Offline examination auxiliary system

Abstract

The invention discloses an offline examination auxiliary system, and belongs to the technical field of offline examination auxiliary equipment, and the system comprises an examination room camera which is used for carrying out the video shooting of an examination room; the acquisition module is used for acquiring examinee registration information and examination requirement information; the detection module is used for performing examination room seat detection on the video key frame in the examination preparation stage and performing examinee face detection and recognition on the video key frame in the examination entrance stage; the processing module is used for arranging examination room seats and generating examination information according to the examinee registration information, the examination requirement information, the detected seats and / or the detected and recognized face information, or checking examinees in an examination entering stage and generating examinee checking results; and the sending module is used for sending the examination information and the examinee checking result generated by the processing module to the invigilator terminal and / or the examinee terminal according to the set requirement at the set time. The method can reduce the manual dependence in the examination process, and guarantees the smooth completion of the examination.

Description

Technical Field

[0001] The present invention belongs to the technical field of offline examination auxiliary equipment, and in particular relates to an offline examination auxiliary system. Background Art

[0002] The traditional offline multi-examination room examination process has the following shortcomings: 1. During the pre-examination preparation, it is impossible to automatically obtain the actual number and layout of seats in each examination room. Non-dedicated examination rooms are usually used for classes, and the number and layout of seats may be adjusted by teachers and students, which is different from the planned number and layout. During the examination, the number of available seats in the examination room and the number of assigned candidates are inconsistent, and manual allocation of examination seats is required.

[0003] 2. In the pre-exam preparation, the examination staff needs to arrange the examination room, post paper examination reminders around the examination room for candidates to refer to, manually select the examination seats in the examination room, and stick paper table stickers on each examination seat to display the seat number and candidate information; after the examination, these paper materials need to be cleaned up, which consumes manpower and material resources;

[0004] 3. Candidates cannot intuitively understand the layout of the examination room and can only find the examination room and seats in the examination room based on their admission tickets; during the examination, the invigilators manually count the number of candidates and the number of absentees. The above process may result in problems such as candidates going to the wrong examination room, sitting in the wrong seat, and invigilator statistics errors;

[0005] 4. Lack of global information makes it impossible to efficiently allocate examination seat resources.

[0006] Therefore, there is an urgent need for an examination assistance system that can reduce manual dependence during the examination process and ensure the smooth completion of the examination. Summary of the invention

[0007] In view of the deficiencies in the prior art, the present invention provides an offline examination auxiliary system, which can reduce manual dependence during the examination process and ensure smooth completion of the examination.

[0008] The present invention provides the following technical solutions:

[0009] An offline examination auxiliary system is provided, comprising:

[0010] Examination room camera, used to capture video of the examination room during the examination preparation and examination entrance stages;

[0011] An acquisition module is used to acquire the candidate registration information and examination requirement information input in advance; the examination requirement information includes the examination subjects and the number of candidates for each subject, and the candidate registration information includes: the candidate's facial image and identity information;

[0012] The detection module is used to detect the seats in the examination room based on the key frames of the video during the examination preparation phase, and to detect and recognize the faces of the examinees based on the key frames of the video during the examination entrance phase;

[0013] A processing module, for arranging seats in the examination room and generating examination information, or checking candidates at the examination entrance stage and generating candidate checking results, based on candidate registration information, examination requirement information, and detected seats and / or detected and recognized face information;

[0014] The sending module is used to send the test information generated by the processing module and the examinee verification results to the invigilator terminal and / or the examinee terminal at a set time and according to set requirements.

[0015] Optionally, the detecting of examination room seats on the video key frames in the examination preparation stage specifically includes:

[0016] Using the trained examination room seat detection model, the seat detection is performed on the extracted key frames of the exam preparation stage video to obtain the pixel coordinates of the upper left corner and lower right corner of each seat.

[0017] The upper left pixel coordinates or upper right pixel coordinates of all seats are converted into world coordinates and output as the position coordinates of the seats.

[0018] Optionally, the step of performing examinee face detection and recognition on the key frames of the video during the entrance phase of the examination specifically includes:

[0019] Extract the key frames of the video after all candidates take their seats during the entrance stage of the exam;

[0020] Using the trained face detection and recognition model, perform face detection on the extracted video key frames to obtain the face position coordinates and size; the face position coordinates are the coordinates obtained by converting the pixel coordinates of the face center point into world coordinates;

[0021] Perform feature extraction on the detected face and perform identity recognition on the detected face based on all the registered face images.

[0022] Optionally, a calibration module is also included for obtaining a distortion correction matrix and a perspective transformation matrix, and using the distortion correction matrix to perform distortion correction on the captured video. During the detection process, the perspective transformation matrix is ​​used to convert the detected pixel coordinates into world coordinates.

[0023] Optionally, the processing module includes:

[0024] The seat arrangement subunit is used to obtain the number of available seats and find the target test seat set according to the number of candidates and the tested test seat positions, arrange the target test seat set in sequence, and generate a test seat layout diagram;

[0025] The candidate information generation subunit is used to assign seats to each candidate and generate test information according to the examination room seat layout, candidate registration information and test requirement information;

[0026] The absence and misplacement detection subunit is used to compare the recognized face identity with the face image of the candidates who have registered for the current examination, and detect the candidates who are absent or misplaced;

[0027] The automatic sign-in subunit is used to determine whether there is a misplacement or absence operation based on the detection results of the absence and misplacement detection subunit when the invigilator starts the automatic sign-in operation, and automatically sign in when there is no absence or misplacement, or complete the sign-in through the intervention of the invigilator when there is no misplacement but absence.

[0028] Optionally, the process of obtaining the number of available seats and finding a target test seat set according to the number of examinees and the detected test seat positions is as follows:

[0029] Obtain all available seats according to the detected examination room seat position coordinates;

[0030] Calculate the distance between any two seats based on the location coordinates of each available seat;

[0031] Construct an adjacency matrix of seats based on the distance between any two seats and the set adjacent distance range;

[0032] Using the greedy algorithm, find all non-adjacent point sets from the seat adjacency matrix. This point set is the target test seat set.

[0033] Arrange the target test seat set in sequence and generate a test room seat layout diagram.

[0034] Optionally, the recognized face identity is compared with the face image of the candidates who registered for the current examination to detect the absent or misplaced candidates, and the specific process is as follows:

[0035] Compare all recognized facial identities with the facial images of all candidates who registered for the test to obtain the identities and total number of absent candidates;

[0036] According to the position of each candidate in the examination room seat layout, obtain the theoretical seat number of each candidate valid-seat;

[0037] Based on the position coordinates of the detected face, find the nearest examination seat number of each detected face in the examination room seat layout map as the actual seat number seat;

[0038] Compare each candidate's theoretical seat number valid-seat and actual seat number seat in turn to see if they are consistent. If they are consistent, the candidate is not sitting in the wrong seat; otherwise, the candidate is sitting in the wrong seat.

[0039] Optionally, based on the position coordinates of the detected face, the nearest examination seat number of each detected face is found in the examination room seat layout map as the actual seat number seat, which is specifically determined by the following formula:

[0040] seat=seat i

[0041]

[0042] Among them, (x face ,y face ) is the face position coordinates obtained by face detection, (x i ,y i ,seat i ) is the position coordinates and number of any available seat in the examination room seat layout diagram, and m is the total number of available seats in the examination room seat layout diagram.

[0043] Optionally, the sending of the test information generated by the processing module and the examinee verification result to the invigilator terminal and / or the examinee terminal at a set time and according to set requirements specifically includes:

[0044] Send the examination subject, examination time, admission ticket number, examination room number, seat number and the candidate's position in the examination room seat layout to the candidate's terminal at the designated time;

[0045] At the designated time, the examination subject, time, examination room number, examination room location, number of students taking the examination, personal information of all candidates in the examination room, and the location of all candidates in the examination room in the examination room seat layout are sent to the invigilator's terminal.

[0046] Optionally, it also includes an examination room projector for publicly displaying part of the examination information during the examination entrance stage; part of the examination information includes a seat layout diagram of all candidates in the examination room, and on the seat layout diagram, the position of each candidate is displayed with a set number.

[0047] Compared with the prior art, the present invention has the following beneficial effects:

[0048] The offline examination auxiliary system of the present invention realizes the comprehensive automation and intelligence of the examination preparation, admission and invigilation processes. In the examination preparation stage, the system can accurately detect the examination room seat layout to ensure the accuracy and rationality of the seat arrangement; in the examination admission stage, the system uses face recognition technology to quickly verify the identity of the examinee, effectively prevent substitute examinations and misplacement, and at the same time reduce the work pressure of the invigilator; in addition, the present invention can realize the automatic arrangement of seats for offline multi-examination rooms, candidate guidance and automatic examination sign-in, and can obtain the examination room situation in real time, reduce manual dependence during the examination process, and ensure the smooth completion of the examination. BRIEF DESCRIPTION OF THE DRAWINGS

[0049] Figure 1 It is a schematic diagram of the overall structure of the offline examination auxiliary system of the present invention;

[0050] Figure 2 It is a schematic diagram of the interior of the examination room of the present invention;

[0051] Figure 3 It is a schematic diagram of the seating layout of the examination room of the present invention;

[0052] Figure 4 It is a flow chart of the automatic arrangement of examination room seats of the present invention;

[0053] Figure 5 It is a flow chart of the training of the examination room seat detection model of the present invention;

[0054] Figure 6 It is a seat detection effect diagram provided by the present invention;

[0055] Figure 7 It is a flow chart of obtaining target examination seats in the examination seat arrangement of the present invention;

[0056] Figure 8 This is a schematic diagram of the automatic arrangement of examination room seats according to the present invention;

[0057] Fig. 9 This is a flow chart of automatic sign-in in the examination room of the present invention;

[0058] Fig.10 It is a flow chart of the examinee misalignment detection of the present invention. DETAILED DESCRIPTION

[0059] The present invention will be further described below in conjunction with the accompanying drawings. The following examples are only used to more clearly illustrate the technical solution of the present invention, and cannot be used to limit the scope of protection of the present invention. It should be noted that the term "comprising" and any variation thereof in the specification and claims of the present invention are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device comprising a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0060] like Figure 1 As shown, an offline examination auxiliary system includes: an examination room camera, an acquisition module, a detection module, a processing module and a sending module.

[0061] The examination room camera is used to capture video of the examination room during the examination preparation and entrance stages.

[0062] The video information of the examination room is obtained through the cameras in each examination room. The examination room camera is installed in the front of the examination room and can shoot the situation of each seat in the examination room from the front. The layout diagram of the examination room camera is shown in the figure below. Figure 2 shown.

[0063] The acquisition module is used to obtain the candidate registration information and examination requirement information input in advance; the examination requirement information includes the examination subjects and the number of candidates for each subject, and the candidate registration information includes: the candidate's registration facial image and identity information.

[0064] The detection module is used to detect the examination room seats in the key frames of the video during the examination preparation stage, and to detect and recognize the faces of the examinees in the key frames of the video during the examination entrance stage.

[0065] In this embodiment, the examination room seats are detected for the video key frames in the examination preparation stage, such as Figure 6 As shown, it specifically includes: using the trained examination room seat detection model to perform seat detection on the extracted video key frames of the examination preparation stage, obtaining the pixel coordinates of the upper left corner and the lower right corner of each seat, converting the upper left corner pixel coordinates or the upper right corner pixel coordinates of all seats into world coordinates, and outputting them as the position coordinates of the seat.

[0066] The examination room seat detection model can refer to the existing technology, specifically, Figure 5 As shown in the figure, the specific steps of training the examination room seat detection model are: select a pre-trained model for seat detection, label seat detection data, fine-tune the pre-trained model, evaluate the performance after fine-tuning, and if the performance meets the requirements, output the trained examination room seat detection model; if the performance does not meet the requirements, adjust the labeled data and training parameters, and re-fine-tune the pre-trained model until the performance meets the requirements.

[0067] In this embodiment, the face detection and recognition of examinees are performed on the key frames of the video at the entrance stage of the examination, specifically including:

[0068] Extract key frames from the video after all candidates have taken their seats during the entrance stage of the test; use the trained face detection and recognition model to perform face detection on the extracted key frames of the video to obtain the face position coordinates and size; the face position coordinates are the coordinates obtained by converting the pixel coordinates of the face center point into world coordinates; perform feature extraction on the detected faces, and perform identity recognition on the detected faces based on all the face images of the candidates.

[0069] The face detection and recognition model can refer to the existing technology, and the specific steps of training the face detection and recognition model can refer to the specific steps of training the examination room seat detection model.

[0070] The processing module is used to arrange the seats in the examination room and generate the examination information according to the candidate registration information, examination requirement information, detected seats and / or detected and recognized facial information, or to check the candidates at the examination entrance stage and generate the candidate verification results.

[0071] More specifically, the processing module includes a seat arrangement subunit, a candidate information generation subunit, an absence and misplacement detection subunit and an automatic sign-in subunit.

[0072] The seat arrangement subunit is used to obtain the number of available seats and find the target test seat set according to the number of candidates and the tested test seat locations, arrange the target test seat set in sequence, and generate a test seat layout diagram.

[0073] In this embodiment, n seats are detected in each examination room, denoted as [(x1, y1), (x2, y2), ... (xn, yn)], and m seats that meet the requirements are selected from the n seats as target examination seats, such as Figure 4 and Figure 7 As shown in FIG. 1 , the specific process of the seat arrangement method based on the examination auxiliary system is as follows:

[0074] According to the detected examination room seat position coordinates, all available seats are obtained; according to the position coordinates of each available seat, the distance between any two seats is calculated; according to the distance between any two seats and the set adjacent distance range, the seat adjacency matrix is ​​constructed; using the greedy algorithm, all non-adjacent point sets are found from the seat adjacency matrix, and this point set is the target examination seat set; the target examination seat set is serially numbered to generate the examination room seat layout diagram.

[0075] Two seats are adjacent, that is, the distance between the two seats is the minimum; the seat adjacency judgment can be changed by selecting or designing different seat distances, so as to obtain different m values ​​(number of examination seats) for a certain examination room. If the sum of the number of examination seats in all examination rooms is not less than the total number of candidates, the automatic arrangement of examination room seats is completed.

[0076] For seat A(x a ,y a ) and seat B(x b ,y b ), commonly used distances include:

[0077] a: Euclidean distance:

[0078]

[0079] b: Manhattan distance:

[0080] distance(A,B)=|x a -x b |+|y a -y b |

[0081] C: Chebyshev distance:

[0082] distance(A,B)=max(|x a -x b |,|y a -y b |)

[0083] The seat selection results based on Euclidean distance are as follows Figure 8 As shown in the figure, the green dot indicates the location of a seat, and the red circle outside the green dot indicates that the seat can be used as the target test seat.

[0084] The candidate information generation subunit is used to allocate seats to each candidate and generate examination information based on the examination room seat layout, candidate registration information and examination requirement information.

[0085] In this embodiment, the seats for each examinee may be classified with reference to the prior art, that is, examinees may be randomly assigned to target examination seats, or the number of each examination seat in the examination room may be obtained by sorting the m examination seats in a specified order.

[0086] The absentee and misplaced detection subunit is used to compare the recognized facial identity with the facial images of candidates for this examination to detect absent or misplaced candidates.

[0087] In this embodiment, if Fig.10 As shown, the specific process of the absence dislocation detection method based on the examination auxiliary system is as follows:

[0088] Compare all recognized facial identities with all facial images of candidates who have registered for the exam to obtain the identities and total number of absent candidates; obtain each candidate's theoretical seat number valid-seat based on the position of each candidate in the exam seat layout; find the nearest exam seat number of each detected face in the exam seat layout as the actual seat number seat based on the position coordinates of the detected face; compare each candidate's theoretical seat number valid-seat and actual seat number seat in turn to see if they are consistent. If they are consistent, the candidate is not sitting in the wrong seat; otherwise, the candidate is sitting in the wrong seat.

[0089] Among them, based on the position coordinates of the detected face, the nearest examination seat number of each detected face is found in the examination room seat layout map as the actual seat number seat, which is specifically determined by the following formula:

[0090] seat=seat i

[0091]

[0092] Among them, (x face ,y face ) is the face position coordinates obtained by face detection, (x i ,y i ,seat i ) is the position coordinates and number of any available seat in the examination room seat layout diagram, and m is the total number of available seats in the examination room seat layout diagram.

[0093] Candidate attendance detection: In order to determine whether all candidates in the examination room are present, the examination assistance system uses face detection and recognition technology to analyze video data, calculate the facial position of each student in the examination room video and identify the student's identity. By comparing with the registered student information of this examination room, it detects absent candidates and candidates who go to the wrong examination room, and counts the number of students who actually take the exam.

[0094] Candidate position check: In order to determine whether there are candidates sitting in the wrong seats in the examination room, the examination auxiliary system checks the position of each candidate based on the candidate's identity obtained by face detection and recognition, as well as the face coordinate results and the candidate's examination seat coordinate information. Candidate position check logic: Face recognition obtains the candidate's face position and identity in the video screen, and then calculates the number of the examination seat closest to the candidate's face position in the video screen. The number of the examination seat should be consistent with the seat number in the candidate's examination information, otherwise the candidate is sitting in the wrong seat.

[0095] The automatic sign-in subunit is used to determine whether there is a misplacement or absence operation based on the detection results of the absence and misplacement detection subunit when the invigilator starts the automatic sign-in operation, and automatically sign in when there is no absence or misplacement, or complete the sign-in through the intervention of the invigilator when there is no misplacement but absence.

[0096] In this embodiment, the automatic check-in in the examination room can automatically detect whether the examinee has walked into the wrong examination room or sat in the wrong seat, and count the actual number of examinees taking the examination; after entering the examination room, the examinee locates his or her examination seat according to the guidance of the projection screen; after all examinees are seated, the invigilator starts the automatic check-in at the invigilator terminal, reminding all examinees in the examination room to face the examination room camera; after receiving the automatic check-in request from the invigilator terminal, the examination auxiliary system starts to analyze the video data captured by the examination room camera and performs automatic check-in.

[0097] like Fig. 9 As shown, the specific process of the automatic sign-in method based on the examination auxiliary system is as follows:

[0098] The invigilator initiates an automatic sign-in command on the invigilator terminal. Based on the face detection and recognition results and the candidate's registration information, the invigilator determines whether there is any misplacement or absence, and automatically signs in when there is no absence or misplacement. When there is an absence but no misplacement, the invigilator intervenes to complete the sign-in. When there is a misplacement, the candidate automatically signs in again after adjustment based on the reported misplacement information. It is worth noting that if the number of automatic sign-ins has reached the upper limit, the invigilator is also required to intervene to complete the sign-in manually or in other set ways.

[0099] The sending module is used to send the test information generated by the processing module and the examinee verification results to the invigilator terminal and / or the examinee terminal at a set time and according to set requirements.

[0100] In this embodiment, sending the test information generated by the processing module and the examinee verification result to the invigilator terminal and / or the examinee terminal at a set time and according to set requirements specifically includes:

[0101] At the designated time, the examination subject, examination time, admission ticket number, examination room number, seat number and the candidate's position in the examination room seat layout will be sent to the candidate's terminal; at the designated time, the examination subject, time, examination room number, examination room location, number of students taking the examination, personal information of all candidates in this examination room, and the position of all candidates in this examination room in the examination room seat layout will be sent to the invigilator's terminal, that is, the candidate only knows his or her seat before entering the examination room, and does not know the position of others.

[0102] In some other embodiments, the examination assistance system also includes a calibration module for obtaining a distortion correction matrix and a perspective transformation matrix, and using the distortion correction matrix to perform distortion correction on the captured video. During the detection process, the perspective transformation matrix is ​​used to convert the detected pixel coordinates into world coordinates, that is, the camera in each examination room is a fixed-angle camera and only needs to be calibrated once.

[0103] In some other embodiments, the examination assistance system also includes an examination room projector, which is used to publicly display part of the examination information during the examination entrance stage; the part of the examination information includes a seating layout of all candidates in the examination room, on which the position of each candidate is displayed with a set number; the projection screen of the projector is hung below the examination room camera.

[0104] During the exam preparation phase before the exam, the exam support system is initialized, the seats in the exam room are automatically arranged, and the exam seat layout diagram is generated. The exam support system obtains the internal information of the exam room, including the number and layout of seats, by analyzing the video data of each exam room, and then summarizes the number and layout of seats in each exam room, and arranges the exam room in combination with the exam subjects and the number of applicants for each subject, and generates exam information, including candidate personal information, exam subject information, exam time, exam room number, exam room layout and seat number, exam room seat layout diagram, allocates the appropriate number of candidates to each exam room, and designates an exam seat for each candidate; ensures that each candidate obtains an exam seat, and that there is enough space between adjacent candidates to prevent plagiarism.

[0105] After initialization, the test assistance system distributes the test information to the candidate terminal and the invigilator terminal at the specified time before the test, and sends the test room seat layout to the projection screen in the test room to help candidates locate their test seats. The candidate terminal runs on the student's mobile phone or other smart device, and can receive the test information issued by the test assistance system at the specified time, including the test subject and test time, admission ticket number, test room number and seat number, and test room seat layout. Before entering the test room, candidates can check the seat layout in the test room through the candidate terminal to find out their test location, but they cannot know the location of others, such as Figure 3 As shown in (a), Figure 3 (a) in the figure is a schematic diagram of the candidate's position in the examination room seat layout received by the candidate's terminal.

[0106] After entering the examination room, candidates can confirm their examination seats again through the projection screen. After entering the examination room, candidates hand in the smart device running the candidate terminal and retrieve it after the examination. The invigilator terminal runs on the invigilator's mobile phone or other smart device, and can receive the examination information of the examination room issued by the examination auxiliary system at the specified time, including the examination subject and time, examination room number and location, number of students taking the examination, personal information of all candidates in the examination room, internal layout of the examination room and seating arrangement of all candidates in the examination room, such as Figure 3 As shown in (b), Figure 3 (b) is a schematic diagram of the positions of all examinees in the examination room seat layout diagram received by the invigilator terminal.

[0107] After the candidates are in their seats, the invigilator instructs the candidates to face the camera in the examination room and start the automatic sign-in process on the examination terminal. If necessary, the invigilator can sign in manually.

[0108] Those skilled in the art can clearly understand that the technology in the embodiments of the present invention can be implemented by means of software plus a necessary general hardware platform. Based on this understanding, the technical solution in the embodiments of the present invention is essentially or the part that contributes to the prior art can be embodied in the form of a software product, which can be stored in a storage medium such as ROM / RAM, a disk, an optical disk, etc., and includes a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment of the present invention or some parts of the embodiments.

[0109] The above are only preferred embodiments of the present invention, and the protection scope of the present invention is not limited to the above embodiments. All technical solutions under the concept of the present invention belong to the protection scope of the present invention. It should be pointed out that for ordinary technicians in this technical field, some improvements and modifications without departing from the principle of the present invention should be regarded as the protection scope of the present invention.

Claims

1. An offline examination auxiliary system, characterized in that: include: Examination room camera, used to capture video of the examination room during the examination preparation and examination entrance stages; The acquisition module is used to obtain the candidate registration information and examination requirement information input in advance; The examination requirement information includes the examination subjects and the number of candidates for each subject, and the candidate registration information includes: the candidate's facial image and identity information; The detection module is used to detect the seats in the examination room based on the key frames of the video during the examination preparation phase, and to detect and recognize the faces of the examinees based on the key frames of the video during the examination entrance phase; A processing module, for arranging seats in the examination room and generating examination information, or checking candidates at the examination entrance stage and generating candidate checking results, based on candidate registration information, examination requirement information, detected seats and / or detected and recognized face information; The sending module is used to send the test information generated by the processing module and the examinee verification results to the invigilator terminal and / or the examinee terminal at a set time and according to set requirements.

2. The offline examination auxiliary system according to claim 1, characterized in that: The detecting of examination room seats for the video key frames in the examination preparation stage specifically includes: Using the trained examination room seat detection model, the seat detection is performed on the extracted key frames of the exam preparation stage video to obtain the pixel coordinates of the upper left corner and lower right corner of each seat. The upper left pixel coordinates or upper right pixel coordinates of all seats are converted into world coordinates and output as the position coordinates of the seats.

3. The offline examination auxiliary system according to claim 1, characterized in that: The test taker face detection and recognition of the key frames of the video during the test entrance stage specifically includes: Extract the key frames of the video after all candidates take their seats during the entrance stage of the exam; Using the trained face detection and recognition model, perform face detection on the extracted video key frames to obtain the face position coordinates and size; the face position coordinates are the coordinates obtained by converting the pixel coordinates of the face center point into world coordinates; Perform feature extraction on the detected face and perform identity recognition on the detected face based on all the registered face images.

4. The offline examination auxiliary system according to claim 2 or 3, characterized in that: It also includes a calibration module for obtaining a distortion correction matrix and a perspective transformation matrix, and using the distortion correction matrix to perform distortion correction on the captured video. During the detection process, the perspective transformation matrix is ​​used to convert the detected pixel coordinates into world coordinates.

5. The offline examination auxiliary system according to claim 1, characterized in that: The processing module comprises: The seat arrangement subunit is used to obtain the number of available seats and find the target test seat set according to the number of candidates and the tested test seat positions, arrange the target test seat set in sequence, and generate a test seat layout diagram; The candidate information generation subunit is used to assign seats to each candidate and generate test information according to the examination room seat layout, candidate registration information and test requirement information; The absence and misplacement detection subunit is used to compare the recognized face identity with the face image of the candidates who have registered for the current examination, and detect the candidates who are absent or misplaced; The automatic sign-in subunit is used to determine whether there is a misplacement or absence operation based on the detection results of the absence and misplacement detection subunit when the invigilator starts the automatic sign-in operation, and automatically sign in when there is no absence or misplacement, or complete the sign-in through the intervention of the invigilator when there is no misplacement but absence.

6. The offline examination auxiliary system according to claim 5, characterized in that: The specific process of obtaining the number of available seats and finding the target test seat set according to the number of examinees and the detected test seat positions is as follows: Obtain all available seats according to the detected examination room seat position coordinates; Calculate the distance between any two seats based on the location coordinates of each available seat; Construct an adjacency matrix of seats based on the distance between any two seats and the set adjacent distance range; Using the greedy algorithm, find all non-adjacent point sets from the seat adjacency matrix. This point set is the target test seat set. Arrange the target test seat set in sequence and generate a test room seat layout diagram.

7. The offline examination auxiliary system according to claim 5, characterized in that: The identified face identity is compared with the face image of the candidates who registered for the current exam to detect the absent or misplaced candidates. The specific process is as follows: Compare all recognized facial identities with the facial images of all candidates who registered for the test to obtain the identities and total number of absent candidates; According to the position of each candidate in the examination room seat layout, obtain the theoretical seat number of each candidate valid-seat; Based on the position coordinates of the detected face, find the nearest examination seat number of each detected face in the examination room seat layout map as the actual seat number seat; Compare each candidate's theoretical seat number valid-seat and actual seat number seat in turn to see if they are consistent. If they are consistent, the candidate is not sitting in the wrong seat; otherwise, the candidate is sitting in the wrong seat.

8. The offline examination auxiliary system according to claim 7, characterized in that: Based on the position coordinates of the detected face, the nearest test seat number of each detected face is found in the test room seat layout map as the actual seat number seat, which is specifically determined by the following formula: seat=seat i Among them, (x face ,y face ) is the face position coordinates obtained by face detection, (x i ,y i ,seat i ) is the position coordinates and number of any available seat in the examination room seat layout diagram, and m is the total number of available seats in the examination room seat layout diagram.

9. The offline examination auxiliary system according to claim 1, characterized in that: The step of sending the test information generated by the processing module and the test taker verification results to the invigilator terminal and / or the test taker terminal at a set time and according to set requirements specifically includes: Send the examination subject, examination time, admission ticket number, examination room number, seat number and the candidate's position in the examination room seat layout to the candidate's terminal at the designated time; At the designated time, the examination subject, time, examination room number, examination room location, number of students taking the examination, personal information of all candidates in the examination room, and the location of all candidates in the examination room in the examination room seat layout are sent to the invigilator's terminal.

10. The offline test assistance system according to claim 1, characterized in that: It also includes an examination room projector, which is used to publicly display part of the examination information during the examination entrance stage; part of the examination information includes a seat layout diagram of all candidates in the examination room, and on the seat layout diagram, the position of each candidate is displayed with a set number.

Images