Homework correction mark-remaining printing method, storage medium and equipment

By performing structured processing and similarity matching on the assignment images, the problem of incorrect order and direction in the marking and recording process of the marking machine was solved, realizing unified marking and recording printing of assignments from multiple classes, and improving the accuracy and operational flexibility of marking and recording.

CN120825549APending Publication Date: 2025-10-21读书郎教育科技有限公司
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202510945949.7
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2025-07-09
Publication Date
2025-10-21

AI Technical Summary

Technical Problem

Existing grading machines are prone to errors in order and direction during the grading and recording process, and cannot scan multiple classes' assignments at once for unified recording and printing, making the operation inflexible.

Method used

By structuring the images of assignments, calculating the density of handwritten characters and the frequency of responses, selecting target matching areas, and using similarity matching technology to accurately locate correction marks during printing, the system supports paper assignments in both directions and enables unified printing of correction labels.

Benefits of technology

It improves the accuracy and operational flexibility of marking and recording, allowing multiple classes' assignments to be scanned at once and printed together, reducing operational errors and human intervention, and improving efficiency.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure SMS_1
    Figure SMS_1
  • Figure SMS_2
    Figure SMS_2
  • Figure SMS_3
    Figure SMS_3
Patent Text Reader

Abstract

The invention discloses a homework correction mark-leaving printing method, a storage medium and equipment, and the method achieves the correspondence between a correction label and a paper homework to be subjected to correction mark-leaving printing through similarity matching, thereby allowing a teacher to scan and correct all batches of paper homework at a time, and then carrying out the mark-leaving printing operation in a unified manner, and improving the efficiency. The method is more flexible, accurate correspondence between the correction label and the paper homework can be ensured, and the accuracy of mark-leaving printing is ensured. Besides, the blocks of one or more test questions with high handwriting density and high student answering frequency in the homework images are selected as the target matching areas, and only the target matching areas of the homework images need to be subjected to feature matching in the subsequent similarity matching retrieval, so that the matching can be more accurate, and the matching speed is higher. The paper feeding direction of the paper operation of mark-leaving printing is supported to be positively placed, reversely placed or reversely placed, the limitation on paper feeding during printing is reduced, the requirement for an operator is lowered, and the problem of mark-leaving staggering is reduced.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the field of intelligent education technology, and in particular to a homework correction trace printing method, storage medium and device. Background Art

[0002] With the rapid development of information-based education, the powerful support of large-scale models has made intelligent grading of test papers / homework a reality. Grading traces for test papers / homeworks means that after students complete their work on paper test papers / homeworks, the teacher scans the paper test papers / homeworks to be graded into the grading system, completes the grading on the grading system, and finally prints the teacher's grading traces on the corresponding paper test papers / homeworks. Currently, there are grading machines on the market that have integrated grading traces for test papers / homeworks, but these grading machines have the following problems with grading trace printing:

[0003] 1. The scanning and correction of homework and the printing of the homework mark must ensure that the homework is in the same order, front and back, and direction in order to accurately leave the mark. Otherwise, the mark will be misplaced or wrong.

[0004] 2. Usually a teacher will teach multiple classes. The current marking machine only supports scanning and marking the test papers / homework of a class and then printing the trace of the class immediately. It cannot scan and mark the test papers / homework of multiple classes at one time and then perform the trace printing operation uniformly, which is not flexible enough. Summary of the Invention

[0005] In view of the deficiencies in the prior art, the present invention aims to provide a method, storage medium and device for printing traces of homework corrections.

[0006] In order to achieve the above object, the present invention adopts the following technical solutions:

[0007] A method for printing homework correction traces, comprising the following steps:

[0008] S1, scan the paper homework to be corrected to obtain the homework image set;

[0009] S2. Performing a first image preprocessing on each job image in the job image set to obtain a first job image set;

[0010] S3. Perform second image preprocessing on each job image in the first job image set to obtain a target matching area for each job image:

[0011] S3.1. Structuring the homework image: Segment the homework image into test questions to obtain blocks of each test question in the homework image;

[0012] S3.2, performing handwriting OCR detection on the cutouts of each test question obtained in step S3.1 to detect the handwriting area of ​​the cutouts of each test question;

[0013] S3.3. Calculate and select the target matching area:

[0014] S3.3.1. Calculate the handwriting density of each test question's cutout according to the following formula to obtain the handwriting density value of each test question's cutout;

[0015]

[0016] Where Density is the handwriting density value of the test question block, Pixb is the area of ​​each handwriting area in the test question block, and Spaper is the total area of ​​the test question block;

[0017] S3.3.2. Calculate the answer frequency value for each question using the following formula:

[0018]

[0019] Where Fre is the answer frequency value of a certain test question, Nstu is the number of students who answered the test question, and Ntotal is the total number of paper assignments. The number of students who answered a test question is the number of handwritten areas detected on the cut-out of each assignment image.

[0020] S3.3.3. Calculate the weight of each question's slice as follows:

[0021] Rtarget=m*Density+n*Fre

[0022] m+n=1

[0023] Among them, Rtarget is the weighted value of the test question, and m and n are weighting coefficients;

[0024] Sort the slices of each test question in the job image according to the weighted value Target, and use the slices of the first or multiple test questions with the largest weighted value Target as the target matching area of ​​the job image;

[0025] S4. The teacher corrects each homework image and generates correction marks left by the teacher on each homework image as correction labels for the test questions in the homework image; the correction labels include each correction mark data and the relative position of each correction mark data on the homework image;

[0026] S5. Place the paper job to be printed into the paper inlet of the printer, use the printer to capture an image of the paper job to be printed, retrieve a corresponding job image based on the target matching area, and obtain a corresponding correction label: perform similarity matching on each job image in the first job image set with a corresponding area in the image of the paper job to be printed, use the job image with the highest similarity as the retrieval result, and use the correction label of the retrieved job image as the correction label to be printed;

[0027] S6. Print the correction label obtained in step S5 onto the paper job to be printed: print each correction trace in the correction label at a corresponding position on the paper job according to its relative position on the job image.

[0028] Furthermore, in step S2, the first image preprocessing includes uniformly scaling all job images to a preset first size.

[0029] Furthermore, in step S2, the first image preprocessing further includes image binarization or grayscale conversion.

[0030] Furthermore, in step S3.3.3, the blocks of the test questions in the work image whose weighted values ​​are greater than a preset threshold are used as target matching areas of the work image.

[0031] Furthermore, in step S4, the correction traces include symbols and / or evaluation statements used by the teacher to correct the homework.

[0032] Furthermore, in step S5, the SSIM algorithm is used to perform similarity matching.

[0033] Furthermore, in step S5, before printing, the paper job is first detected to be inverted according to the image of the paper job to be printed. If the paper job is inverted, the correction trace data in the correction label is inverted before printing on the paper job.

[0034] The present invention also provides a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, and the computer program implements the above method when executed by a processor.

[0035] The present invention also provides a computer device, comprising a processor and a memory, wherein the memory is used to store a computer program; and when the processor is used to execute the computer program, the above method is implemented.

[0036] The beneficial effects of the present invention are:

[0037] 1. The present invention achieves the correspondence between the correction label and the paper homework that is currently to be corrected and printed through similarity matching, thereby allowing teachers to scan and correct all batches of paper homework at one time, and then perform the trace printing operation uniformly. This is not only more flexible, but also can ensure the precise correspondence between the correction label and the paper homework, and ensure the accuracy of the trace printing.

[0038] 2. The present invention selects one or more test questions with high handwriting density and high student answer frequency in the homework image as the target matching area. In the subsequent similarity matching retrieval, only the target matching area of ​​the homework image needs to be feature matched, without the need to perform feature matching on the entire homework image, which can achieve more accurate matching and faster matching speed.

[0039] 3. The present invention supports paper feeding directions of paper jobs for trace printing, such as upright, reverse or upside down feeding, which reduces restrictions on paper feeding during printing, lowers requirements for operators, and reduces the problem of staggered traces. DETAILED DESCRIPTION

[0040] The present invention will be further described below. It should be noted that this embodiment is based on the technical solution and provides a detailed implementation method and specific operation process, but the protection scope of the present invention is not limited to this embodiment.

[0041] This embodiment provides a method for printing a trace of homework corrections, including the following steps:

[0042] S1. Scan the paper assignments to be corrected to obtain a set of assignment images. If the paper assignments are double-sided, double-sided scanning can be used to obtain the images of the front and back sides of each paper assignment.

[0043] S2. Perform a first image preprocessing on each job image in the job image set to obtain a first job image set; the first image preprocessing includes uniformly scaling all job images to a preset first size to eliminate the influence of size differences.

[0044] In this embodiment, the first image preprocessing may further include image binarization or grayscale conversion.

[0045] S3. Perform second image preprocessing on each job image in the first job image set to obtain a target matching area for each job image:

[0046] S3.1. Structuring the homework image: Segment the homework image into test questions to obtain blocks of each test question in the homework image;

[0047] S3.2, performing handwriting OCR detection on the cutouts of each test question obtained in step S3.1 to detect the handwriting area of ​​the cutouts of each test question;

[0048] S3.3. Calculate and select the target matching area:

[0049] S3.3.1. Calculate the handwriting density of each test question's cutout according to the following formula to obtain the handwriting density value of each test question's cutout;

[0050]

[0051] Wherein, Density is the handwriting density value of the test question block, Pixb is the area of ​​each handwriting area in the test question block, and Spaper is the total area of ​​the test question block.

[0052] S3.3.2. Calculate the answer frequency value for each question using the following formula:

[0053]

[0054] Where Fre is the answer frequency value of a certain test question, Nstu is the number of students answering the corresponding test question, and Ntotal is the total number of paper assignments (or the total number of students). The number of students answering a test question is the number of handwritten areas detected on the cut-out of each assignment image for that test question.

[0055] S3.3.3. Calculate the weight of each question's slice as follows:

[0056] Rtarget=m*Density+n*Fre

[0057] m+n=1

[0058] Among them, Rtarget is the weighted value of the test question, and m and n are weighting coefficients;

[0059] The slices of each test question in the work image are sorted according to the weighted value Target, and the slices of the first one or more test questions with the largest weighted value Target are used as the target matching area of ​​the work image. Specifically, the slices of the test questions in the work image with weighted values ​​greater than a preset threshold can be used as the target matching area of ​​the work image.

[0060] It should be noted that the purpose of the second image preprocessing is to obtain a block of one or more test questions with high handwriting density and high frequency of student answers in the homework image as the target matching area. In the subsequent image retrieval, feature matching is only performed on the target matching area of ​​this homework image, without the need for feature matching of the entire homework image, which can be more accurate and faster.

[0061] S4. The teacher corrects each assignment image and generates a correction label for the test question in the assignment image using the correction traces left by the teacher. The correction label contains the data for each correction trace and its relative position on the assignment image. The correction trace includes the symbols and / or evaluation statements used by the teacher to correct the assignment, such as any one or a combination of ticks, crosses, circles, slashes, reasons for errors, and answers.

[0062] S5. Place the paper job to be printed into the paper inlet of the printer, use the printer to capture an image of the paper job to be printed, retrieve the corresponding job image according to the target matching area, and obtain the corresponding correction label.

[0063] S6. Print the correction label obtained in step S5 onto the paper job to be printed: print each correction trace in the correction label at a corresponding position on the paper job according to its relative position on the job image.

[0064] In this embodiment, in step S5, the target matching area of ​​each job image in the first job image set is matched with the corresponding area in the image of the paper job to be printed one by one for similarity, the job image with the highest similarity is used as the retrieval result, and the correction label of the retrieved job image is used as the correction label to be printed.

[0065] Specifically, in this embodiment, the SSIM algorithm is used for similarity matching. The core of the SSIM algorithm is to quantify image similarity through three dimensions: brightness, contrast, and structure.

[0066] Furthermore, in step S5, before printing, the paper job is first detected to be inverted according to the image of the paper job to be printed. If the paper job is inverted, the correction trace data in the correction label is inverted before printing on the paper job.

[0067] Those skilled in the art can make various corresponding changes and modifications based on the above technical solutions and concepts, and all of these changes and modifications should be included in the scope of protection of the claims of the present invention.

Claims

1. A method for printing homework correction traces, characterized in that: The steps include: S1, scan the paper homework to be corrected to obtain the homework image set; S2. Performing a first image preprocessing on each job image in the job image set to obtain a first job image set; S3. Perform second image preprocessing on each job image in the first job image set to obtain a target matching area for each job image: S3.

1. Structuring the homework image: Segment the homework image into test questions to obtain blocks of each test question in the homework image; S3.2, performing handwriting OCR detection on the cutouts of each test question obtained in step S3.1 to detect the handwriting area of ​​the cutouts of each test question; S3.

3. Calculate and select the target matching area: S3.3.

1. Calculate the handwriting density of each test question's cutout according to the following formula to obtain the handwriting density value of each test question's cutout; Where Density is the handwriting density value of the test question block, Pixb is the area of ​​each handwriting area in the test question block, and Spaper is the total area of ​​the test question block; S3.3.

2. Calculate the answer frequency value for each question using the following formula: Where Fre is the answer frequency value of a certain test question, Nstu is the number of students who answered the test question, and Ntotal is the total number of paper assignments. The number of students who answered a test question is the number of handwritten areas detected on the cut-out of each assignment image. S3.3.

3. Calculate the weight of each question's slice as follows: Rtarget=m*Density+n*Fre m+n=1 Among them, Rtarget is the weighted value of the test question, and m and n are weighting coefficients; Sort the slices of each test question in the job image according to the weighted value Target, and use the slices of the first or multiple test questions with the largest weighted value Target as the target matching area of ​​the job image; S4. The teacher corrects each homework image and generates correction marks left by the teacher on each homework image as correction labels for the test questions in the homework image; the correction labels include each correction mark data and the relative position of each correction mark data on the homework image; S5. Place the paper job to be printed into the paper inlet of the printer, use the printer to capture an image of the paper job to be printed, retrieve a corresponding job image based on the target matching area, and obtain a corresponding correction label: perform similarity matching on each job image in the first job image set with a corresponding area in the image of the paper job to be printed, use the job image with the highest similarity as the retrieval result, and use the correction label of the retrieved job image as the correction label to be printed; S6. Print the correction label obtained in step S5 onto the paper job to be printed: print each correction trace in the correction label at a corresponding position on the paper job according to its relative position on the job image.

2. The method according to claim 1, characterized in that In step S2, the first image preprocessing includes uniformly scaling all job images to a preset first size.

3. The method according to claim 2, characterized in that In step S2, the first image preprocessing further includes image binarization or grayscale conversion.

4. The method according to claim 1, wherein In step S3.3.3, the blocks of the test questions in the work image whose weighted values ​​are greater than the preset threshold are used as the target matching areas of the work image.

5. The method according to claim 1, wherein In step S4, the marking traces include the symbols and / or evaluation statements used by the teacher to mark the homework.

6. The method according to claim 1, characterized in that In step S5, the SSIM algorithm is used to perform similarity matching.

7. The method according to claim 1, characterized in that In step S5, before printing, the paper job is first detected to be inverted according to the image of the paper job to be printed. If the paper job is inverted, the correction trace data in the correction label is inverted before printing on the paper job.

8. A computer-readable storage medium, characterized in that The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the method according to any one of claims 1 to 7 is implemented.

9. A computer device, characterized in that: The method comprises a processor and a memory, wherein the memory is used to store a computer program; and when the processor is used to execute the computer program, the method according to any one of claims 1 to 7 is implemented.