CN116303871A - Exercise book reading method - Google Patents
Exercise book reading method Download PDFInfo
- Publication number
- CN116303871A CN116303871A CN202310275971.6A CN202310275971A CN116303871A CN 116303871 A CN116303871 A CN 116303871A CN 202310275971 A CN202310275971 A CN 202310275971A CN 116303871 A CN116303871 A CN 116303871A
- Authority
- CN
- China
- Prior art keywords
- answer
- answer sheet
- student
- correction
- information
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
- G06F16/31—Indexing; Data structures therefor; Storage structures
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
- G06F16/34—Browsing; Visualisation therefor
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F40/00—Handling natural language data
- G06F40/10—Text processing
- G06F40/103—Formatting, i.e. changing of presentation of documents
- G06F40/106—Display of layout of documents; Previewing
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F40/00—Handling natural language data
- G06F40/10—Text processing
- G06F40/166—Editing, e.g. inserting or deleting
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/044—Recurrent networks, e.g. Hopfield networks
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/0464—Convolutional networks [CNN, ConvNet]
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/048—Activation functions
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/764—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using classification, e.g. of video objects
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/82—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using neural networks
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V30/00—Character recognition; Recognising digital ink; Document-oriented image-based pattern recognition
- G06V30/10—Character recognition
- G06V30/19—Recognition using electronic means
- G06V30/191—Design or setup of recognition systems or techniques; Extraction of features in feature space; Clustering techniques; Blind source separation
- G06V30/19173—Classification techniques
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02D—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE
- Y02D10/00—Energy efficient computing, e.g. low power processors, power management or thermal management
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- General Health & Medical Sciences (AREA)
- General Engineering & Computer Science (AREA)
- Health & Medical Sciences (AREA)
- Artificial Intelligence (AREA)
- Evolutionary Computation (AREA)
- Software Systems (AREA)
- Computing Systems (AREA)
- Data Mining & Analysis (AREA)
- Computational Linguistics (AREA)
- Biophysics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Biomedical Technology (AREA)
- Databases & Information Systems (AREA)
- Molecular Biology (AREA)
- Mathematical Physics (AREA)
- Multimedia (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Audiology, Speech & Language Pathology (AREA)
- Medical Informatics (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
- Electrically Operated Instructional Devices (AREA)
Abstract
The invention discloses a training book reading method, which comprises the steps of storing data in a database, generating answer sheets, uploading the answer sheets, automatically reading and manually reading, wherein the answer sheets can be automatically set through student information, test question numbers, answers and teacher custom information in the database; after uploading answer sheets for answering, the answer area of each question of the student answer sheet picture is subjected to screenshot, the difference between the subjective questions and the objective questions is identified, and the automatic correction of the objective questions is realized by utilizing OCR and confusion-prone analysis; the subjective questions are displayed independently, and the teacher manually reads the subjective questions. According to the scheme, manual correction is arranged, so that manual correction of subjective questions can be realized, in addition, objective questions can be automatically corrected, multi-disciplinary selection and gap filling can be simultaneously supported, the limitation that a traditional intelligent correction system only supports correction of the selection questions is expanded, and the workload of a teacher in correction operation is reduced; the earlier workload of the traditional intelligent reading system and the use cost of teachers are greatly reduced.
Description
Technical Field
The invention relates to the technical field of operation reading, in particular to an exercise book reading method.
Background
At present, educational software in China only can solve the correction of selection questions and judgment questions, but cannot automatically identify and read calculation questions, gap filling questions and other questions, and the complete test questions are required to be input into a question library in advance; the research and development of the related fields outside the country are almost the same as those of the domestic situation; for example, e-book communication of a science net only supports automatic correction of objective questions, and a required test question is required to be input into a question library in advance, and then the automatic correction function can be used after the book is assembled, so that flexibility is not provided; the bee operation supports the correction of various questions, but the bee operation also needs to be manually corrected or proofread after the automatic correction, and the teacher workload is reduced, but the bee operation also needs to be corrected by the teaching and research assistance of a company, the labor cost is only transferred from the teacher to the teaching and research of the company, the bee operation is not reduced per se, and additional economic burden is brought to schools or teachers; the traditional exercise books are required to be added with dot matrix codes, the special exercise books are used, the limitation is very high, the network-accessible intelligent learning terminal is slightly better than the previous schemes, objective questions can be automatically corrected on the basis of the original paper exercise books, subjective questions are corrected by teachers, but the technology based on photographing and searching questions is adopted, so that the system is inevitably required to record all test question information in advance, and only twenty books are provided for each school for free recording every year, and the use of different exercise books with different grades and different subjects is also limited.
The chinese patent discloses a general automatic reading method for exercise books, publication No. CN114419639a. When the modification is carried out, the whole body relies on automatic reading, but the answer of the part of subjective questions is not unique, and the reading of the subjective questions is difficult to realize.
The invention mainly aims to provide a exercise book reading method, which aims to solve the problems that in the related technology, when the exercise book is modified, the whole exercise book is automatically read, but the answer is not unique for part of subjective questions, and the reading of the subjective questions is difficult to realize.
In order to achieve the above object, the present invention provides an exercise book reading method, including S1, database storing data: firstly, student information, test question numbers, answers and teacher custom information are stored in a database;
s2, generating an answer sheet: the answer sheet can be automatically set through student information, test question numbers, answers and teacher custom information in the database;
s3, uploading an answer sheet: after the answer sheet is generated, the answer sheet can be uploaded;
s4, automatic reading: after uploading answer sheets for answering, the answer area of each question of the student answer sheet picture is subjected to screenshot, the difference between the subjective questions and the objective questions is identified, and the automatic correction of the objective questions is realized by utilizing OCR and confusion-prone analysis;
s5, manual reading: the subjective questions are displayed independently, and the teacher manually reads the subjective questions.
In the embodiment of the present invention, in the step S1, the database storage data includes storage student UUID information, storage exercise book key information, storage generated answer sheets and storage correction results, where the storage student UUID information corresponds to a school, a grade and a class where a student is located, is used to generate a student information identifier, classifies uploaded test papers according to the student, then automatically sorts and counts and analyzes the answer situation of each student, the storage exercise book key information is used to locate each exercise question in the exercise exercises, automatically set the size of the answer sheet answer area and automatically correct the answer sheet, the storage generated answer sheets are used for a teacher to quickly download and use at any time, and the storage correction results are used to subsequently check and count each student, score situation of each question, and further analyze the depth of the student knowledge point grasp situation.
In the embodiment of the invention, the key information of the storage exercise book comprises chapters, page numbers, question numbers and answers, wherein the chapters, the page numbers and the question numbers are used for positioning each exercise question in exercise questions, and the answers are used for automatically setting the size of an answer sheet answer area and automatically correcting; the answer sheet generated by storage comprises an answer sheet and an answer sheet template, wherein the answer sheet is used for a teacher to download and use quickly at any time, and the answer sheet template is used for cutting out an answer area in the answer sheet after the student answers.
In the embodiment of the invention, in the step S2, the generating the answer sheet includes generating a reserved identifier, generating an answer area, generating a complete answer sheet and a template, wherein the generating the reserved identifier generates the reserved identifier according to a UUID of a student, a chapter, a page number and a question number of a problem in a database, the generating the answer area automatically sets the answer area of each problem and a color block area corresponding to the template according to the question number and answer information in the database and the individualized setting of a teacher, and the answer sheet is used for being downloaded by the student and then is used for answering on paper on the template in the generating the complete answer sheet and the template, and the template is used for cutting the answer area of each problem of the student.
In the embodiment of the invention, the generated reserved identifier comprises a student information identifier, an answer sheet correction identifier and a page number identifier, wherein the student information identifier is positioned in a two-dimensional code at the upper right corner of the answer sheet or the test paper, the content is a 32-bit UUID, the answer sheet correction identifier is positioned in the upper left corner, the upper right corner and the lower left corner of the answer sheet or the test paper and is used for identifying the square at the lower right corner of the answer sheet or the test paper, if inclination exists, the correction is performed, and the page number identifier is positioned at the top end identifier of the answer sheet or the test paper and is a hollow or solid square in the middle and is used for identifying and judging the page number of the answer sheet.
In the embodiment of the invention, in the step S3, the uploading answer sheet includes answer sheet scanning, answer sheet correction, obtaining page number information, obtaining student information and obtaining answer content, the answer sheet scanning is performed on the answer sheet by a teacher by using scanning equipment, the answer sheet correction corrects the answer sheet according to a correction identifier, the obtaining page number information determines a position according to the page number identifier, the obtaining student information binds a corresponding student according to the student identifier, the obtaining answer content obtains a corresponding template from a database, matching is performed, and an answer area is obtained by comparing the template.
In the embodiment of the invention, the scanning equipment comprises a mobile phone and a scanner, and a teacher takes a picture by using the mobile phone or acquires an answer sheet image by using the scanner.
In the embodiment of the invention, in the step S4, the automatic reading includes performing OCR on an answer area, performing confusing analysis and automatic correction, wherein a model adopted by the OCR on the answer area is a cyclic convolutional neural network-based and connection-based time sequence classification technology, the confusing analysis is used for further compensating for accurate recognition of a student writing area under a complex scene, compensating for the situation of low OCR recognition accuracy caused by redundant handwriting, different writing habits and writing, irregular light and shade changes of an answer card, and the like, the automatic correction is that the answer content of the student is matched with an answer in a database, the answer is used for objective question judgment, and the correction result is stored in the database.
In an embodiment of the invention, the recurrent convolutional neural network is a combination of a convolutional neural network and a recurrent neural network.
In the embodiment of the present invention, in S5, after the subjective questions are modified, the results are also stored in the database, and are combined with the scores of the corresponding objective questions for statistics.
Compared with the prior art, the invention has the beneficial effects that:
the manual correction is arranged, so that the subjective questions can be manually corrected, in addition, the objective questions can be automatically corrected, the selection and filling of multiple subjects can be simultaneously supported, the limitation that the traditional intelligent correction system only supports correction of the selection questions is expanded, and the workload of a teacher in correction operation is reduced; the earlier workload of the traditional intelligent reading system and the use cost of teachers are greatly reduced.
Drawings
FIG. 1 is a flow chart of an exercise book reading method according to an embodiment of the present invention;
FIG. 2 is a schematic diagram of the contents of database storage data of the exercise book reading method according to the embodiment of the invention;
fig. 3 is a schematic content diagram of an answer sheet generated by the exercise book reading method according to an embodiment of the present invention;
fig. 4 is a schematic content diagram of an uploading answer sheet of an exercise book reading method according to an embodiment of the present invention;
FIG. 5 is a schematic diagram of automatic reading content of an exercise book reading method according to an embodiment of the present invention;
fig. 6 is a schematic diagram of a complete answer sheet of an exercise book reading method according to an embodiment of the present invention.
Description of the embodiments
In order that those skilled in the art will better understand the present invention, a technical solution in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in which it is apparent that the described embodiments are only some embodiments of the present invention, not all embodiments. All other embodiments, which can be made by those skilled in the art based on the embodiments of the present invention without making any inventive effort, shall fall within the scope of the present invention.
It should be noted that the terms "first," "second," and the like in the description and the claims of the present invention and the above figures are used for distinguishing between similar objects and not necessarily for describing a particular sequential or chronological order. It is to be understood that the data so used may be interchanged where appropriate in order to describe the embodiments of the invention herein. Furthermore, the terms "comprises," "comprising," and "having," and any variations thereof, are intended to cover a non-exclusive inclusion, such that a process, method, system, article, or apparatus that comprises a list of steps or elements is not necessarily limited to those steps or elements expressly listed but may include other steps or elements not expressly listed or inherent to such process, method, article, or apparatus.
In the present invention, the terms "upper", "lower", "left", "right", "front", "rear", "top", "bottom", "inner", "outer", "middle", "vertical", "horizontal", "lateral", "longitudinal" and the like indicate an azimuth or a positional relationship based on that shown in the drawings. These terms are only used to better describe the present invention and its embodiments and are not intended to limit the scope of the indicated devices, elements or components to the particular orientations or to configure and operate in the particular orientations.
Also, some of the terms described above may be used to indicate other meanings in addition to orientation or positional relationships, for example, the term "upper" may also be used to indicate some sort of attachment or connection in some cases. The specific meaning of these terms in the present invention will be understood by those of ordinary skill in the art according to the specific circumstances.
In addition, the term "plurality" shall mean two as well as more than two.
It should be noted that, without conflict, the embodiments of the present invention and features of the embodiments may be combined with each other. The invention will be described in detail below with reference to the drawings in connection with embodiments.
Examples
Referring to fig. 1 to 6, the invention provides a exercise book reading method, which includes S1, storing data in a database: firstly, student information, test question numbers, answers and teacher custom information are stored in a database;
s2, generating an answer sheet: the answer sheet can be automatically set through student information, test question numbers, answers and teacher custom information in the database;
s3, uploading an answer sheet: after the answer sheet is generated, the answer sheet can be uploaded;
s4, automatic reading: after uploading answer sheets for answering, the answer area of each question of the student answer sheet picture is subjected to screenshot, the difference between the subjective questions and the objective questions is identified, and the automatic correction of the objective questions is realized by utilizing OCR and confusion-prone analysis;
s5, manual reading: the subjective questions are displayed independently, and the teacher manually reads the subjective questions.
The answer sheet can be automatically set through student information, test question numbers, answers and teacher custom information in the database, and automatic correction of objective questions can be realized by utilizing OCR and confusion analysis after the answer sheet is uploaded;
OCR (Optical Character Recognition ) refers to the process of an electronic device (e.g., a scanner or digital camera) checking characters printed on paper, determining their shape by detecting dark and light patterns, and then translating the shape into computer text using a character recognition method; that is, for the print character, the technology of converting the characters in the paper document into the image file of black-white lattice by adopting the optical mode, and converting the characters in the image into the text format by the recognition software for further editing and processing by the word processing software;
the invention automatically corrects objective questions, can simultaneously support the selection and filling of multiple disciplines, expands the limitation that the traditional intelligent reading system only supports the correction of the selection questions, and reduces the workload of teacher correction operation;
the invention only needs to input the question number and answer information in the system, and does not need to input all the information of the stem, the option, the position relation and the like of the exercise questions in advance, thereby greatly reducing the earlier-stage workload of the traditional intelligent reading and writing system and the use cost of teachers.
Referring to fig. 1 and 2, in step S1, the database stores data including storing UUID information of students, key information of exercise books, answer sheets generated by storing, and correction results;
different students have different UUID information, the UUID information of the stored students corresponds to schools, grades and classes where the students are located, the UUID information is used for generating student information identifiers, the system can classify the uploaded test paper according to the students according to the UUID information of the students, then automatically sort and read the test paper, and count and analyze the response condition of each student;
UUID is an abbreviation for universally unique identification code (Universally Unique Identifier);
the key information of the storage exercise book is used for positioning each exercise question in the exercise exercises and automatically setting the size and automatic correction of the answer sheet answer area, the key information of the storage exercise book comprises chapters, page numbers, question numbers and answers, wherein the chapters, page numbers and the question numbers are used for positioning each exercise question in the exercise exercises, the answers are used for automatically setting the size and automatic correction of the answer sheet answer area, the chapters, page numbers and question numbers of the test questions in the exercise book are firstly associated with the answer sheet page numbers and then are associated with the question numbers in the answer sheets, and finally, the answers of the exercise books and the answer sheet question numbers can be directly associated, and system input of the question information or careful design of a specific exercise book is avoided;
the generated answer sheet is stored for a teacher to quickly download and use at any time, the generated answer sheet comprises an answer sheet and an answer sheet template, the answer sheet is used for the teacher to quickly download and use at any time, the answer sheet template is used for cutting out an answer area in an answer sheet after the student answers, the relevance between the answer sheet and an exercise book of the last step is provided, the answer sheet and the template can be generated, that is, when a publisher finally determines an exercise book, the answer sheet and the answer sheet template of the whole exercise book can be designed in batch and are stored in a database, the teacher can download and print all the answer sheets by one key, can set and edit the appointed answer sheet according to teaching requirements on one hand, then download and use the answer sheet again, and after the answer sheet is edited, the template can be updated, and a new answer sheet and the template can be bound with the teacher and the exercise book, so that the use and automatic reading of the answer sheets of other teachers can not be affected;
storing correction results for later checking and counting the scoring condition of each student and each question, and further carrying out deep analysis on the knowledge point mastering condition of the student, wherein the correction results are applied in two aspects: 1. the answer conditions of the whole class and even the whole age are counted, and data with instructive significance for teaching such as lowest score, highest score, average score, median, error-prone questions and the like are analyzed; 2. the method is used for analyzing the answering situation of the single student, providing the learning situation analysis of the single student, including word answering situation, past score curve, knowledge point mastering situation and the like, and being capable of being used for the follow-up personalized recommended exercises of the student and carrying out strengthening exercises aiming at the short board.
Referring to fig. 1, 3 and 6, in step S2, generating an answer sheet includes generating a reserved identifier, generating an answer area, and generating a complete answer sheet and a template;
generating a reserved identifier according to a UUID of a student, chapters, page numbers and question numbers in a database, generating the reserved identifier, wherein the reserved identifier comprises a student information identifier, an answer sheet correction identifier and a page number identifier, the student information identifier is positioned at a two-dimensional code at the upper right corner of the answer sheet or the test paper, the content is a 32-bit UUID corresponding to personal information of the student, the answer sheet correction identifier is positioned at the upper left corner, the upper right corner, the lower left corner and the lower right corner of the answer sheet or the test paper and is used for identifying the direction of a scroll, if inclination exists, correction is carried out, the page number identifier is positioned at the top end identifier of the answer sheet or the test paper and is a square hollowed or solid in the middle and used for identifying and judging the page number of the answer sheet;
generating an answer area, automatically setting the answer area of each question and a color block area corresponding to a template according to the question number and answer information in a database and the individualized setting of a teacher, wherein the answer area is not only a place where a student answers but also a place where each answer correction result of the student and a teacher comment are fed back after homework is automatically corrected, so that the size of the answer area cannot be violently reserved in a exercise book, the individualized setting of the teacher is considered on the basis of the adjustment of the length of the answer, and the teacher can also adjust the specific answer area on the basis of the automatically generated answer sheet, thereby greatly improving the practicability of the answer sheet;
generating complete answer sheets and templates, reserving identifiers and answer areas, generating answer sheets of exercise books and corresponding templates in batches by utilizing a front-end technology, wherein the answer sheets can be used for students to answer on paper after being downloaded, and the templates are used for cutting answer areas of each question of the students;
as shown in FIG. 6, the two-dimensional code at the upper right corner is a student information identifier, the middle square nested pattern at the four corners is an answer sheet correction identifier, the middle hollowed-out or solid square at the upper part is a page identifier, the student information identifier can correspond to and change student's call A', the school is 'middle education and cloud middle school', the grade is 'seven grade', the answer sheet correction identifier is respectively positioned at four corners, the lower right corner is different from the patterns of the other three corners, if the test paper rotates, the test paper can also be corrected according to the designed position relation, and the different answer sheet correction identifier is ensured to be kept at the lower right corner of the corrected image, so that the test paper can be prevented from rotating by 180 degrees, the page identifier adopts a binary mode, the middle hollowed-out square represents 0, the solid square represents 1, and the page represented here is 00000001, namely page 1;
two-dimensional code: the two-dimensional bar code is characterized in that a bar code with readability in another dimension is expanded on the basis of the one-dimensional bar code, black-and-white rectangular patterns are used for representing binary data, information contained in the bar code can be obtained after the bar code is scanned by equipment, the width of the one-dimensional bar code records the data, the length of the one-dimensional bar code does not record the data, the length and the width of the two-dimensional bar code record the data, the two-dimensional bar code is provided with a locating point and a fault-tolerant mechanism which are not provided with the one-dimensional bar code, and the fault-tolerant mechanism can correctly restore the information on the bar code even if all the bar codes are not identified or the bar is stained.
Referring to fig. 1, 4 and 6, in step S3, uploading an answer sheet includes scanning the answer sheet, correcting the answer sheet, acquiring page information, acquiring student information and acquiring answer content;
the answer sheet scanning is carried out by a teacher by using scanning equipment, wherein the scanning equipment comprises a mobile phone and a scanner, the teacher takes pictures by using the mobile phone or obtains answer sheet images by using the scanner, and the mobile phone is portable, but has the problem of slow speed for scanning a plurality of answer sheets, so the teacher can carry out batch scanning on the answer sheets by using the high-speed scanner when in actual use;
because the invention provides the student information identifier, the answer sheet correction identifier and the page number identifier, when a teacher prepares to scan the test paper, the teacher does not need to perform complicated operations such as sequencing, specifying the front and the back, upside down and the like on the test paper, and can directly talk the received test paper into the scanner to obtain all electronic version scanning files;
correcting the answer sheet according to the correction identifier, uploading the electronic version scanning file obtained in the previous step, and correcting the answer sheet according to the answer sheet correction identifier, for example correcting an image with inclined or even reversed position when photographing or scanning the mobile phone, so as to ensure the normal identification and display of the test paper;
acquiring page number information, determining a position according to a page number identifier, wherein the page number identifier adopts a binary mode, a square hollowed out in the middle represents 0, a square solid represents 1, and as shown in fig. 6, the page number is 00000001, namely page 1;
the acquired student information is bound to the corresponding student according to the student identifier, and the student information identifier in FIG. 6 can correspond to the student called "student A", the school is "middle education cloud intelligent middle school", and the grade is "seven grades";
the answering content is acquired from the database, the corresponding templates are matched, the answering area is acquired by comparing the templates, and the acquired answering area can effectively retain the answering information of students and remove irrelevant information.
Referring to fig. 1, 5 and 6, in step S4, after uploading the answer sheet for answering, the answer area of each question of the student answer sheet picture is captured, the difference between the subjective questions and the objective questions is identified, and the automatic reading includes OCR, confusing analysis and automatic correction of the answer area;
the model adopted by OCR in answer area is a cyclic convolutional neural network based on a connective timing classification technology, and the current more mature cyclic convolutional neural network (Convolutional Recurrent Neural Network, CRNN) based on a connective timing classification (Connectionist Temporal Classification, CTC) technology, which is called CRNN+CTC technology for short;
cyclic convolutional neural networks, CRNN, is a combination of convolutional neural networks (Convolutional Neural Networks, CNN) and cyclic neural networks (Recurrent neural networks, RNN);
the scheme of crnn+ctc includes a three-layer structure: CNN layer, RNN layer and CTC layer. The CNN layer is used for extracting depth features of an input image, a VGG network is adopted in an original paper, and ResNet is adopted in the invention, so that the purpose of extracting features by adopting a deeper network is achieved, the RNN layer adopts deep bidirectional RNNs, predicts a tag sequence of a feature sequence obtained from a convolution layer, and the CTC layer is used for replacing a traditional softmax layer;
the data sources of the present invention are three: the method comprises the steps of collecting real data, a public data set and generated data;
the real data acquires a large amount of student writing data of desensitized information from the existing digital course teaching material cloud platform of the company, and the acquired data comprises Chinese and English, mathematics, physics, chemistry and the like, so that the total data volume of effective training reaches tens of millions on the premise of guaranteeing the privacy of the students;
the public data set comprises data sets disclosed by more than ten years of domestic and foreign OCR project competition, and handwriting and printing data of common Chinese and English and formulas are covered;
the generated data mainly aims at the Chinese and English data of handwriting, single characters are extracted through the handwriting data collected in the front, handwriting fonts of more than 2000 people are manufactured, and the designated number of data can be generated according to specific texts, lengths and sizes, so that the data volume is greatly expanded;
the server for training the model comprises 1 server provided with RTX A4000, 4 servers provided with RTX 3090, 6 servers provided with RTX 3080 and a plurality of servers provided with RTX 2080, and can support distributed training;
the mature technology and the optimization scheme, together with huge data volume and the holding of a plurality of large-scale deep learning servers, can recognize handwriting and achieve good effect, can recognize Chinese and English and mathematical formulas, and is compatible with multiple subjects;
the confusing analysis is used for further making up for accurate recognition of the writing area of the student under the complex scene, making up for the condition of low OCR recognition accuracy caused by redundant handwriting, writing habits and writing of different people, irregular brightness change of the answering card and the like, and the confusing characters are difficult to distinguish because the information of the character image is only analyzed without considering front and rear semantics due to the limitation of an OCR network. So that the confusing characters are processed by our confusing analysis, the characters are matched with the rules, and the semantic analysis is carried out according to the front and back writing contents, so that the confusing characters are accurately distinguished;
the automatic correction is that the answer content of the student is matched with the answer in the database, the correction result is used for objective question judgment, the correction result is stored in the database, the correction result is fed back to the student and the teacher, the correct and wrong answer sheets submitted by the student are marked, and the correction data and the correction statistical data of each student are displayed on the interface of the teacher and the background.
Finally, manual reading: the observation questions are displayed independently, the teacher manually reviews, after the subjective questions are revised, the results are stored in the database, and the results are combined with the scores of the corresponding objective questions for statistics.
Specifically, the working principle of the universal exercise book automatic reading method is as follows: when the device is used, firstly, the database stores data, including student UUID information, key information of a training book, answer sheets generated by storage and correction results, namely, student information, test question numbers, answers and teacher custom information are stored in the database, and then, answer sheets are generated and uploaded: generating the answer sheet comprises generating a reserved identifier, generating an answer area, generating a complete answer sheet and a template, uploading the answer sheet comprises answer sheet scanning, answer sheet correcting, obtaining page number information, obtaining student information and obtaining answer contents, namely, the answer sheet can be automatically set through student information, test question numbers, answers and teacher custom information in a database, and after the answer sheet is generated, the answer sheet can be uploaded, and automatic reading is performed: the automatic reading comprises OCR, confusing analysis and automatic correction of answer areas, after the answer sheet is uploaded, the answer area of each question of the student answer sheet picture is subjected to screenshot, the difference between the subjective questions and the objective questions is identified, the automatic correction of objective questions can be realized by utilizing the OCR and the confusing analysis, and finally, the automatic reading is manually carried out: the subjective questions are displayed independently, and the teacher manually reads the subjective questions.
The above description is only of the preferred embodiments of the present invention and is not intended to limit the present invention, but various modifications and variations can be made to the present invention by those skilled in the art. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention should be included in the protection scope of the present invention.
Claims (10)
1. The exercise book reading method is characterized by comprising
S1, storing data in a database: firstly, student information, test question numbers, answers and teacher custom information are stored in a database;
s2, generating an answer sheet: the answer sheet can be automatically set through student information, test question numbers, answers and teacher custom information in the database;
s3, uploading an answer sheet: after the answer sheet is generated, the answer sheet can be uploaded;
s4, automatic reading: after uploading answer sheets for answering, the answer area of each question of the student answer sheet picture is subjected to screenshot, the difference between the subjective questions and the objective questions is identified, and the automatic correction of the objective questions is realized by utilizing OCR and confusion-prone analysis;
s5, manual reading: the subjective questions are displayed independently, and the teacher manually reads the subjective questions.
2. The method for reading exercise books according to claim 1, wherein in the step S1, the database stores data including information of UUID of the students, key information of the exercise books, answer sheets generated by storing, and correction results stored, the information of UUID of the students corresponds to schools, grades and classes where the students are located, the information identifiers of the students are generated, the uploaded test papers are classified according to the students, and then the reading is automatically arranged, and statistics and analysis are performed on the answer situation of each student, the key information of the exercise books is used for positioning each exercise question in the exercise books, automatically setting the size and automatic correction of the answer area of the answer sheets, the answer sheets generated by storing are used by a teacher at any time, and the correction results stored are used for subsequent checking and statistics of the score situation of each student, the answer situation of each question, and further deep analysis of the knowledge point grasp situation of the students.
3. The method for reading exercise book according to claim 2, wherein the key information of the stored exercise book includes chapters, page numbers, question numbers and answers, the chapters, the page numbers and the question numbers are used for positioning each exercise question in exercise questions, and the answers are used for automatically setting the size of answer areas of the answer sheet and automatically correcting the answer areas; the answer sheet generated by storage comprises an answer sheet and an answer sheet template, wherein the answer sheet is used for a teacher to download and use quickly at any time, and the answer sheet template is used for cutting out an answer area in the answer sheet after the student answers.
4. The exercise book reading method according to claim 1, wherein in the step S2, the generating an answer sheet includes generating a reserved identifier, generating an answer area, generating a complete answer sheet and a template, the generating the reserved identifier generates the reserved identifier according to a UUID of a student, a chapter, a page number and a question number of a problem in a database, the generating the answer area automatically sets the answer area of each problem and a color block area of a corresponding template according to the question number and answer information in the database and individualized settings of a teacher, and the answer sheet is used for the student to answer on the paper after being downloaded in the complete answer sheet and the template, and the template is used for cutting the answer area of each problem of the student.
5. The method for reading exercise book according to claim 4, wherein the generating the reserved identifier includes a student information identifier, an answer sheet correction identifier and a page identifier, the student information identifier is a two-dimensional code of an upper right corner of the answer sheet or the test paper, the content is a UUID of 32 bits, corresponding to personal information of the student, the answer sheet correction identifier is a square of an upper left corner, an upper right corner, a lower left corner and a lower right corner of the answer sheet or the test paper, and is used for identifying a direction of a roll, if an inclination exists, correction is performed, and the page identifier is a square of which the middle is hollowed or solid and is used for identifying a page number of the answer sheet.
6. The method for reading and by wholesale of exercise book according to claim 1, wherein in the step S3, the uploading of the answer sheet includes answer sheet scanning, answer sheet correction, obtaining page number information, obtaining student information and obtaining answer contents, the answer sheet scanning is performed batch scanning on the answer sheet by a teacher using a scanning device, the answer sheet correction corrects the answer sheet according to a correction identifier, the obtaining of the page number information determines a position according to the page number identifier, the obtaining of the student information binds to a corresponding student according to the student identifier, the obtaining of the answer contents obtains a corresponding template from a database, and the matching is performed, and the answer area is obtained against the template.
7. The exercise book reading method of claim 6, wherein the scanning device comprises a cell phone and a scanner, and the teacher takes a photograph of the cell phone or the scanner obtains the answer sheet image.
8. The method for reading exercise books according to claim 1, wherein in the step S4, the automatic reading includes performing OCR on an answer area, performing confusing analysis on the model used for performing OCR on the answer area, and performing automatic correction on the model based on a cyclic convolutional neural network and a connection-based time sequence classification technology, the confusing analysis is used for further compensating for accurate recognition of a student writing area in a complex scene, compensating for a situation of low OCR recognition accuracy due to redundant handwriting, writing habits and writing of different people, irregular light and dark changes of an answer card, and the like, and the automatic correction is that the answer content of the student is matched with the answer in a database for objective judgment, and the correction result is stored in the database.
9. The exercise book reading method of claim 8, wherein the cyclic convolutional neural network is a combination of a convolutional neural network and a cyclic neural network.
10. The exercise book reading method according to claim 1, wherein in S5, after the subjective questions are corrected, the results are also stored in the database and combined with scores of the corresponding objective questions for statistics.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310275971.6A CN116303871A (en) | 2023-03-21 | 2023-03-21 | Exercise book reading method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310275971.6A CN116303871A (en) | 2023-03-21 | 2023-03-21 | Exercise book reading method |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN116303871A true CN116303871A (en) | 2023-06-23 |
Family
ID=86830217
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310275971.6A Pending CN116303871A (en) | 2023-03-21 | 2023-03-21 | Exercise book reading method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116303871A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117333334A (en) * | 2023-10-13 | 2024-01-02 | 可之(宁波)人工智能科技有限公司 | Intelligent job submitting system based on two-dimension code |
| CN118379754A (en) * | 2024-04-25 | 2024-07-23 | 红心动力(北京)科技有限公司 | Exercise book detection method and system based on cloud computing and artificial intelligence |
-
2023
- 2023-03-21 CN CN202310275971.6A patent/CN116303871A/en active Pending
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117333334A (en) * | 2023-10-13 | 2024-01-02 | 可之(宁波)人工智能科技有限公司 | Intelligent job submitting system based on two-dimension code |
| CN118379754A (en) * | 2024-04-25 | 2024-07-23 | 红心动力(北京)科技有限公司 | Exercise book detection method and system based on cloud computing and artificial intelligence |
| CN118379754B (en) * | 2024-04-25 | 2025-02-14 | 红心动力(北京)科技有限公司 | A method and system for detecting exercise books based on cloud computing and artificial intelligence |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US11790641B2 (en) | Answer evaluation method, answer evaluation system, electronic device, and medium | |
| CN110705534B (en) | A method for generating wrong question book suitable for electronic vision aids | |
| CN111753767A (en) | A method, device, electronic device and storage medium for automatic job correction | |
| CN110597806A (en) | A system and method for wrong question set generation and answer statistics based on review recognition | |
| CN106781784A (en) | A kind of intelligence correction system | |
| CN109712456A (en) | System is intelligently read and made comments in a kind of student's papery operation based on camera | |
| CN109033046A (en) | Structuring visible document snap information input system and method | |
| US20150187219A1 (en) | Systems and methods for computer-assisted grading of printed tests | |
| CN102436547A (en) | Teaching wrong question statistical method and system | |
| CN107784264A (en) | Test paper analysis method and system based on image procossing | |
| CN110826872A (en) | Method, system and device for generating examination result analysis report | |
| CN116303871A (en) | Exercise book reading method | |
| CN110837793A (en) | Intelligent recognition handwriting mathematical formula reading and amending system | |
| KR20130021684A (en) | System for managing answer paper and method thereof | |
| CN107067399A (en) | A kind of paper image segmentation processing method | |
| CN114757152A (en) | Method for acquiring and printing wrong questions in teaching scene | |
| CN114708606A (en) | Intelligent automatic marking service method, system and device | |
| CN112396897A (en) | Teaching system | |
| CN115620332B (en) | Automatic reading and amending method and equipment based on paper operation | |
| KR101479444B1 (en) | Method for Grading Examination Paper with Answer | |
| CN113903039A (en) | Color-based answer area acquisition method for answer sheet | |
| CN116363657A (en) | Automatic correcting system and correcting method for paper English composition | |
| CN112164262A (en) | Intelligent paper reading tutoring system | |
| CN114663891A (en) | Daily work wrong question book generation method and system based on configured answer sheet | |
| CN117975477A (en) | Test question input method and device, electronic equipment and storage medium |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |