Objective question marking method and system based on machine vision

The method and system use multiple electronic terminals with varying performance levels to efficiently and accurately grade objective test papers, addressing the inefficiencies and costs of existing technologies by utilizing common mobile devices and high-performance terminals for machine vision-based grading.

CN119251854AInactive Publication Date: 2025-01-03BEIJING FENGHUANG XUE YI SCI & TECH CO LTD
View PDF 4 Cites 0 Cited by

Patent Information

Application Number
CN202411303251.7
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-09-19
Publication Date
2025-01-03
Estimated Expiration
Not applicable · inactive patent

AI Technical Summary

Technical Problem

Existing automatic grading technologies for objective questions in examinations are costly, require specialized hardware, are inflexible, and prone to misjudgment due to answer sheet format changes, leading to inefficiencies and high maintenance costs.

Method used

A method and system utilizing multiple groups of electronic terminals with varying performance levels, including common mobile devices for image capture, higher-performance terminals for image processing, and a scoring unit, to efficiently and accurately mark objective test papers using machine vision, ensuring objectivity and fairness while reducing hardware costs.

Benefits of technology

The solution enables efficient, accurate, and cost-effective grading of large-scale examinations by leveraging common mobile devices and high-performance terminals, ensuring timely data transmission and maintaining privacy, thus enhancing marking efficiency and reducing hardware costs.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119251854A_ABST
    Figure CN119251854A_ABST
Patent Text Reader

Abstract

The invention provides an objective question marking method and system based on machine vision, and belongs to the technical field of image recognition processing and educational informatization. The method comprises the steps of test paper electronization, objective question positioning, answer area recognition and the like executed by mobile terminals or electronic terminals of different groups, and the final test paper marking and scoring step. In the test paper marking and scoring step, objective questions of each test paper to be read are scored based on a pre-stored objective question standard answer set; the objective question standard answer set comprises a plurality of answer conjunctions. The system comprises M1 first mobile terminals, M2 second electronic terminals, M3 third electronic terminals and a scoring unit, the M < 1 > < gt >; m < 2gt >; 1; m < 1gt >; m3gt; 1. According to the technical scheme, the hardware arrangement cost of objective question marking is reduced, the method is suitable for objective question answer sheets in various template formats and easy to popularize, and the marking and scoring efficiency of objective questions in large-scale multi-frequency examinations is improved.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention belongs to the field of image recognition processing and educational informatization technology, and in particular relates to an objective test paper marking method and system based on machine vision, a computer-readable storage medium for implementing the method, a computer program product and an electronic device. Background Art

[0002] The marking of the examination papers after the exam is an important part of the examination process. Teachers and other examiners will give students and candidates grades after marking the papers. For a long time, examiners have to mark all paper examination papers one by one. The longer the marking time and the more examination papers they need to mark, the more likely they are to feel tired, which will affect the efficiency and accuracy of marking.

[0003] In recent years, various automatic grading technologies based on OMR cursor reading have gradually entered the automatic grading of objective questions in large-scale examinations, but they also have various disadvantages. This automatic grading technology cannot be used alone and needs to be paired with a dedicated cursor reader with a relatively high price; secondly, to use this technology, you need to be familiar with the filling rules of the answer sheet, otherwise it is easy to cause system misjudgment; at the same time, if there are few objective questions in the exam, then this technology will cause a waste of answer sheet resources; in addition, answer sheets need to be specially made and required to be arranged in a fixed pattern. Once the style of the answer sheet changes, it will be difficult for the system to recognize, and the upgrade is difficult and the maintenance cost is high.

[0004] In the actual education process, examinations, as an important means of education evaluation, need to be conducted frequently and in large quantities. If the above-mentioned automatic answer sheet grading technology is used every time, the hardware cost will be high and the flexibility will be poor, which cannot meet the needs of teachers' independent question setting and personalized teaching evaluation. Summary of the invention

[0005] In order to solve the above technical problems, the present invention proposes an objective test paper marking method and system based on machine vision, a computer-readable storage medium, a computer program product and an electronic device for implementing the method.

[0006] In a first aspect of the present invention, a method for marking objective test papers based on machine vision is proposed. The method is executed based on multiple electronic terminal devices including a mobile terminal, and specifically comprises the following steps:

[0007] Electronic examination papers: The first group of mobile terminals is used to capture images of the examination papers, and N electronic examination papers to be reviewed are obtained; N>1;

[0008] Objective question positioning: receiving the N electronic test papers to be reviewed through the second group of electronic terminals, and positioning the objective question areas in the test papers to be reviewed, and obtaining N objective question area images;

[0009] Answer area identification: receiving N objective question area images through a third group of electronic terminals, and identifying the answer set corresponding to each objective question area image;

[0010] Marking and scoring: Based on the pre-stored standard answer set of objective questions, score the objective questions of each test paper to be marked;

[0011] The standard answer set for objective questions includes multiple answer combinations;

[0012] Each of the answer combinations includes a number, a fraction and a combination of:

[0013] (1) Single letter; (2) Multiple letters; (3) Predicate.

[0014] Wherein, the second group of electronic terminals and the third group of electronic terminals are different from the first group of mobile terminals;

[0015] The first group of mobile terminals is equipped with an image acquisition device and an image preprocessing device;

[0016] The image acquisition device is used to take pictures of each test paper to be reviewed to obtain an image of the test paper to be reviewed;

[0017] The image preprocessing device preprocesses the image of the test paper to be reviewed to obtain an electronic test paper to be reviewed, and sends it to the second group of electronic terminals;

[0018] The image processing performance of the second group of electronic terminals is higher than that of the first group of mobile terminals.

[0019] In the specific implementation of the present invention, the method is performed by three groups of mobile terminals or electronic terminal devices of different groups;

[0020] Specifically, the first group of mobile terminals includes M1 first mobile terminals {D1, D2, . . . , D M1};

[0021] The second group of electronic terminals includes M2 second electronic terminals {T1, T2, . . . , T M2}; The third group of electronic terminals includes M3 third electronic terminals {S1, S2, ..., S M3};

[0022] M1>M2>1; M1>M3>1;

[0023] To reduce hardware costs, the first group of mobile terminals may be common mobile terminal devices with lower performance, that is, the image processing performance of the second group of electronic terminals and the third group of electronic terminal devices is higher than that of the first group of mobile terminals.

[0024] The first mobile terminal D iThe test papers J i The size is L i ; 1≤i≤N

[0025] The first mobile terminal D i According to the electronic examination papers to be reviewed i Size L i , the size of the electronic examination papers to be reviewed currently being processed by the second group of electronic terminals, and determining to receive the examination papers to be reviewed i The target second electronic terminal T j ,1≤j≤M2.

[0026] Assume that the second electronic terminal T x Locate the test papers to be read y The time spent on the objective questions section is t xy , waiting for examination paper J y The size of the objective question area is L y0 ;

[0027] Among them, 1≤x≤M2,1≤y≤N;

[0028] The second electronic terminal T x According to the test paper J y The size of the objective question area image L y0 , the average recognition speed of the objective question area picture currently being processed by the third group of electronic terminals, and determining the receiving of the test paper to be read y The objective question area picture of the target third electronic terminal S z ,1≤z≤M3. The objective question location step locates the objective question area in the test paper to be reviewed, including:

[0029] For each electronic test paper to be reviewed, perform the following steps:

[0030] S410: Performing perspective transformation processing on the test paper to be reviewed to obtain a corrected image;

[0031] S420: Searching for a handwritten target character region in the corrected image, where the handwritten target character is a single letter, a number, a judgment symbol, or multiple letters or numbers;

[0032] S430: Connect each handwritten target character area with the area serial number to form the objective question area.

[0033] The answer area recognition step identifies the answer set corresponding to each objective question area image, which specifically includes the following steps:

[0034] S510: extracting multiple character features of each handwritten target character area in the objective question area;

[0035] S520: identifying target characters in each handwritten target character area based on a template matching algorithm;

[0036] S530: storing the serial numbers of the upper and lower adjacent regions or the left and right adjacent regions in association with the target characters in the handwritten target character region as an answer combination;

[0037] S540: Summarize all answer combinations to obtain an answer set corresponding to each objective question area image.

[0038] In order to implement the method described in the first aspect, in the second aspect of the present invention, a machine vision-based objective test grading system is proposed, the system comprising: M1 first mobile terminals {D1, D2, ..., D M1}, M2 second electronic terminals {T1, T2, ..., T M2}, M3 third electronic terminals {S1, S2, ..., S M3} and scoring unit; said M1>M2>1; M1>M3>1;

[0039] M1 mobile terminals {D1, D2, ..., D M1}Acquire images of N test papers and obtain N electronic test papers to be reviewed {J1, J2, ..., J N}; N > 1;

[0040] M2 second electronic terminals {T1, T2, ..., T M2}Receive the N electronic examination papers {J1, J2, ..., J N}, and locate the objective question area in the test paper to be reviewed, and obtain N objective question area images;

[0041] M3 third electronic terminals {S1, S2, ..., S M3}Receive N objective question area images and identify the answer set corresponding to each objective question area image;

[0042] The scoring unit scores the objective questions of each test paper to be reviewed based on the pre-stored standard answer set of objective questions;

[0043] Among them, M2 second electronic terminals {T1, T2, ..., T M2 and M3 third electronic terminals {S1, S2, ..., S M3} has a higher image processing performance than the M1 mobile terminals {D1, D2, ..., D M1};

[0044] The pre-stored standard answer set for objective questions includes a plurality of answer associations;

[0045] Each of the answer combinations includes a number, a fraction and a combination of:

[0046] (1) Single letter; (2) Multiple letters; (3) Predicate.

[0047] M2 second electronic terminals {T1, T2, ..., T M2}Receive the N electronic examination papers {J1, J2, ..., J N}, and locate the objective question area in the test paper to be reviewed, and obtain N objective question area images, including:

[0048] For each electronic test paper to be reviewed, perform the following steps:

[0049] S410: Performing perspective transformation processing on the test paper to be reviewed to obtain a corrected image;

[0050] S420: Searching for a handwritten target character region in the corrected image, where the handwritten target character is a single letter, a number, a judgment symbol, or multiple letters or numbers;

[0051] S430: connecting all handwritten target character areas to form the objective question area;

[0052] M3 third electronic terminals {S1, S2, ..., S M3}Receive N objective question area images and identify the answer set corresponding to each objective question area image, including:

[0053] S510: extracting multiple character features of each handwritten target character area in the objective question area;

[0054] S520: identifying target characters in each handwritten target character area based on a template matching algorithm;

[0055] S530: storing the target characters in the upper and lower adjacent or left and right adjacent handwritten target character regions in an associated manner as an answer union;

[0056] S540: Summarize all answer combinations to obtain an answer set corresponding to each objective question area image.

[0057] Part or all of the steps of the objective test marking method based on machine vision described in the first aspect can be automatically implemented through various forms of electronic devices and computer program instructions; the computer program instructions can be stored in different forms of storage media and loaded into computer electronic devices for execution.

[0058] Therefore, in the third aspect of the present invention, a computer-readable storage medium is also provided for storing computer instructions. When the computer instructions are executed on an electronic device, the electronic device executes the objective test marking method based on machine vision as described in the first aspect.

[0059] In the fourth aspect of the present invention, an electronic device is also proposed, which includes a processor and a memory, the memory is used to store instructions, and the processor is used to call the instructions in the memory, so that the electronic device executes the objective test marking method based on machine vision described in the first aspect above.

[0060] In a fifth aspect of the present invention, a computer program product is provided, the product comprising a computer program, and when the computer program is executed, the objective test marking method based on machine vision described in the first aspect is implemented.

[0061] A large number of hardware devices of the technical solution of the present invention can directly apply various existing ordinary (low-performance) mobile terminal devices, without the need to configure additional special equipment, reducing the hardware layout cost and facilitating promotion and application; at the same time, different marking processes are performed based on different groups of mobile terminals or electronic terminal devices, which not only takes into account the different functions (performance) of devices in different groups, but also ensures the privacy processing of the data itself, and ensures the objectivity and fairness of the marking process; in addition, when terminal devices in different groups are performing data exchange transmission, the source device (device to send data) fully considers the current processing performance of the target device before sending the data, and then selects the best target device to send the data, avoiding data transmission blockage, which can ensure the timeliness of the marking data transmission in the large-scale examination marking process, thereby ensuring the marking efficiency.

[0062] Further advantages of the present invention will be further reflected in detail in the specific embodiments section in conjunction with the drawings of the specification. BRIEF DESCRIPTION OF THE DRAWINGS

[0063] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0064] Figure 1 It is a main flow diagram of a method for marking objective questions based on machine vision according to an embodiment of the present invention;

[0065] Figure 2 yes Figure 1A schematic diagram of the arrangement of mobile terminals of different groups in the objective test paper marking method based on machine vision;

[0066] Figure 3 It is a schematic diagram of different template formats of objective test answer sheets applicable to the technical solution of the present invention;

[0067] Figure 4 yes Figure 1 A schematic diagram of a specific execution flow of objective question positioning in the method;

[0068] Figure 5 yes Figure 1 A schematic diagram of a specific execution flow of answer area identification in the method;

[0069] Figure 6 It is a schematic diagram of some principles and effects of implementing the technical solution of the present invention. DETAILED DESCRIPTION

[0070] Here, exemplary embodiments will be described in detail, examples of which are shown in the accompanying drawings. When the following description refers to the drawings, unless otherwise indicated, the same numbers in different drawings represent the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments that are the same as the present application. Instead, they are only examples of devices and methods that are the same as some aspects of the present application as detailed in the attached claims.

[0071] In the embodiment of the present application, the term "module" or "unit" refers to a computer program or a part of a computer program with a predetermined function, and works together with other related parts to achieve a predetermined goal, and can be implemented in whole or in part by using software, hardware (such as processing circuits or memories) or a combination thereof. Similarly, a processor (or multiple processors or memories) can be used to implement one or more modules or units. In addition, each module or unit can be a part of an overall module or unit containing the module or unit function. At the same time, in the specific implementation of the present application, if it is related to user-related data, when the embodiment of the present application is applied to a specific product or technology, it is necessary to obtain user permission or consent, and the collection, use and processing of relevant data need to comply with the relevant laws, regulations and standards of relevant countries and regions.

[0072] The block diagrams shown in the accompanying drawings are merely functional entities and do not necessarily correspond to physically independent entities. That is, these functional entities may be implemented in software form, or in one or more hardware modules or integrated circuits, or in different networks and / or processor devices and / or microcontroller devices.

[0073] See also Figure 1 , Figure 1A schematic diagram of the main flow of an objective test paper marking method based on machine vision according to an embodiment of the present invention is shown.

[0074] Figure 1 The method includes the steps of electronicization of the test paper, positioning of objective questions, identification of answer areas, etc., which are executed by mobile terminals or electronic terminals of different groups, and the final step of marking and grading the test paper.

[0075] The functions of each step are summarized as follows:

[0076] Electronic examination papers: The first group of mobile terminals is used to capture images of the examination papers, and N electronic examination papers to be reviewed are obtained; N>1;

[0077] Objective question positioning: receiving the N electronic test papers to be reviewed through the second group of electronic terminals, and positioning the objective question areas in the test papers to be reviewed, and obtaining N objective question area images;

[0078] Answer area identification: receiving N objective question area images through a third group of electronic terminals, and identifying the answer set corresponding to each objective question area image;

[0079] Marking and scoring: Based on the pre-stored standard answer set of objective questions, score the objective questions of each test paper to be marked;

[0080] Wherein, the second group of electronic terminals and the third group of electronic terminals are different from the first group of mobile terminals;

[0081] The first group of mobile terminals is equipped with an image acquisition device and an image preprocessing device;

[0082] The image acquisition device is used to take pictures of each test paper to be reviewed to obtain an image of the test paper to be reviewed;

[0083] The image preprocessing device preprocesses the image of the test paper to be reviewed to obtain an electronic test paper to be reviewed, and sends it to the second group of electronic terminals;

[0084] The second group of electronic terminals locates the objective question area in the test paper to be reviewed, obtains N objective question area images, and sends them to the third group of electronic terminals;

[0085] The third group of electronic terminals identifies the answer set corresponding to each objective question area picture.

[0086] In order to better explain the method embodiment of the present invention, Figure 2 , the above-mentioned embodiment is continued to be developed in the form of parameterized computer program pseudo code language.

[0087] See also Figure 2 , Figure 2 yes Figure 1A schematic diagram of the arrangement of mobile terminals of different groups in the objective test marking method based on machine vision.

[0088] Specifically, the first group of mobile terminals includes M1 first mobile terminals The second group of electronic terminals includes M2 second electronic terminals ; The third group of electronic terminals includes M3 third electronic terminals ;M1>M2>1;M1>M3>1.

[0090] As an improvement, the first group of mobile terminals includes M1 first mobile terminals {D1, D2, . . . , D M1} are all existing common mobile terminals, such as mobile phones.

[0091] In the specific implementation of the present invention, M1 mobile terminals {D1, D2, ..., D M1}Various existing common mobile phone devices can be directly adopted without additional configuration.

[0092] Specifically, M1 mobile terminals {D1, D2, ..., D M1}It can be a smart phone device of a student, parent or teacher, equipped with an image acquisition device and an image preprocessing device;

[0093] The image acquisition device is used to take pictures of each test paper to be reviewed to obtain an image of the test paper to be reviewed;

[0094] The image preprocessing device preprocesses the image of the test paper to be marked to obtain an electronic test paper to be marked, and sends it to the second group of electronic terminals.

[0095] Optionally, the preprocessing includes grayscale binarization, noise reduction processing, average median filtering, etc., which can be performed on an ordinary smart phone through an ordinary image processing program without the need for special equipment or special models.

[0096] Therefore, M1 first mobile terminals {D1, D2, ..., D M1} can collect images of N test papers and obtain N electronic test papers to be reviewed {J1, J2, ..., J N}; N > 1;

[0097] In practical applications, if the test is conducted in a closed-book format, after the test is over and the answer sheets are collected, a large number of personnel can be assigned to use their ordinary mobile phones to capture images of N test answer sheets to obtain N electronic test papers {J1, J2, ..., J N};

[0098] If the test is open book, each candidate can capture an image of his or her answer sheet through his or her own mobile phone or that of another person (such as a parent) after finishing the test.

[0099] Different from the disadvantage of the prior art that a special answer sheet needs to be configured, the technical solution of the present application can provide various forms of answer sheets for the answer sheets to be reviewed, and the candidates can fill in the answers independently.

[0100] Figure 3 It is a schematic diagram of different template formats of objective question answer sheets to which the technical solution of the present invention is applicable.

[0101] After obtaining the pre-processed electronic examination papers to be reviewed, each first mobile terminal needs to send them to M2 second electronic terminals {T1, T2, ..., T M2} proceed to the next step.

[0102] In the implementation process of this application, in order to ensure data security isolation, M2 second electronic terminals {T1, T2, ..., T M2} is isolated from each first mobile terminal in space, and the two communicate with each other through a wireless network, and each first mobile terminal cannot communicate with each other (is in an independent isolation state),

[0103] At this time, a large number of electronic examination papers to be reviewed of the first mobile terminals are generated almost at the same time and are almost transmitted simultaneously within a predetermined time period; this process is prone to data congestion, resulting in data loss.

[0104] At this time, it is particularly important for each first mobile terminal to correctly select a target second electronic terminal.

[0105] In the case where there is only one or a few mobile terminals, each first mobile terminal only needs to select a target second electronic terminal that is currently in an idle state to avoid data blocking.

[0106] However, in the actual implementation of the present application, since M1 is much larger than M2, usually M1>10*M2; at this time, if a second electronic terminal is detected as idle by a first mobile terminal, it will also be detected as idle by other (assuming 50) first mobile terminals. At this time, there will be a situation where (50+1) mobile terminals simultaneously send their own generated electronic test papers to be marked to the idle second electronic terminal almost at the same time, and communication congestion will also occur, resulting in data loss due to load imbalance.

[0107] Therefore, the processing method of simply determining whether the second electronic terminal is idle cannot be applied to the specific scenario described in the present invention.

[0108] For this reason, the further improvements of the present invention are as follows:

[0109] Assume that the first mobile terminal D i The test papers J i The size is L i ; 1≤i≤N;

[0110] The mobile terminal D i According to the electronic examination papers to be reviewed i Size L i , the size of the electronic examination papers to be reviewed currently being processed by the second group of electronic terminals, and determining to receive the examination papers to be reviewed i The target second electronic terminal T j ,1≤j≤M2.

[0111] Specifically, as a preferred method implemented by a computer, j is determined as follows:

[0112]

[0113] Img k is the kth second electronic terminal T k The size of the electronic test paper currently being processed; 1≤j≤M2, % represents the modulo operation, Indicates rounding down.

[0114] It can be seen that the above formula determines the target second electronic terminal T k , not only taking into account the electronic examination papers to be read i Size L i , the size of the electronic examination papers to be reviewed currently being processed by the second group of electronic terminals is also taken into consideration, and a certain processing margin (rounded down) is reserved, which can maximize the load balancing of the second group of electronic terminals while ensuring that data transmission is not blocked, thereby avoiding data loss.

[0115] Furthermore, the above embodiment also performs different marking processes based on different groups of mobile terminals or electronic terminal devices, which not only takes into account the different functions (performance) of the devices in different groups, but also ensures the privacy processing of the data itself, and ensures that the marking process is objective and fair;

[0116] Similarly, M2 second electronic terminals {T1, T2, ..., T M2}Receive the N electronic examination papers {J1, J2, ..., J N}, and locate the objective question area in the test paper to be reviewed. After obtaining N objective question area images, they need to be sent to M3 third electronic terminals {S1, S2, ..., S M3}, at this time, it is equally important for each second mobile terminal to correctly select the target third electronic terminal.

[0117] For this reason, the further improvements of the present invention are as follows:

[0118] Assume that the second electronic terminal T x Locate the test papers to be read y The time spent on the objective questions section is t xy , waiting for examination paper J y The size of the objective question area is L y0 ;

[0119] Among them, 1≤x≤M2,1≤y≤N;

[0120] The second electronic terminal T x According to the test paper J y The size of the objective question area image L y0 , the average recognition speed of the objective question area picture currently being processed by the third group of electronic terminals, and determining the receiving of the test paper to be read y The objective question area picture of the target third electronic terminal S z ,1≤z≤M3.

[0121] Specifically, as a preferred method implemented by a computer, z is determined as follows:

[0122] Determine each third electronic terminal S a (a=1,2,…,M3) Average recognition speed V for objective question area pictures a ;

[0123]

[0124] t y,j For size L y,j The time it takes from input to output of recognition results for objective question area images.

[0125] calculate

[0126] Pick The smallest a is taken as z.

[0127] It can be seen that when determining T j When determining S z When , this embodiment focuses on the recognition speed.

[0128] This is because the M2 second electronic terminals {T1, T2, ..., T M2} performs the objective question location step, which inputs the entire electronic test paper to be reviewed, so the image size needs to be considered as a whole; and M3 third electronic terminals {S1, S2, ..., S M3}The answer area recognition step is executed. The input of this step is the image of the objective question area that has been segmented. The image size is already very small (relative to the objective question positioning step). Therefore, it is necessary to focus on matching the recognition speed of the target third electronic terminal device at this time. This is another important improvement point of the present invention, which is particularly important for improving the recognition effect and recognition speed.

[0129] In a specific embodiment, M2 second electronic terminals {T1, T2, ..., T M2} and M3 third electronic terminals {S1, S2, ..., S M3} and the M1 first mobile terminals {D1, D2, ..., D M1}They are geographically isolated from each other, and M1 is much larger than M2 and M3, for example, M1>10*M2; M1>10*M3;

[0130] Although the M2 second electronic terminals {T1, T2, ..., T M2} and M3 third electronic terminals {S1, S2, ..., S M3 The processing performance of the first mobile terminals {D1, D2, ..., D M1}.

[0131] However, the number of high-performance second and third electronic terminals that need to be configured is much lower than that of the first mobile terminal, and existing classroom equipment can be used without additional configuration. And the so-called "high performance" actually only requires the second and third electronic terminals to be configured with basic graphics image processing cards (such as general-purpose GPUs) in terms of hardware performance and to load basic image processing code models (such as open source software such as OpenCV) in terms of software, thereby greatly reducing the cost of implementing the environment layout.

[0132] Most importantly, the above process executes different marking processes based on different groups of mobile terminals or electronic terminal devices, which not only takes into account the different functions (performance) of different groups of devices, but also ensures the privacy processing of the data itself, ensuring the objectivity and fairness of the marking process.

[0133] It can be seen that no matter whether it is the first mobile terminal, the second electronic device or the third electronic device, what is processed is the answer sheet to be reviewed, and the actual standard answer matching and scoring process is not involved. The actual scoring process can be performed by a high-level manager with marking authority in the subsequent marking and scoring steps. The marking terminal device is equipped with a user login authority control module. After summarizing the answer sets corresponding to all N objective question area images, it scores the objective questions of each test paper to be reviewed based on the pre-stored objective question standard answer set, thereby avoiding data (answer) leakage and protecting personal privacy (avoiding early leakage of personal scores).

[0134] Next, the specific implementation methods of the above-mentioned objective question locating steps and answer area identification steps are introduced respectively. Figure 4 yes Figure 1 A schematic diagram of a specific execution flow of objective question positioning in the method;

[0135] Specifically, the objective question locating step locates the objective question area in the test paper to be reviewed, including:

[0136] For each electronic test paper to be reviewed, perform the following steps:

[0137] S410: Performing perspective transformation processing on the test paper to be reviewed to obtain a corrected image;

[0138] S420: Searching for a handwritten target character region in the corrected image, where the handwritten target character is a single letter, a number, a judgment symbol, or multiple letters or numbers;

[0139] S430: Connect each handwritten target character area with the area serial number to form the objective question area.

[0140] In step S410, because the photographed picture is inevitably tilted or distorted when taking pictures with a mobile phone, appropriate perspective transformation processing is required to allow the answer sheet photo to proceed to the next step;

[0141] Perspective Transformation is to project an image onto a new viewing plane, also known as Projective Mapping. The specific principle can be found in the prior art, such as the following documents:

[0142] Zhu Yuwan. Perspective transformation and perspective projection[J]. Journal of Engineering Graphics, 1988(9):1-9.

[0143] In step S420, since the embodiment of the present application is aimed at offline handwritten paper answer sheets, what needs to be identified is the relevant handwritten answers. Therefore, it is necessary to search for the handwritten target character area in the corrected image, and the handwritten target character is a single letter, number, judgment symbol or multiple letters and numbers.

[0144] As an example, a single letter is generally a single-choice answer to a multiple-choice question, such as A, B, C, D (when the answer options are A, B, C, D); a single number can be a handwritten 1, 2, 3, 4 (when the answer options are 1, 2, 3, 4); a single judgement symbol can be √, × (the answer to the judgment question is √ for right or × for wrong) or √, ○ (when the answer to the judgment question is √ for right or ○ for wrong); multiple letters can be any two combinations of A, B, C, D in multiple-choice questions or a combination of 1, 2, 3, 4 (when the answer options are 1, 2, 3, 4), etc.

[0145] Step S430: connect each handwritten target character area with the area serial number to serve as the objective question area, specifically, connect the handwritten answer character with the corresponding question serial number (ie, area serial number) to serve as the objective question area.

[0146] See also Figure 3 The handwritten answer characters can be written in the table box (rectangular box) below the question number (i.e. area number), or they can be written on the corresponding underline to the right of the question number (i.e. area number). These are the two most common objective test answer format templates, and teachers can choose or combine them at their own discretion.

[0147] Furthermore, Figure 5 yes Figure 1 A schematic diagram of the specific execution flow of answer area identification in the method.

[0148] The answer area recognition step identifies the answer set corresponding to each objective question area image, which specifically includes the following steps:

[0149] S510: extracting multiple character features of each handwritten target character area in the objective question area;

[0150] S520: identifying target characters in each handwritten target character area based on a template matching algorithm;

[0151] S530: storing the serial numbers of the upper and lower adjacent regions or the left and right adjacent regions in association with the target characters in the handwritten target character region as an answer combination;

[0152] S540: Summarize all answer combinations to obtain an answer set corresponding to each objective question area image.

[0153] Step S510 may extract multiple character features for each handwritten character (e.g., a handwritten A);

[0154] Step S520 identifies the target characters in each handwritten target character area based on a template matching algorithm.

[0155] How to extract the handwritten character features of the target area to identify the target characters can be referred to the prior art. The handwritten answer characters are written into the table box (rectangular box) below the question number (i.e., the area number). For related prior art, see:

[0156] Xia He, Feng Junhuan. Research on table data positioning algorithm for automatic marking of handheld terminals[J]. Microcomputers and Applications, 2014, 33(06).

[0157] Similar feature extraction algorithms and machine learning algorithms can also be used for the answer sheet written on the underline corresponding to the right side of the question number (i.e., area number).

[0158] Therefore, it can be seen that the technical solution of the present invention has strong adaptability and can be applied to objective test answer sheet recognition in various environments, without being limited to answer sheet templates in a specific format. This is because the above process is essentially an image processing process based on machine vision recognition.

[0159] Steps S530 and S540 can be combined Figure 6 Further explanation is as follows:

[0160] S530: Associate the area numbers of the upper and lower adjacent areas or the left and right adjacent areas with the target characters of the handwritten target character area and store them as an answer union.

[0161] See also Figure 6 An example of a handwritten target character region is the handwritten answer ABCD-… corresponding to the single-choice question number 1-2-3-…; 1-2-3-… is the region number; the upper and lower adjacent regions of the recognized target character A are numbered 1, and the upper and lower adjacent regions of the recognized target character D are numbered 4, so an example of an answer union is (1-(A), (4-(D));

[0162] S540: Summarize all answer combinations to obtain the answer set corresponding to each objective question area image. For specific results, see Figure 2 A schematic diagram of the "answer set corresponding to the answer sheet area picture" area is shown.

[0163] On this basis, the pre-set standard answer set for objective questions includes multiple answer combinations;

[0164] Each of the answer combinations includes a number, a fraction and a combination of:

[0165] (1) Single letter; (2) Multiple letters; (3) Predicate.

[0166] For example, the standard answer set for objective questions is (1-(A)-2 points; 4-(C)-2 points; 15-(AD)-3 points, 21-(√)-2 points);

[0167] Then, marking and scoring: based on the pre-stored set of standard answers to objective questions, score the objective questions of each test paper to be marked.

[0168] In the above example, the score for Multiple Choice Question 1 is 2.

[0169] After introducing the method embodiment of the present invention, the present invention further introduces a system embodiment as follows:

[0170] A machine vision-based objective test grading system, the system comprising: M1 first mobile terminals {D1, D2, ..., D M1}, M2 second electronic terminals {T1, T2, ..., T M2}, M3 third electronic terminals {S1, S2, ..., S M3} and scoring unit; said M1>M2>1; M1>M3>1;

[0171] M1 mobile terminals {D1, D2, ..., D M1}Acquire images of N test papers and obtain N electronic test papers to be reviewed {J1, J2, ..., J N}; N > 1;

[0172] M2 second electronic terminals {T1, T2, ..., T M2}Receive the N electronic examination papers {J1, J2, ..., J N}, and locate the objective question area in the test paper to be reviewed, and obtain N objective question area images;

[0173] M3 third electronic terminals {S1, S2, ..., S M3}Receive N objective question area images and identify the answer set corresponding to each objective question area image;

[0174] The scoring unit scores the objective questions of each test paper to be reviewed based on the pre-stored standard answer set of objective questions;

[0175] Among them, M2 second electronic terminals {T1, T2, ..., T M2 and M3 third electronic terminals {S1, S2, ..., S M3 The image processing performance is higher than that of the M1 mobile terminals {D1, D2, ..., D M1};

[0176] The pre-stored standard answer set for objective questions includes a plurality of answer associations;

[0177] Each of the answer combinations includes a number, a fraction and a combination of:

[0178] (1) Single letter; (2) Multiple letters; (3) Predicate.

[0179] Mobile terminal D i The test papers J i The size is L i ; 1≤i≤N;

[0180] The mobile terminal D i According to the test paper J i Size L i , M2 second electronic terminals {T1, T2, ..., T M2} The size of the test paper currently being processed to be read, determine the size of the test paper to be read i The target second electronic terminal T j ;

[0181] Img k is the kth electronic terminal T k The size of the electronic test paper currently being processed; 1≤j≤M2, % represents the modulo operation, Indicates rounding down.

[0182] M2 second electronic terminals {T1, T2, ..., T M2}Receive the N electronic examination papers {J1, J2, ..., J N}, and locate the objective question area in the test paper to be reviewed, and obtain N objective question area images, including:

[0183] For each electronic test paper to be reviewed, perform the following steps:

[0184] S410: Performing perspective transformation processing on the test paper to be reviewed to obtain a corrected image;

[0185] S420: Searching for a handwritten target character region in the corrected image, where the handwritten target character is a single letter, a number, a judgment symbol, or multiple letters or numbers;

[0186] S430: connecting all handwritten target character areas to form the objective question area;

[0187] M3 third electronic terminals {S1, S2, ..., S M3}Receive N objective question area images and identify the answer set corresponding to each objective question area image, including:

[0188] S510: extracting multiple character features of each handwritten target character area in the objective question area;

[0189] S520: identifying target characters in each handwritten target character area based on a template matching algorithm;

[0190] S530: storing the target characters in the upper and lower adjacent or left and right adjacent handwritten target character regions in an associated manner as an answer union;

[0191] S540: Summarize all answer combinations to obtain an answer set corresponding to each objective question area image.

[0192] It is understandable that after the embodiment in the form of method flow is elaborated in detail, the corresponding embodiment in the form of system function architecture does not need to be repeated again, but those skilled in the art can directly and unambiguously determine that the embodiment in the form of system function architecture also corresponds to all the contents of the embodiment in the form of method flow. The same is true for the embodiment in the form of computer program code pseudo language.

[0193] The present invention provides multiple embodiments, each of which can constitute an independent technical solution and may contribute to the prior art and solve corresponding technical problems. However, it should be pointed out that different embodiments can be combined with each other without violating logic; at the same time, each embodiment can solve at least one technical problem, but it is not required that each individual embodiment solves multiple or all technical problems.

[0194] For other technologies, principles, algorithms or models not elaborated in detail in this application, please refer to the prior art.

[0195] A large number of hardware devices of the technical solution of the present invention can directly apply various existing ordinary (low-performance) mobile terminal devices, without the need to configure additional special equipment, reducing the hardware layout cost and facilitating promotion and application; at the same time, different marking processes are performed based on different groups of mobile terminals or electronic terminal devices, which not only takes into account the different functions (performance) of devices in different groups, but also ensures the privacy processing of the data itself, and ensures the objectivity and fairness of the marking process; in addition, when terminal devices in different groups are performing data exchange transmission, the source device (device to send data) fully considers the current processing performance of the target device before sending the data, and then selects the best target device to send the data, avoiding data transmission blockage, which can ensure the timeliness of the marking data transmission in the large-scale examination marking process, thereby ensuring the marking efficiency.

[0196] In the above embodiments, the description of each embodiment has its own emphasis. For parts that are not described in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0197] It should be noted that, for the objective test paper marking method based on machine vision in the embodiment of the present application, ordinary testers in the field can understand that all or part of the process of implementing the objective test paper marking method based on machine vision in the embodiment of the present application can be completed by controlling the relevant hardware through a computer program, and the computer program can be stored in a computer-readable storage medium, such as stored in the memory of an electronic device, and executed by at least one processor in the electronic device, and the execution process may include the process of the embodiment of the objective test paper marking method based on machine vision. Among them, the storage medium can be a disk, an optical disk, a read-only memory, a random access memory, etc.

[0198] The foregoing has shown and described the method embodiments and system of the present invention, but it is understandable to those skilled in the art that various changes, modifications, substitutions and variations may be made to these embodiments without departing from the principles and spirit of the present invention, and the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A method for marking objective questions based on machine vision, characterized in that: The method comprises the following steps: Electronic examination papers: The first group of mobile terminals is used to capture images of the examination papers, and N electronic examination papers to be reviewed are obtained; N>1; Objective question positioning: receiving the N electronic test papers to be reviewed through the second group of electronic terminals, and positioning the objective question areas in the test papers to be reviewed, and obtaining N objective question area images; Answer area identification: receiving N objective question area images through a third group of electronic terminals, and identifying the answer set corresponding to each objective question area image; Marking and scoring: Based on the pre-stored standard answer set of objective questions, score the objective questions of each test paper to be marked; Wherein, the second group of electronic terminals and the third group of electronic terminals are different from the first group of mobile terminals; The first group of mobile terminals is equipped with an image acquisition device and an image preprocessing device; The image acquisition device is used to take pictures of each test paper to be reviewed to obtain an image of the test paper to be reviewed; The image preprocessing device preprocesses the image of the test paper to be reviewed to obtain an electronic test paper to be reviewed, and sends it to the second group of electronic terminals; The image processing performance of the second group of electronic terminals is higher than that of the first group of mobile terminals.

2. The objective test paper marking method based on machine vision as claimed in claim 1, characterized in that: The first group of mobile terminals includes M1 first mobile terminals {D1, D2, . . . , D M1 }; The second group of electronic terminals includes M2 second electronic terminals {T1, T2, . . . , T M2 };M1>M2>1; The first mobile terminal D i The test papers J i The size is L i ; 1≤i≤N The first mobile terminal D i According to the electronic examination papers to be reviewed i Size L i , the size of the electronic examination papers to be reviewed currently being processed by the second group of electronic terminals, and determining to receive the examination papers to be reviewed i The target second electronic terminal T j ,1≤j≤M2.

3. The objective test paper marking method based on machine vision as claimed in claim 1, characterized in that: The first group of mobile terminals includes M1 first mobile terminals {D1, D2, . . . , D M1 }; The second group of electronic terminals includes M2 second electronic terminals {T1, T2, . . . , T M2 }; The third group of electronic terminals includes M3 third electronic terminals {S1, S2, ..., S M3 }; M1>M2>1; M1>M3>1; Assume that the second electronic terminal T x Locate the test papers to be read y The time spent on the objective questions section is t xy , waiting for examination paper J y The image size of the objective question area is Among them, 1≤x≤M2,1≤y≤N; The second electronic terminal T x According to the test paper J y The size of the objective question area image L y0 , the average recognition speed of the objective question area picture currently being processed by the third group of electronic terminals, and determining the receiving of the test paper to be read y The objective question area picture of the target third electronic terminal S z ,1≤z≤M3.

4. The objective test paper marking method based on machine vision as claimed in claim 1, characterized in that: Objective question location step locates the objective question area in the test paper to be reviewed, including: For each electronic test paper to be reviewed, perform the following steps: S410: Performing perspective transformation processing on the test paper to be reviewed to obtain a corrected image; S420: Searching for a handwritten target character region in the corrected image, where the handwritten target character is a single letter, a number, a judgment symbol, or multiple letters or numbers; S430: Connecting all handwritten target character areas and area serial numbers to form the objective question area.

5. The objective test paper marking method based on machine vision as claimed in claim 4, characterized in that: The answer area recognition step identifies the answer set corresponding to each objective question area image, which specifically includes the following steps: S510: extracting multiple character features of each handwritten target character area in the objective question area; S520: identifying target characters in each handwritten target character area based on a template matching algorithm; S530: storing the serial numbers of the upper and lower adjacent regions or the left and right adjacent regions in association with the target characters in the handwritten target character region as an answer combination; S540: Summarize all answer combinations to obtain an answer set corresponding to each objective question area image.

6. The objective test paper marking method based on machine vision as claimed in claim 5, characterized in that: The standard answer set for objective questions includes multiple answer combinations; Each of the answer combinations includes a number, a fraction and a combination of: (1) Single letter; (2) Multiple letters; (3) Predicate.

7. An objective test paper marking system based on machine vision, the system comprising: M1 first mobile terminals {D1, D2, ..., D M1 }, M2 second electronic terminals {T1, T2, ..., T M2 }, M3 third electronic terminals {S1, S2, ..., S M3 } and scoring unit; M1>M2>1; M1>M3>1; Features: M1 first mobile terminals {D1, D2, ..., D M1 }Acquire images of N test papers and obtain N electronic test papers to be reviewed {J1, J2, ..., J N }; N > 1; M2 second electronic terminals {T1, T2, ..., T M2 }Receive the N electronic examination papers {J1, J2, ..., J N }, and locate the objective question area in the test paper to be reviewed, and obtain N objective question area images; M3 third electronic terminals {S1, S2, ..., S M3 }Receive N objective question area images and identify the answer set corresponding to each objective question area image; The scoring unit scores the objective questions of each test paper to be reviewed based on the pre-stored standard answer set of objective questions; Among them, M2 second electronic terminals {T1, T2, ..., T M2 and M3 third electronic terminals {S1, S2, ..., S M3 The image processing performance is higher than that of the M1 mobile terminals {D1, D2, ..., D M1 }; The pre-stored standard answer set for objective questions includes a plurality of answer associations; Each of the answer combinations includes a number, a fraction and a combination of: (1) Single letter; (2) Multiple letters; (3) Predicate.

8. The objective test paper marking system based on machine vision as claimed in claim 7, characterized in that: M2 second electronic terminals {T1, T2, ..., T M2 }Receive the N electronic examination papers {J1, J2, ..., J N }, and locate the objective question area in the test paper to be reviewed, and obtain N objective question area images, including: For each electronic test paper to be reviewed, perform the following steps: S410: Performing perspective transformation processing on the test paper to be reviewed to obtain a corrected image; S420: Searching for a handwritten target character region in the corrected image, where the handwritten target character is a single letter, a number, a judgment symbol, or multiple letters or numbers; S430: connecting all handwritten target characters and area numbers to form the objective question area; M3 third electronic terminals {S1, S2, ..., S M3 }Receive N objective question area images and identify the answer set corresponding to each objective question area image, including: S510: extracting multiple character features of each handwritten target character area in the objective question area; S520: identifying target characters in each handwritten target character area based on a template matching algorithm; S530: storing the serial numbers of the upper and lower adjacent regions or the left and right adjacent regions in association with the target characters in the handwritten target character region as an answer combination; S540: Summarize all answer combinations to obtain an answer set corresponding to each objective question area image.

9. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, it is used to implement the objective test marking method based on machine vision as described in any one of claims 1 to 6.

Citation Information

Patent Citations

  • Answer paper processing system

    CN106503637A

  • Automatic examination paper objective question marking method, electronic device, equipment and storage medium

    CN110879965A

  • Method for collecting handwritten answer area in test paper

    CN116486411A

  • Network exanmination paper go-over method for subjective question

    CN1598845A