Paper sheet double-side correction judgment method, device and equipment and readable medium

Abstract

The invention discloses a paper sheet double-face correction judgment method, device and equipment and a readable medium, and the method comprises the steps: scanning a to-be-corrected paper, and uploading the to-be-corrected paper; performing sample extraction on the uploaded to-be-corrected paper to obtain a first sample paper set; the method comprises the following steps: performing paper determination and judgment on a first sample paper set, performing first judgment logic after judgment, and automatically and accurately judging whether test paper scanned by a current class needs single-side correction or double-side correction through sample extraction, paper determination and judgment and combination of two-stage judgment logic. The problem that an existing intelligent correcting machine cannot accurately judge a correcting object when single-face and double-face mixed placement is carried out is solved, the actual requirement of a school for printing different test papers on the front face and the back face of one piece of paper for saving consumables is effectively met, two-time independent correcting mark reserving on the front face and the back face is supported, the function of single-page paper face correcting mark reserving is achieved, and the practicability is high. And the adaptability of the intelligent correcting machine to diversified teaching scenes is improved.

Description

TECHNICAL FIELD

[0001] The present application belongs to the field of educational technology, and particularly relates to a single / double-sided correction judgment method and device for a roll, an equipment and a readable medium. BACKGROUND

[0002] Under the background of rapid development of smart education, homework intelligent correction machine as an important tool to improve teaching efficiency has attracted widespread attention and application in the field of education. Such equipment can realize full-intelligent correction of student homework and test papers, and the core workflow mainly consists of three key links of scanning test papers, intelligent correcting test papers and test paper printing, which greatly reduces the workload of manual correction by teachers through automatic processing.

[0003] However, the existing smart correction machine has significant technical limitations in actual application. The current equipment is generally designed for double-sided scanning of test papers, and can only identify and correct the front and back of the same paper at the same time. However, in actual teaching scenarios, there are two outstanding problems:

[0004] The placement of test papers or homework submitted by students is random, which can be face up or back up. When test papers in these two placement modes are mixed together, the existing correction machine cannot accurately determine the specific page that needs to be corrected, resulting in ambiguous correction targets.

[0005] Schools often print different contents on the front and back of the same paper to save the cost of consumables. In this case, the front and back of the paper need to be corrected separately and leave correction marks. However, the existing equipment cannot meet the demand of single-page roll surface independent correction and mark leaving, and it is difficult to adapt to this practical application scenario of saving consumables.

[0006] Therefore, how to solve the problem of single / double-sided correction judgment of test papers, especially when the front and back of student test papers are mixed, to intelligently and accurately determine the corresponding page that needs to be corrected, has become a key difficulty that needs to be broken through in the development of homework intelligent correction technology in the field of smart education, which is of great significance to improve the practicality and adaptability of intelligent correction machine. Therefore, we need to provide a single / double-sided correction judgment method and device for a roll, an equipment and a readable medium. SUMMARY

[0007] The present application aims to provide a single / double-sided paper face correction determination method, device, equipment and readable medium, which can automatically and accurately determine whether the current class scanned test paper needs single-sided correction or double-sided correction through sample extraction, paper face determination and two-level determination logic, solves the problem that the existing intelligent correction machine cannot accurately determine the correction object when facing single / double-sided mixed placement, effectively meets the actual needs of schools to print different test papers on the front and back of a piece of paper to save consumables, supports two separate corrections and marks on the front and back, realizes the function of single-page paper face correction and marking, and improves the adaptation ability of the intelligent correction machine to diversified teaching scenarios, to solve the problem of how to determine whether the current class scanned test paper needs single-sided correction or double-sided correction in the above background art, if it is determined that the single-sided correction is needed, how to intelligently and accurately determine the corresponding page that needs to be corrected when the student's test paper front and back is mixed.

[0008] To achieve the above-mentioned purpose, the present application adopts the following technical scheme: a single / double-sided paper face correction determination method, comprising the following steps:

[0009] The paper face to be corrected is scanned and uploaded;

[0010] The uploaded paper face to be corrected is sample extracted to obtain a first sample paper face set;

[0011] The first sample paper face set is subjected to paper face determination, and after determination, a first determination logic is executed;

[0012] If the first determination logic determines the correction mode, the determination is ended;

[0013] If the first determination logic does not confirm the correction mode, the sample extraction is performed on the paper face to be corrected to obtain a second sample paper face set;

[0014] The obtained second paper face set is subjected to paper face determination, and after determination, a second determination logic is executed to confirm the correction mode.

[0015] Preferably, the first sample paper face set and the second sample paper face set can be the same sample paper face set.

[0016] Preferably, the paper face determination further refers to paper face determination of the paper face information scanned by each student, that is, the first test paper image scanned by each student is defined as the front, and the second image is defined as the back, which is essentially a front and back identification of the student paper face in all samples. After this step, the first face image and the second face image in the student image will be explained in the following.

[0017] Preferably, the paper face determination can be determined by image similarity recognition and / or text ocr recognition;

[0018] Wherein, the image similarity refers to obtaining the front and back images of the student A, randomly defining the first one as the front and the second one as the back, comparing the similarity of the images of other students in the sample with the front and / or back of the student A, so as to confirm the front and back of other students in the sample; the text ocr recognition and determination refers to obtaining the front and back images of the student A, randomly defining the first one as the front and the second one as the back, obtaining the ocr text in the front and the ocr text in the back, then performing ocr recognition on each image of other students in the sample, and matching the similarity with the ocr text of the front and / or back of the student A, so as to confirm the front and back of other students.

[0019] Preferably, the first determination logic is a mark determination logic, and the corresponding second determination logic is a whether to answer determination logic; the mark determination logic refers to determining whether there is an already corrected mark in the front and back images: obtaining the front and back of the first sample, if both the front and back have been marked, neither of the front and back continues to perform the recognition and correction logic; if only one of the front and back is marked, it is determined that the single-page correction is performed, and the page to be corrected is specified as the unmarked front or back; if neither of the front and back is marked, the second determination logic is continued to perform; the whether to answer determination logic further refers to determining whether there is a student's handwriting trace in the front and back of the student's paper; if neither of the front and back is answered, it is determined that the current paper is not corrected; if only one of the front and back is answered, it is determined that the single-page correction is performed, and the page to be corrected is specified as the answered page; if both of the front and back are answered, it is determined that the double-page correction logic is performed, and the front and back are both recognized and corrected.

[0020] Preferably, the first determination logic is a whether to answer determination logic, and the corresponding second determination logic is a mark determination logic; the whether to answer determination logic further refers to determining whether there is a student's handwriting trace in the front and back of the student's paper; if neither of the front and back is answered, it is determined that the current paper is not corrected; if only one of the front and back is answered, it is determined that the single-page correction is performed, and the page to be corrected is specified as the answered page; if both of the front and back are answered, the second determination logic is continued to perform; the mark determination logic refers to determining whether there is an already corrected mark in the front and back images: obtaining the front and back of the first sample, if both the front and back have been marked, neither of the front and back continues to perform the recognition and correction logic; if only one of the front and back is marked, it is determined that the single-page correction is performed, and the page to be corrected is specified as the unmarked front or back; if neither of the front and back is marked, it is determined that the double-page correction logic is performed, and the front and back are both recognized and corrected.

[0021] Preferably, in the first determination logic, if the image ratio of the sample front or back leaving marks is greater than a first preset threshold, it is determined that the front has left marks or the back has left marks; in the second determination logic, if the image ratio of the student's answer on the front or back of the sample is greater than a second preset threshold, it is determined that the front has left marks or the back has answered.

[0022] The single / double-sided sheet grading determination device comprises:

[0023] The scanning unit is configured to scan the to-be-graded sheet;

[0024] The identification unit is configured to perform sample extraction, sheet determination, and determination logic to determine the grading mode on the scanned to-be-graded sheet, including front and back identification of the sample sheet, sheet determination through image similarity identification and / or text ocr identification, execution of the leaving mark determination logic and the answer determination logic.

[0025] The sample extraction unit is configured to extract a first sample sheet set and a second sample sheet set from the scanned to-be-graded sheet;

[0026] The determination logic execution unit is configured to execute the first determination logic and the second determination logic, wherein the first determination logic can be the leaving mark determination logic or the answer determination logic, the second determination logic is another determination logic corresponding to the first determination logic, and the grading mode is determined according to the determination result;

[0027] The threshold setting unit is configured to preset various determination thresholds, including a first preset threshold, a second preset threshold, a third preset threshold, a fourth preset threshold, a fifth preset threshold, a sixth preset threshold, etc., to provide a quantitative standard for the execution of the determination logic.

[0028] An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the sheet single / double-sided grading determination method when executing the computer program.

[0029] A computer readable storage medium storing non-transitory computer readable instructions, which, when executed by a computer, can implement the sheet single / double-sided grading determination method.

[0030] The technical effects and advantages of the present application: compared with the prior art, the sheet single / double-sided grading determination method, device, equipment and readable medium provided by the present application have the following advantages:

[0031] The present application can automatically and accurately determine whether the current class scanned test paper needs single-sided correction or double-sided correction by combining sample extraction, roll surface determination and two-level determination logic, solves the problem that the existing intelligent correction machine cannot accurately determine the correction object when facing single-sided and double-sided mixed placement, effectively meets the actual needs of schools to print different test papers on the front and back of a piece of paper to save consumables, supports two separate corrections of the front and back, realizes the function of single-page roll surface correction, improves the adaptation ability of the intelligent correction machine to diversified teaching scenes, reduces the operation steps and workload of teachers without manual selection of correction pages, avoids errors that may occur during manual operation, and improves the use efficiency of users; meanwhile, by supporting flexible single-sided and double-sided correction, the consumable saving demand of schools is further met, and the consumable use cost and equipment use difficulty in the teaching process are reduced.

[0032] Other features and advantages of the present application will be set forth in the following description, and in part will become apparent to those skilled in the art from the description, or can be learned by practice of the present application. The objects and other advantages of the present application will be realized and attained by the structure particularly pointed out in the description and the appended drawings. BRIEF DESCRIPTION OF DRAWINGS

[0033] Figure 1 The step flowchart of the present application. DETAILED DESCRIPTION

[0034] The technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all the embodiments. The specific embodiments described herein are only used to explain the present application, and are not used to limit the present application. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor are within the scope of protection of the present application.

[0035] The present application provides a roll surface single / double-sided correction determination method as shown in Figure 1 The method comprises the following steps:

[0036] The roll surface to be corrected is scanned and uploaded;

[0037] The uploaded roll surface to be corrected is sampled to obtain a first sample roll surface set;

[0038] The first sample roll surface set is subjected to roll surface determination, and after determination, the first determination logic is executed;

[0039] If the first determination logic determines the correction mode, the determination is ended;

[0040] If the first determination logic does not confirm the correction mode, a second sample paper set is obtained by sampling the paper to be corrected;

[0041] The second paper set is used to perform paper determination, and the second determination logic is executed after the determination to confirm the correction mode.

[0042] The first sample paper set and the second sample paper set can be the same sample paper set;

[0043] Specifically, when the first determination logic does not confirm the correction mode, the first sample paper set is directly used as the second sample paper set without re-sampling the paper to be corrected, thereby reducing the redundant operation of sample extraction and improving the determination efficiency;

[0044] The same sample paper set can share the completed paper determination result (i.e., the identification result of the front and back of each student's paper) when used for the first determination logic and the second determination logic, so that the paper determination step does not need to be repeated for the second determination logic, thereby avoiding waste of data processing resources;

[0045] If the first sample paper set and the second sample paper set are the same sample paper set, the sample coverage, the number of student papers, and the image information of a single sample remain consistent during the execution of the two determination logics, which ensures the continuous execution of the determination logic based on the same sample data and guarantees the coherence and accuracy of the correction mode determination result.

[0046] The paper determination further refers to the paper determination of the scanned paper information of each student, that is, the first scanned paper image of each student is defined as the front, and the second scanned paper image is defined as the back. In essence, the front and back of the student paper in all samples are identified once. After this step, the first and second images in the student image will be explained as the front and back in the following description;

[0047] Specifically, for the scanning data of a single student, the first scanned paper image is defined as the front, and the second scanned paper image is defined as the back. If a single student scans only a single paper (i.e., there is only one image), the image is defaulted as the front, and the back is marked as "none".

[0048] A unified front and back identification process is performed on the paper information of all students in the first sample paper set or the second sample paper set: based on the front and back definition of a single student, the reference identification of "front" and "back" in the sample is established to clearly distinguish the first face (front) and the second face (back) corresponding to each student's paper image.

[0049] After the roll surface determination step, the first and second determination logics executed subsequently both take "front" and "back" as the standard terms to describe the first and second surface images of the student's roll surface, ensuring consistency in the reference to the roll surface in different determination logics and avoiding determination errors caused by term confusion.

[0050] The front and back identification results will serve as the basis data for the first and second determination logics. When the first and second sample roll surface sets are the same sample set, the identification results can be directly reused without repeated calculation.

[0051] The roll surface determination can be further determined by image similarity recognition and / or text ocr recognition;

[0052] Among them, the image similarity refers to obtaining the front and back images of student A, randomly defining the first image as the front and the second image as the back, comparing the similarity of the images of other students in the sample with the front and / or back of student A, and thus confirming the front and back of other students in the sample; the text ocr recognition determination refers to obtaining the front and back images of student A, randomly defining the first image as the front and the second image as the back, obtaining the ocr text in the front and the ocr text in the back, and then performing ocr recognition on each image of other students in the sample, and matching the similarity with the ocr text of the front and / or back of student A, so as to confirm the front and back of other students;

[0053] Specifically, the image similarity recognition determination: select any student (such as student A) in the sample as the reference object, obtain all the roll surface images scanned by the student, and randomly define the first image as the front image of student A and the second image as the back image of student A; if student A only scans a single image, only define the image as the front image, and the back image is marked as "none".

[0054] Extract the image features (such as layout, question type distribution, frame style, fixed pattern, etc.) of the front and back images of student A, and establish the front and back feature libraries.

[0055] For each scanned image of other students in the first or second sample roll surface set, extract the image features respectively, and perform similarity calculation (such as through histogram matching, feature point matching, etc.) with the front and back feature libraries of student A.

[0056] If the similarity of an image of a student to the front image of student A is higher than a preset similarity threshold (e.g., 80%), the image is determined to be the front of the student; if the similarity to the back image of student A is higher than the preset similarity threshold, it is determined to be the back of the student; if both the similarity thresholds to the front and back are met, the scanning order is combined to assist in the determination (e.g., the image in the scanning order at the front is matched to the front).

[0057] Text OCR recognition determination: Similarly, student A is selected as the reference object, and OCR text recognition is performed on the front image and back image defined for student A to extract front text information (e.g., test paper title, question number range, specific instruction text, etc.) and back text information, and to establish a front text library and a back text library.

[0058] OCR text recognition is performed on each scanned image of other students in the sample to obtain the text content, and text preprocessing (e.g., removing redundant symbols, standardizing formats, etc.) is performed.

[0059] The preprocessed text content is compared with the front text library and the back text library of student A for text similarity (e.g., the text matching degree is calculated by cosine similarity, edit distance, etc.).

[0060] If the text of an image of a student has a similarity to the front text library of student A that is higher than a preset text threshold (e.g., 70%), it is determined to be the front of the student; if the similarity to the back text library is higher than the preset text threshold, it is determined to be the back; if there is a cross-matching case, unique identification information (e.g., specific page number, chapter number) in the text can be used for further differentiation.

[0061] Fusion determination logic: Image similarity recognition or text OCR recognition can be used alone for test paper determination; both methods can be combined, when the two recognition results are consistent, the front and back are directly confirmed; when the results are inconsistent, the final determination result can be determined by weighted calculation (e.g., image similarity weight 60%, text similarity weight 40%) or preset priority (e.g., text OCR recognition result priority higher than image similarity recognition) to improve the accuracy of front and back recognition.

[0062] The first determination logic is a trace determination logic, and the corresponding second determination logic is a determination logic of whether to answer. The trace determination logic refers to determining whether there is an identification of having been marked in the front and back images. The front and back of the first sample are obtained. If the front and back have been marked, neither the front nor the back continues to execute the identification and marking logic. If only one page of the front and back is marked, it is determined that the single-page marking mode is used, and the page without the mark is specified as the target page of the marking. If neither the front nor the back is marked, the second determination logic is continued to execute. The determination logic of whether to answer further refers to determining whether there is a trace of student handwriting answer in the front and back of the student. If neither the front nor the back is answered, it is determined that the paper is not marked. If only one page of the front and back is answered, it is determined that the single-page marking mode is used, and the page with the answer is specified as the top page of the marking. If the front and back are answered, it is determined that the double-page marking logic is used, and the front and back are both identified and marked.

[0063] Specifically, the trace determination logic specifically includes:

[0064] The "identification of having been marked" is defined as a visual mark with a clear marking attribute, including but not limited to a red cross, a two-dimensional code with a marking record, a top display of a marking image (such as a teacher's handwritten annotation screenshot), and a specific format of a marking text mark (such as a "read" "corrected" and the like).

[0065] The trace identification of each student's front and back images in the first sample paper is detected, and the above-mentioned identification features are located and extracted through image recognition technology (such as a target detection algorithm).

[0066] Based on the detection result, it is determined that if the front and back images of a student both detect the defined trace identification, and the integrity of the trace identification (such as the complete display of the cross, the recognizability of the two-dimensional code) meets the preset standard, it is determined that the front and back of the student have been marked, and the subsequent identification and marking logic is not executed for the two pages of the student.

[0067] If only the front or the back detects the trace identification, and the other page does not detect any trace identification, it is determined that the single-page marking mode is used, and the page without the mark is specified as the target page of the marking.

[0068] If neither the front nor the back detects the trace identification, or the trace identification does not meet the integrity standard (such as only a part of the cross trace is left), it is determined that the trace determination logic does not confirm the marking mode, and the second determination logic is triggered.

[0069] The determination logic of whether to answer specifically includes: the "trace of student handwriting answer" is defined as the handwriting, symbols, drawings and the like written by the student in the answer area of the paper (such as the answer box, the horizontal line area and the like), excluding printed texts, blank areas and the like non-answer traces.

[0070] Handwriting detection is performed on the sample cover of the second determination logic (including the first sample cover set or the second sample cover set), and handwriting features in the answer area are extracted by a handwriting recognition algorithm (such as a handwriting segmentation model based on deep learning).

[0071] Based on the detection result, a determination is made: if no valid handwriting is detected in the answer area of the front and back images, or the area ratio of the handwriting is lower than a preset answer threshold (such as 5%), it is determined to be a blank cover, and this time no correction is performed.

[0072] If only the front or back answer area detects valid handwriting, and the handwriting area ratio of the other page is lower than the preset threshold, it is determined to be a single-page correction mode, and the answer page is marked as the top correction target.

[0073] If valid handwriting is detected in the answer area of the front and back, and both satisfy the preset answer threshold, it is determined to be a double-page correction mode, and the front and back pages are both executed with the recognition correction logic.

[0074] The first determination logic is a whether-to-answer determination logic, and the corresponding second determination logic is a trace determination logic; the whether-to-answer determination logic further refers to determining whether there is a trace of student handwriting answer in the front and back cover of the student; if neither the front nor the back is answered, it is determined that the test paper is not corrected this time; if only one of the front and back is answered, it is determined to be a single-page correction, and the page to be corrected is the answered page; if both the front and back are answered, the second determination logic is executed; the trace determination logic refers to determining whether there is a trace of correction in the front and back images: obtaining the front and back of the first sample, if both the front and back have left traces, neither the front nor the back continues to execute the recognition correction logic; if only one of the front and back has left traces, it is determined to be a single-page correction, and the page to be corrected is the front or back that has not left traces; if neither the front nor the back has left traces, it is determined to be a double-page correction logic, and both the front and back are recognized and corrected.

[0075] Specifically, the determination range of the "trace of student handwriting answer" is limited to the pre-set answer area on the test paper (such as the answer box below the question, the specified horizontal line area, the graph drawing area, etc.), and the handwriting traces in the non-answer area (such as student's name, class, etc. Identifying text) such as the edge of the test paper, the header and footer are excluded.

[0076] The handwriting recognition technology (such as a handwriting detection model based on deep learning) is used to scan the front and back images of each student in the first sample cover set, extract handwriting features (including stroke trajectory, ink color, writing coherence, etc.) in the answer area, and distinguish them from the original printed text features of the test paper.

[0077] Combining the quantitative standard to execute the judgment: if the handwriting in the answer area of the front and back images does not meet the characteristics, or the cumulative length and area ratio of the handwriting is less than the preset answer threshold (such as the handwriting coverage area of a single page answer area being less than 5%), it is determined to be a blank paper, and the correction logic is not executed this time;

[0078] If only the front or back answer area detects handwriting that meets the characteristics, and the handwriting on the other page does not reach the answer threshold, it is determined to be a single-page correction mode, the answer page is placed on top as the target page for this correction, and the front and back properties of the page (based on the previous paper determination result) are recorded;

[0079] If the answer area of the front and back images detects handwriting that meets the characteristics and reaches the answer threshold, the whether answer determination logic does not confirm the correction mode, and triggers the second determination logic (trace determination logic);

[0080] The trace determination logic specifically includes: defining the "already corrected trace" as a mark with a clear correction attribute, including but not limited to the correction symbols used by the teacher when correcting, such as red (or other preset color) hooks, crosses, wavy lines, etc., two-dimensional codes with correction records (containing correction time, corrector, etc.), system-generated top correction marks (such as text boxes with "corrected" words, special style border marks), etc.;

[0081] Using image recognition technology (such as target detection algorithm) to scan the sample paper (which can be the same as the first sample paper set) entering the second determination logic, locate and identify the above-mentioned trace marks, and verify the validity of the marks (such as two-dimensional codes can be parsed, hook and cross symbols are complete and can be distinguished);

[0082] Based on the identification result, if the front and back images of a student both detect valid trace marks, and the number and coverage of the marks meet the preset standard (such as at least one complete mark on each side), it is determined that the front and back of the student have been corrected, and the subsequent recognition and correction logic is not executed on these two pages;

[0083] If only the front or back detects valid trace marks, and the other page does not detect any trace marks, it is determined to be a single-page correction mode, and the page without trace marks is specified as the target page for this correction.

[0084] In the first determination logic, if the image ratio of the sample front or back trace is greater than the first preset threshold, it is determined that the front has been traced or the back has been traced; in the second determination logic, if the image ratio of the student's answer in the sample is greater than the second preset threshold, it is determined that the front has been traced or the back has been answered;

[0085] Specifically, the image ratio of the front or back of the sample leaving marks refers to the proportion of the number of student images with valid mark identification detected on the front to the total number of student images on the front in the first sample set, and the proportion of the number of student images with valid mark identification detected on the back to the total number of student images on the back in the first sample set.

[0086] The first preset threshold is a proportion threshold (such as 60%) set based on sample representativeness. When the front mark image ratio exceeds the threshold, it is determined that the front of the current scanned test paper of the class has been marked as a whole. When the back mark image ratio exceeds the threshold, it is determined that the back of the current scanned test paper of the class has been marked as a whole.

[0087] If the mark image ratios of the front and back both exceed the first preset threshold, the supplementary determination is performed on the individual student pages without marks in combination with the mark detection results of individual students (such as giving priority to the class overall determination result, and defaulting that the student page has been marked).

[0088] Application of the second preset threshold: The front or back image ratio of the student's answer in the sample refers to the proportion of the number of student images with valid answer marks detected on the front to the total number of student images on the front in the second sample set (which can be the same as the first sample set), and the proportion of the number of student images with valid answer marks detected on the back to the total number of student images on the back in the sample set.

[0089] The second preset threshold is a proportion threshold (such as 50%) set based on answer universality. When the front answer image ratio exceeds the threshold, it is determined that the front of the current scanned test paper of the class has answers as a whole. When the back answer image ratio exceeds the threshold, it is determined that the back of the current scanned test paper of the class has answers as a whole.

[0090] If the answer image ratios of the front and back both exceed the second preset threshold, the double-sided correction logic is directly triggered. If only the ratio of one side exceeds the threshold, the secondary verification is performed on the individual student pages that do not meet the standard in combination with the answer detection results of individual students (such as confirming whether it is a false judgment caused by scanning blur), and finally determining the single-page correction mode of the side.

[0091] The paper face single / double-sided correction determination device comprises:

[0092] The scanning unit is used for scanning the test paper to be corrected.

[0093] The identification unit is used for sample extraction, paper face determination, and execution of determination logic to determine the correction mode on the scanned test paper to be corrected, including front and back identification of the sample paper, paper face determination through image similarity recognition and / or text ocr recognition, execution of mark determination logic and answer determination logic.

[0094] a sample extraction unit configured to extract a first sample set and a second sample set from the scanned to-be-corrected manuscript;

[0095] a determination logic execution unit configured to execute a first determination logic and a second determination logic, wherein the first determination logic is a trace determination logic or a whether-to-answer determination logic, the second determination logic is another determination logic corresponding to the first determination logic, and a correction mode is determined according to a determination result;

[0096] a threshold setting unit configured to preset various determination thresholds, including a first preset threshold, a second preset threshold, a third preset threshold, a fourth preset threshold, a fifth preset threshold, a sixth preset threshold, and the like, to provide a quantitative standard for execution of the determination logic.

[0097] An electronic device includes a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the manuscript single / double-sided correction determination method when executing the computer program.

[0098] A computer readable storage medium stores non-transitory computer readable instructions, which, when executed by a computer, can implement the manuscript single / double-sided correction determination method.

[0099] Working principle: scan and upload the to-be-corrected manuscript; perform sample extraction on the uploaded to-be-corrected manuscript to obtain a first sample set; perform manuscript determination on the first sample set, and then execute a first determination logic after the determination; if the first determination logic determines a correction mode, the determination ends; if the first determination logic does not confirm the correction mode, perform sample extraction on the to-be-corrected manuscript to obtain a second sample set; perform manuscript determination on the obtained second sample set, and then execute a second determination logic after the determination to confirm the correction mode.

[0100] In addition, the scanning unit, the recognition unit, the sample extraction unit, and the threshold setting unit determination logic execution unit are also used to implement other functions of the manuscript single / double-sided correction determination method when executed, which are not described here.

[0101] In addition, the present application also provides a terminal device, and the manuscript single / double-sided correction determination method involved in the present embodiment is mainly applied to the terminal device, which can be a PC, a portable computer, a mobile terminal, or the like.

[0102] Specifically, the terminal device can include a processor (for example, a CPU), a communication bus, a user interface, a network interface, and a memory. The communication bus is used to realize the connection communication between these components; the user interface can include a display screen (Display) and an input unit such as a keyboard (Keyboard); the network interface can optionally include a standard wired interface and a wireless interface (such as a WI-FI interface); the memory can be a high-speed RAM memory, or a stable memory (non-volatile memory) such as a disk memory, and the memory can also be a storage device independent of the aforementioned processor.

[0103] The memory stores a readable storage medium, and the readable storage medium stores a single / double-sided correction judgment program. The processor can call the single / double-sided correction judgment program stored in the memory and execute the single / double-sided correction judgment method provided by the embodiment of the application.

[0104] It can be understood that the readable storage medium can be a tangible device that can retain and store instructions for use by an instruction execution device. The computer readable storage medium can be, for example but not limited to, an electronic storage device, a magnetic storage device, an optical storage device, an electromagnetic storage device, a semiconductor storage device, or any suitable combination of the above. More specific examples (a non-exhaustive list) of the computer readable storage medium include a portable computer diskette, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), a static random access memory (SRAM), a portable compact disc read-only memory (CD-ROM), a digital versatile disk (DVD), a memory stick, a floppy disk, a mechanical encoding device, such as a punched card or a concave-convex structure in a slot, and any suitable combination of the above. The computer readable storage medium used herein is not to be interpreted as a transitory signal per se, such as a radio wave or other freely propagating electromagnetic wave, an electromagnetic wave propagating through a waveguide or other transmission medium (for example, an optical pulse through an optical fiber cable), or an electrical signal transmitted through a wire.

[0105] The computer readable program instructions described herein can be downloaded from the computer readable storage medium to each computing / processing device, or to an external computer or an external storage device through a network, such as the Internet, a local area network, a wide area network, and / or a wireless network. The network can include copper transmission cables, fiber-optic transmission, wireless transmission, routers, firewalls, switches, gateway computers, and / or edge servers. The network adapter card or network interface in each computing / processing device receives the computer readable program instructions from the network and forwards the computer readable program instructions to the computer readable storage medium in each computing / processing device for storage.

[0106] Computer readable program instructions for carrying out operations of the present disclosure can be assembly instructions, instruction-set-architecture (ISA) instructions, machine instructions, machine dependent instructions, microcode, firmware instructions, state-setting data, or either source code or object code written in any combination of one or more programming languages, including an object oriented programming language such as Smalltalk, C++ or the like, and conventional procedural programming languages, such as the "C" programming language or similar programming languages. The computer readable program instructions can execute entirely on the user's computer, partly on the user's computer, as a stand-alone software package, partly on the user's computer and partly on a remote computer or entirely on the remote computer or server. In the latter scenario, the remote computer can be connected to the user's computer through any type of network, including a local area network (LAN) or a wide area network (WAN), or the connection can be made to an external computer (for example, through the Internet using an Internet Service Provider). In some embodiments, electronic circuitry including, for example, programmable logic circuitry, field-programmable gate arrays (FPGA), or programmable logic arrays (PLA) can execute the computer readable program instructions by utilizing state information of the computer readable program instructions to personalize the electronic circuitry, in order to perform aspects of the present disclosure.

[0107] Finally, it should be noted that the above-mentioned only is the preferred embodiment of the present application, and does not limit the present application, although the foregoing embodiments of the present application are described in detail, for those skilled in the art, it still can be modified to the technical solution recorded in the foregoing embodiments, or equivalent replacement of part of the technical features, within the spirit and principles of the present application, any modification, equivalent replacement, improvement, etc. made, should be included in the protection scope of the present application.

Claims

1. A method for judging whether a paper is single-sided or double-sided, characterized in that: Includes the following steps: Scan the papers to be graded and upload them; The uploaded papers to be graded are sampled to obtain the first sample paper set. Perform a paper determination judgment on the first sample paper set, and then execute the first judgment logic; If the first judgment logic determines the correction mode, the judgment ends. If the first judgment logic fails to identify the grading mode, then a second sample set of papers will be obtained by sampling the papers to be graded. The second set of papers will be used to determine the paper type. After the determination, the second determination logic will be executed to confirm the grading mode.

2. The method for determining single-sided or double-sided marking of a test paper according to claim 1, characterized in that: The first sample set and the second sample set can be the same sample set.

3. The method for judging single-sided and double-sided marking of a test paper according to claim 1, characterized in that: The execution of the paper surface determination further refers to performing paper surface determination on the paper surface information scanned by each student. That is, the first image of the test paper scanned by each student is defined as the front side and the second image as the back side. In essence, the front and back sides of the student's paper surface in all samples are identified. After this step, the first and second images in the student's image will be described in terms of front and back sides in the following text.

4. The method for determining single-sided or double-sided marking according to claim 3, characterized in that: The determination of the paper surface can be further made through image similarity recognition and / or text OCR recognition; Image similarity refers to acquiring front and back images of student A, randomly defining the first image as the front and the second as the back, and comparing the similarity of other students' images in the sample with the front and / or back images of student A to confirm the front and back of other students in the sample. Text OCR recognition refers to acquiring front and back images of student A, randomly defining the first image as the front and the second as the back, acquiring the OCR text in the front and the OCR text in the back, and then performing OCR recognition on each image of other students in the sample. By matching the similarity with the OCR text of student A's front and / or back images, the front and back of other students can be confirmed.

5. The method for judging single-sided and double-sided marking of a test paper according to claim 1, characterized in that: The first determination logic is the trace-keeping determination logic, and the corresponding second determination logic is the answer-whether-answered determination logic. The trace-keeping determination logic refers to determining whether there are any marking marks in the front and back images: obtain the front and back of the first sample. If both the front and back have marking marks, then the recognition and marking logic will not continue to be executed on either the front or back. If only one page of the front and back has marking marks, then it is determined to be a single-page marking, and the page to be marked is the front or back without marking marks. If neither the front nor back has marking marks, then the second determination logic will continue to be executed. The answer-whether-answered determination logic further refers to determining whether there are any traces of handwritten answers in the front and back of the student's paper. If neither the front nor back has answers, then it is determined that the paper will not be marked. If only one page of the front and back has answers, then it is determined to be a single-page marking, and the page marked at the top is the page with answers. If both the front and back have answers, then it is determined to be a double-sided marking logic, and recognition and marking will be performed on both the front and back.

6. The method for determining single-sided or double-sided marking of a test paper according to claim 1, characterized in that: The first determination logic is the "whether an answer has been given" logic, and the corresponding second determination logic is the "traceability" logic. The "whether an answer has been given" logic further refers to determining whether there are traces of handwritten answers on both sides of the student's exam paper. If neither side has been answered, the exam paper is determined not to be graded. If only one page on both sides has been answered, it is determined to be a single-page grade, and the page marked at the top is the page with the answer. If both sides have been answered, the second determination logic continues to be executed. The "traceability" logic refers to determining whether there are any marking marks on the images of both sides: obtain the front and back sides from the first sample. If both sides have been marked, the marking logic will not continue to be executed on either side. If only one page on both sides has been marked, it is determined to be a single-page grade, and the page to be marked is the front or back side without a mark. If neither side has been marked, it is determined to be a double-sided grade, and both sides will be marked.

7. The method for judging single-sided and double-sided marking of a test paper according to claim 5, characterized in that: In the first determination logic, if the ratio of images with marks on the front or back of the sample is greater than the first preset threshold, it is determined that marks have been left on the front or back. In the second determination logic, if the ratio of images of the front or back of the student's answer in the sample is greater than the second preset threshold, it is determined that marks have been left on the front or the answer has been given on the back.

8. A device for determining whether a paper is single-sided or double-sided, comprising: Scanning unit: Used to scan the paper to be graded; Recognition Unit: Used to extract samples from the scanned papers to be graded, determine the papers, and execute judgment logic to determine the grading mode. This includes front and back recognition of the sample papers, paper determination through image similarity recognition and / or text OCR recognition, execution of trace determination logic, and whether an answer has been given. Sample extraction unit: used to extract the first sample set and the second sample set from the scanned test papers to be graded; Judgment Logic Execution Unit: Used to execute the first judgment logic and the second judgment logic, wherein the first judgment logic can be the record-keeping judgment logic or the whether to answer judgment logic, and the second judgment logic is another judgment logic corresponding to the first judgment logic, and the grading mode is determined according to the judgment result; Threshold setting unit: used to preset various judgment thresholds, including the first preset threshold, the second preset threshold, the third preset threshold, the fourth preset threshold, the fifth preset threshold, the sixth preset threshold, etc., to provide quantitative standards for the execution of judgment logic.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the method for determining single-sided or double-sided marking of a paper as described in any one of claims 1-8.

10. A computer-readable storage medium storing non-transitory computer-readable instructions, which, when executed by a computer, can implement the single-sided / double-sided marking determination method according to any one of claims 1-8.

Images