A system and method for correcting a topic based on a review system

Abstract

The application discloses a kind of based on the correction system and method of subject checking of reading system, image acquisition module is used to collect test paper image;Subject block segmentation module is used to locate each test question area on test paper image and cut out each test question area corresponding answer area, and the answer area obtained by segmentation is framed with rectangular frame;Checking module is used to check the accuracy of the answer area of each test question segmented;Correction module is used to correct the test paper image that test question segmentation exists exception.This application detects the rectangular answer area of the test question segmented, and checks whether the test question is correct by the area calculation of answer area rectangular frame, effectively guarantees the accuracy of the test question, and timely corrects the test question of the abnormal sheet, effectively guarantees the stability and accuracy of reading system.

Description

Technical Field

[0001] This invention relates to the field of intelligent marking technology, and specifically to a topic-related verification and correction system and method based on a marking system. Background Technology

[0002] Under the premise of Education Informatization 2.0, marking systems and assignment systems are increasingly integrated into schools. The current main workflow of marking systems typically involves a test paper acquisition platform capturing images of exam papers and uploading them to a marking platform. The marking platform then segments each paper into question blocks, and the segmented questions are graded online. Therefore, the accuracy of question segmentation is crucial throughout the marking process. Inaccurate segmentation may result in the loss or incomplete display of answers, affecting the marking results. Thus, verifying and correcting the segmented question blocks is essential. Currently, most marking systems do not verify the segmentation results, making it impossible to detect errors in timely manner. Summary of the Invention

[0003] To address the shortcomings of existing technologies, this invention aims to provide a topic-related verification and correction system and method based on a marking system.

[0004] To achieve the above objectives, the present invention adopts the following technical solution:

[0005] A question-based verification and correction system based on a marking system includes an image acquisition module, a question segmentation module, a verification module, and a correction module;

[0006] The image acquisition module is used to acquire images of the exam paper; the image acquisition module scans the exam paper images into a designated folder using a document scanner or a high-speed scanner;

[0007] The question segmentation module is used to locate each question area on the test paper image and segment out the corresponding answer area for each question area. The segmented answer area is outlined with a rectangle.

[0008] The verification module is used to verify the accuracy of question segmentation in the answer areas of each segmented question; the specific process is as follows:

[0009] The system detects continuous rectangles within each question area and calculates the area of ​​the largest rectangle detected in each question area. If the area of ​​the largest rectangle is greater than or equal to a preset minimum area threshold and less than or equal to a preset maximum area threshold, the verification passes. If the area of ​​the largest rectangle is less than the preset minimum area threshold or greater than the preset maximum area threshold, the verification fails, indicating that there is an anomaly in the segmentation.

[0010] The correction module is used to correct the question segmentation of test paper images that are abnormal.

[0011] Furthermore, the correction module provides users with two correction methods: manual correction and automatic correction. In the manual correction method, the correction module provides users with an interface to manually correct the rectangles with abnormal segmentation. In the automatic correction method, the correction module provides users with an interface to manually select the answer area of ​​each question on the original test paper as a question-segmentation template. Then, the correction module re-segments all test papers with abnormal segmentation according to the question-segmentation template.

[0012] Furthermore, the test papers that have been automatically corrected by the correction module need to be verified again by the verification module.

[0013] Furthermore, the correction module allows users to choose between manual or automatic correction, or the correction module can automatically start the manual correction mode when the number of test papers with abnormalities is less than or equal to the preset quantity threshold, otherwise it can automatically start the automatic correction mode.

[0014] The present invention also provides a method using the above system, the specific process of which is as follows:

[0015] The image acquisition module scans the exam paper images into a designated folder using a document scanner or a high-speed scanner;

[0016] The question segmentation module locates each question area on the test paper image and segments the corresponding answer area for each question area. The segmented answer area is outlined with a rectangle.

[0017] The verification module checks the accuracy of question segmentation in the answer areas of each segmented question:

[0018] The system detects continuous rectangles within each question area and calculates the area of ​​the largest rectangle detected in each question area. If the area of ​​the largest rectangle is greater than or equal to a preset minimum area threshold and less than or equal to a preset maximum area threshold, the verification passes. If the area of ​​the largest rectangle is less than the preset minimum area threshold or greater than the preset maximum area threshold, the verification fails, indicating that there is an anomaly in the segmentation.

[0019] The correction module corrects the question segmentation of test paper images that have abnormalities.

[0020] Furthermore, in the above method, the correction module provides users with two correction methods, including manual correction and automatic correction. In the manual correction method, the correction module provides users with an interface for manually correcting the rectangles with abnormal segmentation. In the automatic correction method, the correction module provides users with an interface for manually selecting the answer area of ​​each question on the original test paper as a question-segmentation template. Then, the correction module re-segments all test papers with abnormal segmentation according to the question-segmentation template.

[0021] The beneficial effects of this invention are as follows: This invention effectively ensures the accuracy of question segmentation by detecting the rectangular answer area of ​​the segmented test questions and verifying the correctness of question segmentation by calculating the area of ​​the rectangular answer area. Furthermore, it promptly corrects any abnormalities in the question segmentation of the test paper, thus effectively ensuring the stability and accuracy of the marking system. Detailed Implementation

[0022] The present invention will be further described below. It should be noted that this embodiment is based on the present technical solution and provides detailed implementation methods and specific operation processes, but the protection scope of the present invention is not limited to this embodiment.

[0023] Example 1

[0024] This embodiment provides a question-based verification and correction system for a marking system, including an image acquisition module, a question block segmentation module, a verification module, and a correction module;

[0025] The image acquisition module is used to acquire images of the exam paper; the image acquisition module scans the exam paper images into a designated folder using a document scanner or a high-speed scanner.

[0026] The question segmentation module is used to locate each question area on the test paper image and segment out the corresponding answer area for each question area. The segmented answer area is outlined with a rectangle.

[0027] The verification module is used to verify the accuracy of question segmentation in the answer areas of each segmented question; the specific process is as follows:

[0028] The system detects continuous rectangles within each question area and calculates the area of ​​the largest rectangle detected in each question area. If the area of ​​the largest rectangle is greater than or equal to a preset minimum area threshold and less than or equal to a preset maximum area threshold, the verification passes. If the area of ​​the largest rectangle is less than the preset minimum area threshold or greater than the preset maximum area threshold, the verification fails, indicating an anomaly in the segmentation.

[0029] It should be noted that, in addition to the rectangular frame in the answer area, there may be continuous rectangular frames in the question stem, attached diagrams, etc. The method in this embodiment can filter out these interfering rectangular frame factors by verifying the area of ​​the largest rectangular frame. Since the area of ​​the rectangular frame in the answer area is fixed, the area of ​​the divided rectangular frames needs to be within a preset range.

[0030] Specifically, the detection of rectangular boxes can be achieved through OpenCV's rectangular outline detection.

[0031] The correction module is used to correct the question segmentation of test paper images that are abnormal.

[0032] Specifically, in this embodiment, the correction module provides users with two correction methods: manual correction and automatic correction. In the manual correction method, the correction module provides users with an interface for manually correcting the rectangles with abnormal segmentation (manual correction is sufficient for a small number of rectangles). In the automatic correction method, the correction module provides users with an interface for manually selecting the answer area of ​​each question on the original test paper image as a question-segmentation template. Then, the correction module re-segments all test papers with abnormal segmentation based on the question-segmentation template.

[0033] In this embodiment, the test paper that has been automatically corrected by the correction module needs to be verified again by the verification module.

[0034] In this embodiment, the correction module allows the user to choose between manual correction or automatic correction. Alternatively, the correction module can automatically activate the manual correction mode when the number of test papers with abnormalities is less than or equal to the preset quantity threshold, or automatically activate the automatic correction mode otherwise.

[0035] Example 2

[0036] This embodiment provides a method using the system described in Embodiment 1, the specific process of which is as follows:

[0037] The image acquisition module scans the test paper images into a designated folder using a document scanner or a high-speed scanner.

[0038] The question segmentation module locates each question area on the test paper image and segments the corresponding answer area for each question area. The segmented answer area is outlined with a rectangle.

[0039] The verification module checks the accuracy of question segmentation in the answer areas of each segmented question:

[0040] The system detects continuous rectangles within each question area and calculates the area of ​​the largest rectangle detected in each question area. If the area of ​​the largest rectangle is greater than or equal to a preset minimum area threshold and less than or equal to a preset maximum area threshold, the verification passes. If the area of ​​the largest rectangle is less than the preset minimum area threshold or greater than the preset maximum area threshold, the verification fails, indicating an anomaly in the segmentation.

[0041] The correction module corrects the question segmentation of test paper images that have abnormalities.

[0042] Specifically, the correction module provides users with two correction methods: manual correction and automatic correction. In the manual correction method, the correction module provides users with an interface to manually correct the rectangles with abnormal segmentation. In the automatic correction method, the correction module provides users with an interface to manually select the answer area of ​​each question on the original test paper as a question-segmentation template. Then, the correction module re-segments all test papers with abnormal segmentation according to the question-segmentation template.

[0043] In this embodiment, the test paper that has been automatically corrected by the correction module needs to be verified again by the verification module.

[0044] In this embodiment, the correction module allows the user to choose between manual correction or automatic correction. Alternatively, the correction module can automatically activate the manual correction mode when the number of test papers with abnormalities is less than or equal to the preset quantity threshold, or automatically activate the automatic correction mode otherwise.

[0045] For those skilled in the art, various corresponding changes and modifications can be made based on the above technical solutions and concepts, and all such changes and modifications should be included within the protection scope of the claims of this invention.

Claims

1. A topic-related verification and correction system based on a marking system, characterized in that, It includes an image acquisition module, a question segmentation module, a verification module, and a correction module; The image acquisition module is used to acquire images of the exam paper; The question segmentation module is used to locate each question area on the test paper image and segment out the corresponding answer area for each question area. The segmented answer area is outlined with a rectangle. The verification module is used to verify the accuracy of question segmentation in the answer areas of each segmented question; the specific process is as follows: The system detects continuous rectangles within each question area and calculates the area of ​​the largest rectangle detected in each question area. If the area of ​​the largest rectangle is greater than or equal to a preset minimum area threshold and less than or equal to a preset maximum area threshold, the verification passes. If the area of ​​the largest rectangle is less than the preset minimum area threshold or greater than the preset maximum area threshold, the verification fails, indicating that there is an anomaly in the segmentation. The correction module is used to correct the question segmentation of test paper images that are abnormal.

2. The system according to claim 1, characterized in that, The correction module provides users with two correction methods: manual correction and automatic correction. In manual correction, the module provides an interface for users to manually correct the rectangles with abnormal segmentation. In automatic correction, the module provides an interface for users to manually select the answer area of ​​each question on the original test paper as a question-segmentation template. Then, the correction module re-segments all test papers with abnormal segmentation based on the question-segmentation template.

3. The system according to claim 1, characterized in that, The test paper that has been automatically corrected by the correction module needs to be verified again by the verification module.

4. The system according to claim 1, characterized in that, The correction module allows users to choose between manual or automatic correction. Alternatively, the correction module can automatically activate manual correction when the number of test papers with abnormalities is less than or equal to the preset threshold, or automatically activate automatic correction if the number of abnormal test papers is less than or equal to the threshold.

5. A method using the system according to any one of claims 1-4, characterized in that, The specific process is as follows: The image acquisition module scans the exam paper images into a designated folder using a document scanner or a high-speed scanner; The question segmentation module locates each question area on the test paper image and segments the corresponding answer area for each question area. The segmented answer area is outlined with a rectangle. The verification module checks the accuracy of question segmentation in the answer areas of each segmented question: The system detects continuous rectangles within each question area and calculates the area of ​​the largest rectangle detected in each question area. If the area of ​​the largest rectangle is greater than or equal to a preset minimum area threshold and less than or equal to a preset maximum area threshold, the verification passes. If the area of ​​the largest rectangle is less than the preset minimum area threshold or greater than the preset maximum area threshold, the verification fails, indicating that there is an anomaly in the segmentation. The correction module corrects the question segmentation of test paper images that have abnormalities.

6. The method according to claim 5, characterized in that, The correction module provides users with two correction methods: manual correction and automatic correction. In manual correction, the module provides an interface for users to manually correct the rectangles with abnormal segmentation. In automatic correction, the module provides an interface for users to manually select the answer area of ​​each question on the original test paper as a question-segmentation template. Then, the correction module re-segments all test papers with abnormal segmentation based on the question-segmentation template.