analyze the personnel information set of the same examinee and construct the cheating matrix Z of the corresponding examinee; Wherein, An1 represents the behavior sub-matrix of the corresponding examinee in the n1th examination of subject A obtained according to the analysis results; Bn1 represents the behavior sub-matrix of the corresponding examinee in the n1th examination of subject B obtained according to the analysis results; Obtaining behavioral sub-features of each behavioral sub-matrix in the cheating matrix Z, and constructing a feature array for the same test subject; Calculate the behavioral reference value X for each feature array; Among them, sim(s i1 ,s j1 ) represents the similarity value between the i1th behavioral sub-feature and the j1th behavioral sub-feature in the corresponding feature array, and i1≠j1; sim(s i1 ,s j1 ) max Indicates that from all sim(s i1 ,s j1 ); m represents the maximum similarity value obtained from all sim(s i1 ,s j1 ) to obtain satisfaction The number of behavioral sub-features; n1 represents the total number of behavioral sub-features in the corresponding feature array, and is consistent with the number of behavioral sub-matrices in the corresponding row; When the behavior reference value is greater than or equal to the preset value, all sim(s) corresponding to the same test subject are selected. i1 ,s j1 ) to extract the largest number of consistent sub-behaviors; When the behavior reference value is less than the preset value, all sim(s) corresponding to the same test subject are i1 ,s j1 ) to sort by size and get the first sim(s i1 ,s j1 ) corresponds to the existing matching sub-behavior, where [] represents the rounding symbol; According to the last sub-behavior corresponding to each row in the cheating matrix and the subject attributes of the examination subject in each row, the cheating characteristics of the examinee for different examination subjects are obtained.

2. The 5G-based intelligent inspection and proctoring method according to claim 1, characterized in that: The personnel information set includes the examination status of the corresponding examinee during the historical examination process, wherein the examination status includes whether the corresponding examinee cheated and the cheating pattern corresponding to each examination subject in each historical examination, wherein the cheating pattern is obtained by extracting and analyzing the local movements of the examinee.

3. The 5G-based intelligent inspection and proctoring method according to claim 1, characterized in that: The third inspection method is used to realize intelligent inspection and proctoring, which also includes: Determining the examination inspection equipment at the examination site based on the third inspection method; Controlling the examination inspection equipment to capture videos and conduct behavior analysis on the persons to be monitored during the examination; Remote proctoring of examinees at the same examination location based on cameras installed at each examination location; When the person to be monitored is a first person who continues to answer questions at the examination location, a first event of interest for the first person is determined based on the corresponding behavior analysis result, and remote proctoring information for the first person is obtained to obtain a second event of interest; Based on the first event of concern and the second event of concern, extracting and compressing a first abnormal video segment, and transmitting it to a remote end for storage and a first early warning reminder; When the person to be monitored is the second person who has left the examination location to answer questions, the second person is dynamically tracked and monitored to obtain a third event of interest, and a second abnormal video segment is extracted and compressed, and transmitted to a remote terminal for storage and a second warning reminder; When the person to be monitored is an invigilator, abnormal invigilation behavior is determined based on the behavior analysis result corresponding to the invigilator and transmitted to the remote end for a third early warning reminder.

4. The 5G-based intelligent inspection and proctoring method according to claim 1, characterized in that: According to the first sorting result and the second sorting result, the key supervision features and non-key supervision features are locked, including: Determine a supervision value J for the same cheating feature in the same examination venue based on the number of occurrences u1 of the same cheating feature determined by the first sorting result and the feature importance u2 of the same cheating feature determined by the second sorting result; According to the supervision value J, the number of occurrences u1 and the feature importance u2, a locking function S(J,u1,u2) for the same cheating feature is established; Determine a lock symbol corresponding to the same cheating feature according to the lock function S(J, u1, u2), wherein the lock symbol is composed of three characters, and the first character matches the supervision level of J, the second character matches the quantity level of u1, and the third character matches the importance level of u2; The locking symbol is allocated based on a symbol allocation database to obtain key supervision features and non-key supervision features.

5. The 5G-based intelligent inspection and proctoring method according to claim 1, characterized in that: Based on the invigilation mode, the corresponding first inspection mode is adjusted to obtain an adapted second inspection mode, including: Identify the invigilators present in the same examination venue; When the number of invigilators is 0, the strictness of the invigilation mode is determined to be 0, and the first inspection mode is used as the second inspection mode; When the number of invigilators is not 0, obtaining the current strictness of the invigilation method and the inspection strictness of the first inspection method; When the current strictness is greater than or equal to the preset strictness of the corresponding examination venue, the corresponding first inspection mode is adjusted downward to obtain a second inspection mode; When the current strictness is less than the preset strictness of the corresponding examination venue, the corresponding first inspection method is adjusted upward to obtain a second inspection method.

6. The 5G-based intelligent inspection and proctoring method according to claim 5, characterized in that: The corresponding first inspection mode is adjusted downward to obtain a second inspection mode, including: Obtaining a first degree difference between the current strictness and a preset strictness, and simultaneously obtaining the inspection strictness of the first inspection mode; Extracting the set inspection indicators that exceed the standard values ​​of the standard inspection indicators in the first inspection mode; Determine a first comparison degree corresponding to all set inspection indicators, and when the first comparison degree is greater than the first degree difference, perform an indicator combination on all set inspection indicators in the first inspection mode, and a second comparison degree corresponding to the indicator combination is less than or equal to the first degree difference; When there is one indicator combination, the set inspection indicator in the indicator combination is replaced with the corresponding standard inspection indicator; When there are multiple indicator combinations, obtain the cumulative sum of the indicator weights of each indicator combination, and replace the set inspection indicator in the indicator combination corresponding to the minimum cumulative sum with the corresponding standard inspection indicator; When the first comparison degree is less than or equal to the first degree difference, the set inspection indicator is replaced with the corresponding standard inspection indicator.

7. The 5G-based intelligent inspection and proctoring method according to claim 6, characterized in that: When the number of invigilators is not 0, obtaining the current strictness of the invigilation method and the inspection strictness of the first inspection method includes: Extract the historical proctoring status of each proctor at the same examination venue, obtain the historical proctoring coefficient of each proctor, and obtain the current strictness; The inspection index corresponding to each cheating feature in the first inspection method is extracted, and the inspection strictness is obtained according to the index coefficient of each inspection index.

8. A 5G-based intelligent inspection and proctoring system, characterized in that: include: A personnel acquisition module is used to construct an examination room inspection map and obtain the examinees of each examination venue in the examination room inspection map; An information determination module is used to retrieve a set of personnel information related to each examinee from a personnel database; A feature acquisition module is used to construct a cheating matrix corresponding to the examinee based on the personnel information set and obtain the cheating features of the examinee; A sorting module is used to obtain all cheating features of the corresponding examination venue based on the obtained cheating features, and sort the cheating features of the same type from the largest to the smallest number to obtain a first sorting result; At the same time, according to the pre-set weight allocation of different cheating characteristics, the characteristics of the corresponding examination venues are ranked from large to small in importance to obtain a second ranking result; A feature locking module, configured to lock key supervision features and non-key supervision features according to the first sorting result and the second sorting result, and obtain a first inspection mode; The inspection determination module is used to obtain the historical invigilation status of the invigilators assigned to each examination venue, obtain the invigilation method of the corresponding examination venue, and adjust the corresponding first inspection method based on the invigilation method to obtain an adapted second inspection method; An intelligent patrol and invigilation module is configured to obtain a third patrol mode according to the second patrol mode of all examination venues in the examination venue patrol map and the venue deployment of the examination venues in the examination venue patrol map, thereby realizing intelligent patrol and invigilation; Among them, the feature acquisition module is used to: Analyze the personnel information set of the same examinee and construct the cheating matrix Z of the corresponding examinee; Wherein, An1 represents the behavior sub-matrix of the corresponding examinee in the n1th examination of subject A obtained according to the analysis results; Bn1 represents the behavior sub-matrix of the corresponding examinee in the n1th examination of subject B obtained according to the analysis results; Obtaining behavioral sub-features of each behavioral sub-matrix in the cheating matrix Z, and constructing a feature array for the same test subject; Calculate the behavioral reference value X for each feature array; Among them, sim(s i1 ,s j1 ) represents the similarity value between the i1th behavioral sub-feature and the j1th behavioral sub-feature in the corresponding feature array, and i1≠j1; sim(s i1 ,s j1 ) max Indicates that from all sim(s i1 ,s j1 ); m represents the maximum similarity value obtained from all sim(s i1 ,s j1 ) to obtain satisfaction The number of behavioral sub-features; n1 represents the total number of behavioral sub-features in the corresponding feature array, and is consistent with the number of behavioral sub-matrices in the corresponding row; When the behavior reference value is greater than or equal to the preset value, all sim(s) corresponding to the same test subject are selected. i1 ,s j1 ) to extract the largest number of consistent sub-behaviors; When the behavior reference value is less than the preset value, all sim(s) corresponding to the same test subject are i1 ,s j1 ) to sort by size and get the first sim(s i1 ,s j1 ) corresponds to the existing matching sub-behavior, where [] represents the rounding symbol; According to the last sub-behavior corresponding to each row in the cheating matrix and the subject attributes of the examination subject in each row, the cheating characteristics of the examinee for different examination subjects are obtained.

9. An electronic device, characterized in that: include: a processor and a memory, the memory being configured to store one or more program instructions; The processor is configured to execute one or more program instructions to perform the method according to any one of claims 1 to 7.

Images