Method, apparatus and system for electronic processing of jobs

Abstract

The present disclosure provides a method, device, system and computer readable storage medium for electronic processing of a task, wherein the method comprises: receiving first information associated with a first trajectory written by a first user on a paper from a first intelligent terminal; receiving second information associated with a second trajectory written by a second user on the paper from a second intelligent terminal; determining subject information associated with the first trajectory based on trajectory shape recognition of the first trajectory according to the first information; determining subject information associated with the second trajectory based on trajectory position recognition of the second trajectory according to the second information; and associating the determined subject information associated with the first trajectory, the determined subject information associated with the second trajectory with a corresponding subject of an electronic template file corresponding to the paper.

Description

Technical Field

[0001] This disclosure relates to the field of electronic information technology, specifically to a method, apparatus, system, and computer-readable storage medium for electronically processing work content. Background Technology

[0002] With the development of network and computer technology, smart teaching has been implemented in the education field. Students can use smart teaching systems to write and submit assignments, while teachers can use them to grade assignments, provide tutoring, and diagnose and analyze student assignments. Combined with big data platforms, this can generate not only individual student learning reports but also overall learning reports for the entire class, grade, or even a specific region. Accurate, rapid, and convenient collection of student assignment data and corresponding teacher grading data is fundamental to achieving assignment diagnosis and subsequent learning analysis. Summary of the Invention

[0003] According to one aspect of this disclosure, a method for digitizing homework content is proposed, comprising: receiving first information transmitted by a first smart terminal as associated with a first trajectory written by a first user on paper; receiving second information transmitted by a second smart terminal as associated with a second trajectory written by a second user on the paper; performing trajectory shape recognition on the first trajectory based on the first information to determine question information associated with the first trajectory; performing trajectory position recognition on the second trajectory based on the second information to determine question information associated with the second trajectory; and associating the determined question information associated with the first trajectory and the question information associated with the second trajectory with corresponding questions in an electronic template file corresponding to the paper.

[0004] Optionally, in the above method, an invisible dot matrix pattern with positional coordinates is pre-laid on the paper, and the first smart terminal and the second smart terminal are dot matrix pens; the first trajectory is the answer handwriting of the first user using the first smart terminal to answer the question on the paper; the second trajectory is the correction handwriting of the second user using the second smart terminal to correct the first user's answer on the paper, wherein the electronic template file is an electronic original corresponding to the question stored in the database.

[0005] Optionally, in the above method, the first information includes at least one of information about the first trajectory and information about the position coordinates of the first trajectory on the paper; the second information includes at least one of information about the second trajectory and information about the position coordinates of the second trajectory on the paper.

[0006] Optionally, the above method further includes: generating a reconstructed answer handwriting on a first image based on the first information, wherein the position of the reconstructed answer handwriting on the first image is determined based on information about the position coordinates of the first trajectory on the paper; and generating a reconstructed correction handwriting on a second image based on the second information, wherein the position of the reconstructed correction handwriting on the second image is determined based on information about the position coordinates of the second trajectory on the paper, wherein the first image and the second image are the same image or different images.

[0007] Optionally, the above method further includes: performing image grayscale processing on the first image / second image to obtain a corresponding grayscale image; performing binarization processing on the grayscale image to obtain a pre-processed image; performing structured analysis on the pre-processed image to segment it into multiple regions, wherein the segmentation of the multiple regions is based on the positions of the reconstructed answer handwriting / the reconstructed correction handwriting on the first image and the second image, respectively; performing trajectory shape recognition on the reconstructed answer handwriting to determine the question information associated with the reconstructed answer handwriting or the answer information associated with the reconstructed answer handwriting; performing trajectory position recognition on the reconstructed correction handwriting to determine the question information associated with the reconstructed correction handwriting, and performing trajectory shape recognition on the reconstructed correction handwriting to determine the correction information associated with the reconstructed correction handwriting.

[0008] Optionally, in the above method, the plurality of areas includes at least one of the following: an assignment number area, a question number area, a first user identity information identification area, an answer area, an objective question grading area, and a subjective question grading area. The question information includes the question number, the answer information includes the answers written by the first user for the objective questions, and the grading information includes information on the second user grading the answers written by the first user for the subjective questions.

[0009] Optionally, the above method further includes: performing trajectory location recognition on the reconstructed correction handwriting to determine at least one of the following associated with the reconstructed correction handwriting: the job number, the identity information of the first user.

[0010] Optionally, the above method further includes: performing trajectory shape recognition on the reconstructed handwriting to determine at least one of the following associated with the reconstructed handwriting: the assignment number and the question number.

[0011] Optionally, the above method further includes: using a pre-trained trajectory shape recognition model to perform trajectory shape recognition on the reconstructed handwriting in at least one of the assignment number area, the question number area, and the answer area; and using the pre-trained trajectory shape recognition model to perform trajectory shape recognition on the reconstructed correction handwriting in the subjective question correction area; wherein the trajectory recognition model includes at least one of the following: a number trajectory shape recognition model, a letter trajectory shape recognition model, a checkmark / cross trajectory shape recognition model, and a special marker trajectory shape recognition model.

[0012] Optionally, the above method further includes: generating statistical data on the first user's answers to the questions based on analysis and statistics of at least one of the question information, answer information, and grading information.

[0013] Optionally, in the above method, the first information further includes identification information about the first smart terminal, which is associated with the first user; the second information further includes identification information about the second smart terminal, which is associated with the second user; the method further includes: associating the reconstructed handwriting with the first user based on the identification information about the first smart terminal; and associating the reconstructed correction handwriting with the second user based on the identification information about the second smart terminal.

[0014] According to another aspect of this disclosure, an apparatus for electronically processing job content is proposed, the apparatus comprising a processor and a memory coupled to the processor, wherein the processor is configured to cause the apparatus to implement the method for electronically processing job content as described above when executing program instructions stored in the memory.

[0015] According to another aspect of this disclosure, a computer-readable storage medium storing program instructions is provided, which, when executed by a processor in an electronic device, cause the electronic device to perform the method of electronically processing the work content as described above.

[0016] According to another aspect of this disclosure, a system for electronically processing work content is proposed, comprising: a first smart terminal for acquiring first information associated with a first trajectory written by a first user on paper; a second smart terminal for acquiring second information associated with a second trajectory written by a second user on the paper; and an apparatus for electronically processing work content according to the above.

[0017] According to the method, apparatus, system, and computer-readable storage medium for electronic processing of homework proposed in this disclosure, homework without a specific format can be electronically processed, broadening the application scope of homework answering and grading based on dot matrix pen technology and improving recognition accuracy. In addition, it does not change students' homework writing habits or teachers' homework grading habits, while facilitating the analysis, diagnosis, statistics, and summarization of students' homework answers and teachers' grading, thereby facilitating the diagnosis of students' learning situation and helping teachers to grasp students' current learning status in a timely manner, so as to provide more targeted teaching guidance and next lesson preparation arrangements. Attached Figure Description

[0018] Figure 1 This illustrates a scenario in which the job processing method proposed in this disclosure can be applied;

[0019] Figure 2A-2B A schematic flowchart of the job processing method proposed in this disclosure is shown;

[0020] Figures 3A-3B A schematic example of a job page to which the proposed job processing method can be applied according to embodiments of the present disclosure is shown;

[0021] Figures 4A-4B The following diagram illustrates the respective connections between an embodiment of the present disclosure and... Figures 3A-3B A schematic example of the layout of the dotted code pattern corresponding to the assignment page;

[0022] Figures 5A-5B Schematic examples of a job page and its corresponding dot-code pattern layout applicable to another embodiment of the present disclosure are shown respectively.

[0023] Figure 6 A schematic example of a job processing system according to this disclosure is shown; and

[0024] Figure 7 A schematic block diagram of a job processing apparatus according to the present disclosure is shown. Detailed Implementation

[0025] The subject will now be described with reference to the figures, where similar reference numerals throughout are used to refer to similar elements. In the following description, numerous specific details are elaborated for illustrative purposes to provide a thorough understanding of the subject. However, it will be clear that this principle can be implemented without these specific details.

[0026] This specification illustrates the principles of this disclosure. Therefore, it will be understood that, although not explicitly described or shown herein, those skilled in the art will be able to design various configurations that embody the principles of this disclosure.

[0027] This principle is naturally not limited to the embodiments described herein.

[0028] It should be noted that references to "an embodiment," "embodiment," "example embodiment," and "specific embodiment" in the specification indicate that the described embodiment may include a specific feature, structure, or characteristic, but not every embodiment necessarily includes that specific feature, structure, or characteristic. Furthermore, such phrases do not necessarily refer to the same embodiment. Moreover, when a specific feature, structure, or characteristic is described in connection with an embodiment, it can be assumed that implementing that feature, structure, or characteristic in conjunction with other embodiments (whether explicitly described or not) is within the knowledge of those skilled in the art.

[0029] The traditional method of grading homework involves students writing their assignments on paper, submitting the paper to the teacher, who then grades the work. The graded assignments are then returned to the students, and the teacher manually tallies the completion status. This traditional method is not conducive to analyzing and diagnosing the performance of individual students, nor is it helpful for statistically summarizing and compiling data on the overall performance of all students.

[0030] With the rapid advancement of network and computer technology, especially the widespread adoption of mobile internet, a new method for improving traditional homework grading has emerged. Specifically, students write their assignments on paper, then take photos of them using smartphones or other smart devices and upload them to a designated backend. Teachers then log in to this backend using tablets or other smart devices and grade the assignments online. However, this method requires students to take photos of their work and upload them, while teachers grade the assignments directly based on these images and return the graded images to each student through the backend. This method, however, requires teachers to spend extended periods grading in front of their electronic devices, and the final grading results are presented electronically, requiring students to use their own devices to view the grading. This can easily lead to eye strain for both teachers and students, which is detrimental to their health. Furthermore, this method is not conducive to analyzing and diagnosing individual student work or to statistically summarizing overall student performance.

[0031] Furthermore, according to information obtained by the inventor, some supplementary teaching materials based on dot-matrix pens have emerged. However, these materials not only require the printing of visible question stems, such as the specific content of the questions, but also the printing of visible answer boxes and / or marking boxes next to the question stems. This is to prompt students to write their answers in the answer boxes and teachers to write their corrections in the marking boxes. However, this method essentially associates the trajectory of the dot-matrix pen with specific question information based on the principle of writing on specific areas of paper with a specific format. This requires the supplementary teaching materials to have the question information in a specific format pre-printed, making it inconvenient to adjust the specific content.

[0032] To address the shortcomings of the aforementioned homework grading methods, this disclosure proposes a method for electronically processing homework. This method does not alter students' homework writing habits or teachers' grading habits. Simultaneously, it facilitates the analysis, diagnosis, statistics, and summarization of student responses and teacher grading, thereby enabling the diagnosis of students' learning progress and allowing teachers to promptly grasp students' current learning status for more targeted teaching guidance and lesson planning. Furthermore, the proposed method for electronically processing homework allows for the processing of assignments without a specific format, broadening the application scope of dot-matrix pen technology for homework answering and grading, and improving recognition accuracy.

[0033] Figure 1 The following illustrates a possible application scenario for the job processing method proposed in this disclosure. For example... Figure 1 As shown, on the student side, each student can complete their homework by writing on a paper workbook with invisible dotted patterns printed on it using a smart terminal (e.g., a smart dotted pen). The dotted patterns have corresponding positional coordinates, and the smart terminal can transmit the student's handwriting and corresponding coordinates to the processing backend in real time via the network. After completing the homework, the student submits the paper workbook to the teacher. On the teacher side, the teacher grades the students' homework using a smart terminal, and the teacher's smart terminal can transmit the teacher's grading handwriting and corresponding coordinates to the processing backend in real time via the network. The processing backend processes the handwriting of both the student and teacher to generate a reconstructed writing trajectory, and associates it with the corresponding part of the original electronic homework paper stored in the database. Simultaneously, based on the analysis and processing of the reconstructed writing trajectory, statistics and summaries of the students' homework can be compiled to provide basic data for subsequent learning assessment.

[0034] For example, smart terminals can be used to collect digital information generated based on student and teacher actions. Here, digital information refers to information that can be stored and processed using a computer and transmitted via a computer network. As an example, the processing backend may include processing devices and resources such as servers and databases. The server can be a single server, multiple servers networked together, or a cloud server. The server processes the digital information collected by the smart terminal to form assignment diagnostic data, which may include answer-related data, grading-related data, etc.

[0035] As an example, the processing backend can also connect to the teaching service processing system to utilize its rich stored teaching resources, such as question banks and knowledge bases, and obtain relevant teaching and tutoring information based on the analysis of homework diagnostic data, such as providing supplementary teaching and tutoring for students' weak knowledge points.

[0036] As an example, in embodiments of this disclosure, the smart terminal may include a dot-matrix pen. It is understood that the above application scenarios are merely exemplary, and based on the technical principles of the embodiments of this disclosure, the technical concepts of this disclosure can also be implemented in other application scenarios.

[0037] Figure 2A A schematic flowchart of a method for digitizing work content according to the present disclosure is shown. It includes: S205, receiving first information transmitted from a first smart terminal associated with a first trajectory written by a first user on paper; S210, receiving second information transmitted from a second smart terminal associated with a second trajectory written by a second user on the paper; S215, performing trajectory shape recognition on the first trajectory based on the first information to determine question information associated with the first trajectory; S220, performing trajectory position recognition on the second trajectory based on the second information to determine question information associated with the second trajectory; and S225, associating the determined question information associated with the first trajectory and the question information associated with the second trajectory with corresponding questions in an electronic template file corresponding to the paper.

[0038] Specifically, the embodiments of this disclosure are based on the principle of dot matrix codes. As an example, similar to ordinary exercise books / sheets, no visible homework content is printed on the paper homework pages used by students. For example, there is no visible information such as homework number, question number, or question stem information on the homework page; that is, it appears to be a blank homework sheet. Unlike ordinary exercise books / sheets, the homework paper according to embodiments of this disclosure has a layer of invisible dot matrix pattern with positional coordinates pre-printed or printed on it. The dot matrix pattern includes many dot matrix codes arranged according to a specific pattern, i.e., coded dots arranged on the paper page. The shape of the dots is not limited; for example, they can be square or circular, with very small side lengths or diameters, so as not to affect the user's normal writing on the paper. The user writes normally on the paper using a dot matrix pen. The dot matrix pen captures the dot matrix codes on the page through a camera to obtain relevant information such as the writing trajectory of the pen tip on the page. Accordingly, students are equipped with dot matrix pens. When writing answers on the paper dot matrix homework paper with the dot matrix pen, the corresponding answer data is synchronously transmitted to the processing backend via the network. Since dot-matrix pens are not significantly different from ordinary pens (such as ballpoint pens and pencils) when writing, they leave visible writing traces on the dot-matrix homework paper. Students then submit their completed paper assignments to the teacher, who uses their own dot-matrix pen to directly grade the assignments. Similarly, while the teacher is grading, the corresponding grading data is synchronously transmitted to the processing backend via the network, leaving visible grading marks on the student's dot-matrix homework paper. Furthermore, in the processing backend, based on the processing of the answer data transmitted from the student's dot-matrix pen and the grading data transmitted from the teacher's dot-matrix pen, reconstructed student writing trajectories and reconstructed teacher writing trajectories are generated for both. These reconstructed student and teacher writing trajectories are then identified, allowing them to be associated with the corresponding questions in the electronic templates of the assignments stored in the database. For example, the student's answer data for question 3 of assignment number B003, the teacher's grading data for the student on that question, and the corresponding electronic assignment number 3 can be linked together to facilitate the analysis, diagnosis, statistics, and summarization of student work performance in later stages.

[0039] The above-mentioned methods of answering and grading assignments do not change students' homework-writing habits or teachers' grading habits. At the same time, they facilitate the analysis, diagnosis, statistics, and summarization of student responses and teacher grading, thereby enabling the diagnosis of students' learning progress and allowing teachers to promptly grasp students' current learning status for more targeted teaching guidance and lesson planning. Furthermore, the assignment paper used in these methods appears no different from ordinary assignment paper; there is no need to print the assignment's question information, such as question numbers or stems. It is a type of assignment paper that does not require design or layout based on the question information itself—in other words, it is a non-format assignment paper.

[0040] According to embodiments of this disclosure, teachers can distribute paper dot matrix exercise books to students, which may include multiple assignments numbered sequentially. As an example, Figure 3A The diagram shows a visible student identification area printed on the title page of the workbook, along with a corresponding assignment correction number identification area. The student identification area indicates which student the workbook belongs to, while the assignment correction number identification area is used by the teacher to check a box when grading the student's assignment, allowing the teacher's handwriting to be linked to the student's answer data for that assignment. Accordingly, as an example, Figure 3B The diagram shows the layout of a single assignment page within the workbook. The header of this page features a visible assignment number input area for students to enter when answering the assignment. As an example, the assignment number can consist of a combination of letters and numbers. For instance, letters correspond to subject parameters (e.g., A = Chinese, B = Math, C = English, etc.), and numbers define a unique assignment number (e.g., 003 represents the 3rd assignment of the current academic year). This provides a unique identifier for the subject and number of the assignment; for example, B003 indicates that this is the 3rd math assignment of the current academic year.

[0041] Figures 4A-4B The following diagram illustrates the respective connections between an embodiment of the present disclosure and... Figures 3A-3B A schematic example of the layout of the dotted code pattern corresponding to the applicable job page. For example... Figure 4A As shown, it includes and Figure 3A The visible area corresponds to the dot matrix pattern layout area, such as the student identification information recognition area, the grading confirmation area, etc.; while such as Figure 4B As shown, it includes and Figure 3B The visible area corresponds to the dot matrix pattern layout area, such as the basic information recognition area for assignments, the question number recognition area, and the handwriting recognition area for answers and corrections.

[0042] According to embodiments of this disclosure, optionally, each assignment may not have the aforementioned title page, but instead... Figure 5A As shown, each assignment has a student identification area and an assignment number identification area in the header. The student identification area is for students to affix their identification label when answering each assignment, while the assignment number identification area is for students to write the assignment number. For example, the identification label could be a self-adhesive name tag coded with the student's name, class, student ID, etc., which students can affix to the student identification area when answering each assignment. The assignment number can be a combination of letters and numbers; for example, letters correspond to subject parameters (e.g., A = Chinese, B = Math, C = English, etc.), and numbers define a unique assignment number (e.g., 003 represents the 3rd assignment of the current academic year), thus giving each assignment a unique identifier for its subject and number. For example, B003 indicates that this is the 3rd math assignment of the current academic year. Alternatively, the student identification area could be used for students to write their name, class, student ID, etc., instead of affixing identification labels. Alternatively, a student identification area can be printed on the cover of the dot matrix exercise book, where students can write their name, class, student ID, etc., or affix their identification labels, without having to set up a student identification area on the first page of each assignment.

[0043] Optionally, each student is equipped with a smart dot matrix pen with specific identification information, such as MAC information. When the smart dot matrix pen is distributed to the student, the identification information of the dot matrix pen can be associated with student information such as class, name, and student ID in the processing backend. This allows the writing trajectory transmitted by the dot matrix pen to the processing backend to be identified as the student's writing trajectory when the dot matrix pen is turned on and written on.

[0044] As mentioned above, the pages of the dot matrix exercise book also have an invisible dot matrix pattern with positional coordinates pre-printed / printed. The dot matrix pattern includes many dot matrix codes arranged according to a specific pattern, that is, code dots arranged on the paper page, each code dot having a corresponding positional coordinate on the page. When students use a dot matrix pen to write on the dot matrix exercise book, on the one hand, since the writing experience of the dot matrix pen is not much different from that of ordinary pens such as ballpoint pens, pencils, and fountain pens, it will also leave a writing trace on the page. On the other hand, the dot matrix pen can be used to collect the writing trace information of students on the page and the corresponding code dot positional coordinate information.

[0045] As an example, the dot matrix pen is equipped with a high-speed camera / scanner. When students use the dot matrix pen to write on the pages of the dot matrix exercise book, the high-speed camera / scanner can be activated to continuously scan and record the coordinates of the dot position of the pen tip while writing, thereby collecting the trajectory information of the student's writing on the page and the corresponding dot position coordinate information.

[0046] As another example, the dot matrix pen is also equipped with a pressure sensor. When a student writes on the page of the dot matrix exercise book with the pen, the pressure sensor will detect the pressure of the pen tip, which can then activate a high-speed camera / high-speed scanner to continuously scan and record the coordinates of the dot position of the pen tip while writing, thereby collecting the trajectory information of the student's writing on the page and the corresponding dot position coordinate information.

[0047] Similar to the page layout of ordinary workbooks / worksheets, each worksheet according to embodiments of this disclosure does not have visible work content on its page. For example, it does not have visible question numbers or question stems, and only a worksheet number identification area is printed / printed in the header of the page. The difference from ordinary workbooks / worksheets is that the dot matrix workbook according to embodiments of this disclosure has invisible dot matrix patterns with positional coordinates pre-printed / printed on its page. As an example, the dot matrix pattern on the page can be divided into multiple areas, each of which can be associated with different work content attributes. For example, the dot matrix pattern on page 1 can be divided into area 1, area 2, area 3, and area 4, where area 1 corresponds to the worksheet number identification area, area 2 corresponds to the question number identification area, area 3 corresponds to the student identity information identification area, and area 4 corresponds to the handwriting identification area for answering and grading. Figure 4B or Figure 5B As shown.

[0048] As an example, when students are writing their assignments, they can first affix their identification tags to the student identification information recognition area. Optionally, teachers can pre-distribute printed student identification tags to students for them to affix while writing their assignments. Each student's identification tag not only has visible information such as the student's class, name, and student ID printed on it, but also contains encoded information reflecting the student's class, name, and student ID that can be read by a dot-matrix pen. This allows teachers to associate their corrections with the student's handwriting by marking (e.g., checking) on ​​the student's identification tag with the student's answer handwriting. Figure 3A Or as shown in 5A.

[0049] Optionally, when completing assignments, students can use a dot-matrix pen to write the assignment number in the assignment number recognition area, such as C005, indicating that this is the 5th English assignment of the current academic year. While writing with the dot-matrix pen, the pen simultaneously collects the coordinates of the invisible dot matrix pattern corresponding to the handwriting, for example, by photographing / scanning the handwriting with a high-speed camera / scanner, and transmits the collected trajectory information and corresponding dot coordinate information to the processing backend. Accordingly, the processing backend processes the handwriting and its coordinates to obtain the assignment number, thereby associating the student's completed assignment with the electronic assignments stored in the database corresponding to that assignment number. Figure 3B Or as shown in 5A.

[0050] When answering questions, students write their answers directly on the dot-matrix paper. For example, Figure 3B As shown in 5A, for objective questions, students can directly write the question number and corresponding answer in the answer area. For example, 1.A indicates that the answer to question 1, which is a multiple-choice question, is A; or (3)B indicates that the answer to question 3, which is a multiple-choice question, is B; or (10)16 indicates that the answer to question 10, which is a fill-in-the-blank question, is 16. Alternatively, circled numbers can be used to represent the question number, such as ⑧, which indicates question 8. In other words, there are no restrictions on how the question number is represented. For subjective questions, for example, Figure 3B or Figure 5A As shown, students can also write the question number in the question number area and the corresponding answer in the answer area.

[0051] According to embodiments of this disclosure, for objective questions, students can use a dot-matrix pen to write in the answer area, leaving their handwriting. Simultaneously, the dot-matrix pen collects the coordinates of the code points of an invisible dot-matrix pattern corresponding to the handwriting and transmits the handwriting and its coordinates to the processing backend. The processing backend then processes the handwriting and coordinates to obtain the question number and the student's answer. By comparing this answer with the correct answers stored in the database, it determines whether the student's answer is correct. This achieves automatic grading of objective questions without requiring teachers to grade them, reducing their workload and facilitating statistical analysis of the answers.

[0052] According to embodiments of this disclosure, for subjective questions, students can use a dot-matrix pen to write in the answer area, leaving their handwriting. As an example, teachers use the dot-matrix pen to correct students' answers to subjective questions and leave their corrections on the assignment paper.

[0053] According to embodiments of this disclosure, as described above, it is not necessary to separately provide visible markings on the worksheet to distinguish between the objective question answering area and the subjective question answering area. Instead, it is possible to... Figure 3B or Figure 5A As shown, a separate blank area is set up on the page as the input area for the question numbers of subjective questions. When a student enters the question number in this area and writes the answer in the corresponding answer area, the handwriting corresponding to that question number is recognized by the processing system as an answer to a subjective question. Therefore, the answer to that subjective question is no longer automatically graded, but left for the teacher to grade later. In practice, considering that subjective questions often do not have a single standard answer, and generally require not only a correct result but also a detailed explanation of the steps, the teacher's experience is needed to judge the student's answer.

[0054] Accordingly, when students answer objective questions, they do not need to enter the question number in the question number input area mentioned above. Instead, they can directly enter the question number and their answer in the answer area so that the processing backend can automatically identify and grade them.

[0055] Alternatively, instead of displaying a visible question number input area for subjective questions on the page, other methods can be used to differentiate between the subjective and objective question answer areas. For example, the backend database can determine whether a student's answer pertains to an objective or subjective question based on the question number. For instance, in a given assignment, questions numbered 1-20 might be objective questions, while questions numbered higher than 20 might be subjective questions. Thus, after identifying the question number, the backend will automatically grade questions with numbers between 1 and 20, while questions with numbers higher than 20 will be left for manual grading by the teacher.

[0056] As an example, when a student uses a dot matrix pen to answer questions in the answer area, the dot matrix pen collects the student's writing trajectory information on the page and the corresponding code point position coordinate information in real time, and transmits it to the processing backend.

[0057] As an example, you can refer to Figure 3B or Figure 5A To understand the principles of this disclosure. For example... Figure 3B or Figure 5AAs shown, the assignment page is mostly blank, with no visible assignment content printed on it, such as question stems or question numbers. There are no longer separate answer areas or grading areas for each question. Instead, students write their answers directly in the answer areas, and teachers write their corrections near the students' answers. In other words, from the student's perspective, they can maintain the same writing habits as when writing regular assignments. The processing backend intelligently recognizes and analyzes the received dot-matrix pen writing trajectory to identify the corresponding question, determine whether it's an objective or subjective question, and automatically grade the student's answers for objective questions. The grading results are then linked to the corresponding area (i.e., the corresponding question) in the original electronic assignment in the database for later diagnostic analysis. This method eliminates the need to print prompts on the assignment page for students to input or check boxes to associate their writing trajectory with each assignment question; instead, it directly determines the specific question associated with the identified writing trajectory.

[0058] Therefore, according to the embodiments of this disclosure, students can maintain their current writing habits. That is, the process of completing assignments using a dot-matrix pen on paper dot-matrix assignment paper is no different from the process of completing assignments using a regular pen on regular paper assignment paper. Students do not need to perform any additional operations. However, since the paper dot-matrix assignment paper has invisible dot-matrix patterns pre-printed on it, the trajectory of the dot-matrix pen writing in a certain area can be identified by the processing backend, thereby determining the actual information corresponding to the writing trajectory, such as which assignment it corresponds to and which question in that assignment. This allows the writing trajectory to be associated with the original electronic assignment paper stored in the database, facilitating subsequent analysis, diagnosis, statistics, and summarization operations.

[0059] Optionally, while the student is writing with the dot matrix pen, the pen simultaneously captures the student's writing trajectory through methods such as taking a picture or scanning, and transmits information about the writing trajectory and the coordinates of the code points to the processing backend via a message queue. As an example, and not a limitation, the above information can be transmitted to the processing backend via an MQTT message queue.

[0060] Optionally, at least one area among the multiple areas included in the dot matrix pattern on the page can correspond to the area on the page used for inputting information such as the student's name, class, and student ID. When the student writes their name, class, and student ID in this area, the dot matrix pen simultaneously captures the corresponding writing trajectory and transmits the writing trajectory information and the corresponding code point position coordinates to the processing backend. The processing backend will identify this information to associate it with the corresponding student identity in the database. Unlike the above method of identifying the corresponding student identity through the dot matrix pen's identification information, this method does not require the dot matrix pen's identification information, such as MAC information, to be pre-associated with the student identity in the processing backend's database. Instead, it dynamically associates the dot matrix pen's writing trajectory with the student's identity when the student writes their own identity information.

[0061] After students complete their assignments, they submit their paper-based dot matrix workbooks to the teacher. The teacher can then grade them. Specifically, teachers can be equipped with their own dot matrix pens and use them to grade students' answers on the submitted workbooks. As an example, the smart dot matrix pens provided to teachers have specific identification information, such as MAC addresses. When distributing smart dot matrix pens to teachers, the processing backend can associate the pen's identification information with teacher information such as grade level, subject taught, and name. This allows the writing trajectory transmitted by the pen to the processing backend when the pen is turned on to be recognized as the teacher's writing trajectory.

[0062] As an example, when grading student assignments, teachers can first confirm the student's identity by displaying an area on the assignment page that shows the student's class, name, and student ID on the title page. This area can be used for student identification, for example, by using a dot-matrix pen. Figure 3A Check the box corresponding to the student's identity information in the grading area for this assignment so that the teacher's grading handwriting can be linked to the student's answers to this assignment.

[0063] Alternatively, if the exercise book does not have the aforementioned title page, such as Figure 5A As shown, a student identification area and an assignment number identification area can be set in the header of each assignment. When grading, the teacher marks the student identification area and / or the assignment number identification area to associate the teacher's grading handwriting with the student's answer to the assignment. This makes it easier for the backend to identify which student's assignment the grading trajectory received from the teacher's dot matrix pen corresponds to.

[0064] Then, the teacher can begin grading the students' assignments as usual. For example, the teacher can write their corrections near the student's handwriting on each question. Similar to traditional paper-based assignment grading, students typically use blue or black ink with a dot-matrix pen, while the teacher uses red ink. As mentioned above, the teacher primarily grades the students' answers to subjective questions. Therefore, the teacher's corrections are placed near the student's handwriting. The teacher can grade the student's answers accordingly. For example, if the student's answer is correct, the teacher can mark it with a checkmark; if the answer is incorrect, the teacher can mark it with an X; or if no answer was given, the teacher can mark it with a question mark. Furthermore, the teacher can write the corresponding score, such as "-2" indicating a deduction of 2 points, etc., so that the backend can analyze and statistically process the teacher's grading.

[0065] As mentioned above, the pages of the dot matrix exercise book also have an invisible dot matrix pattern with positional coordinates pre-printed / printed. The dot matrix pattern includes many dot matrix codes arranged according to a specific pattern, that is, code dots arranged on the paper page, each code dot having a corresponding positional coordinate on the page. When teachers use a dot matrix pen to correct on the dot matrix exercise book, on the one hand, since the writing experience of the dot matrix pen is not much different from that of ordinary pens such as ballpoint pens, pencils, and felt-tip pens, it will also leave correction traces on the page. On the other hand, the dot matrix pen can be used to collect the trajectory information of the teacher's correction on the page and the corresponding code dot positional coordinate information.

[0066] As mentioned above, the dot matrix pen can be equipped with a high-speed camera / high-speed scanner. When the teacher uses the dot matrix pen to correct on the dot matrix exercise book page, the high-speed camera / high-speed scanner can be activated to continuously scan and record the coordinates of the dot code position of the pen tip during correction, thereby collecting the trajectory information of the teacher's correction on the page and the corresponding code dot position coordinate information.

[0067] As another example, the dot matrix pen is also equipped with a pressure sensor. When the teacher uses the dot matrix pen to correct on the page of the dot matrix exercise book, the pressure sensor will detect the pressure of the pen tip, which can then activate a high-speed camera / high-speed scanner to continuously scan and record the coordinates of the dot position of the pen tip during correction, thereby collecting the teacher's correction trajectory information on the page and the corresponding dot position coordinate information.

[0068] According to embodiments of this disclosure, after a student completes their assignment, their handwriting is left in the corresponding answer area. Therefore, when a teacher grades the submitted assignment, they can see the student's handwriting in the corresponding answer area. Similar to traditional assignment grading methods, according to embodiments of this disclosure, the teacher uses a dot-matrix pen to grade near the student's handwriting. Since the teacher's grading handwriting is located near the student's handwriting—that is, within the answer area of ​​the assignment question being graded—the dot-matrix pen simultaneously captures the grading handwriting and associates it with the position coordinates of that answer area. This facilitates the backend processing in identifying which specific question the teacher's grading handwriting is associated with, and thus also with the student's answer to that question. For example, when a teacher is grading near a student's handwriting, such as when grading a student's answer to question 1, the teacher's grading trajectory lies within the area of ​​that answer. The position coordinates associated with the dot matrix pattern area corresponding to that answer area can indicate that the grading trajectory of the teacher using a dot matrix pen in that answer area belongs to the grading information for question 1. In fact, as mentioned above, when a teacher uses a dot matrix pen to grade assignments on the page, the dot matrix pen collects the teacher's grading trajectory information on the page and the corresponding code point position coordinate information, and transmits it to the processing backend.

[0069] In other words, from the teacher's perspective, they can maintain the same grading habits as when grading regular paper assignments. The backend processing system, by analyzing the positional coordinates associated with the teacher's dot-matrix pen markings, can identify the corresponding answer question. This allows it to be linked to the corresponding area (i.e., the corresponding question) of the original electronic assignment in the database. Furthermore, it can associate the marking with the student's information for later diagnostic analysis. This method eliminates the need to print prompts on the assignment page for teachers to input or check boxes to associate their marking with each question. Instead, it directly identifies the specific question associated with the marking trajectory based on the relationship between the positional coordinates of the marking trajectory and the corresponding answer handwriting.

[0070] As an example, the teacher's corrections are usually located near the student's answer to a particular question. That is, the teacher's corrections are located after the question number written by the student and before the next question number. Therefore, by analyzing the relationship between the position coordinates of the teacher's corrections and the position coordinates of the question number written by the student, it is possible to identify which specific question the teacher's corrections are for.

[0071] Therefore, the method for digitizing assignments proposed in this disclosure can be used to digitize assignments without a specific format, broadening the application scope of assignment answering and grading based on dot matrix pen technology and improving recognition accuracy. It performs image processing on the answer and grade handwriting acquired using the dot matrix pen, structuring the image region based on dot coordinates, dividing it into sub-regions corresponding to different data types (e.g., numbers, letters, operators, grading marks, etc.), and employs different AI models to recognize each sub-region. This enables accurate identification of answer data and corresponding grade data, associating them with the corresponding question numbers, ultimately completing the digitization of the assignment.

[0072] Therefore, according to the embodiments of this disclosure, not only can students maintain their current homework writing habits, but teachers can also maintain their current homework grading habits. That is, the process of students completing homework on paper dot matrix homework sheets using a dot matrix pen is no different from the process of completing homework on ordinary paper homework sheets using an ordinary pen. Students do not need to perform any additional operations. However, since the paper dot matrix homework sheets have invisible dot matrix patterns with position coordinates pre-printed on them, the processing backend can identify which specific question the student's handwriting is for by analyzing the trajectory information and corresponding position coordinates transmitted by the dot matrix pen. For objective questions, the backend can identify the student's answer for automatic grading. At the same time, teachers only need to grade the student's handwriting without performing any additional operations. Since the teacher's grading handwriting is located near the corresponding student's handwriting, that is, the position coordinates of the teacher's grading handwriting are related to the position coordinates of the student's handwriting for a certain question, the backend can identify which question the teacher's grading handwriting is for, thus facilitating the processing backend to identify the actual information corresponding to the teacher's grading trajectory. Furthermore, by having teachers verify the student's identity in the student's identification area on the workbook, the teacher's correction handwriting can be associated with a specific student's workbook. This links the correction trajectory to the original electronic workbook stored in the database, thus facilitating subsequent analysis, diagnosis, statistics, and summarization.

[0073] Optionally, while the teacher is writing with the dot matrix pen, the pen simultaneously captures the teacher's marking trajectory through methods such as taking a picture or scanning, and transmits information about the marking trajectory and the coordinates of the code points to the processing backend via a message queue. As an example, and not a limitation, the above information can be transmitted to the processing backend via an MQTT message queue.

[0074] According to embodiments of this disclosure, the answer trajectory written by the student using a dot-matrix pen and the corresponding code point coordinate information are transmitted to the processing backend. The processing backend can identify which student wrote the received answer trajectory using the identification information of the dot-matrix pen bound to each student's identity, such as the MAC information of the dot-matrix pen. As mentioned above, as an example, students can also be identified based on the trajectory information of student information written by the student in a specific area of ​​the worksheet, such as the student information input area, including the class, name, and student ID.

[0075] Similarly, the answer trajectory written by the teacher with the dot matrix pen and the corresponding code point coordinate information are transmitted to the processing backend. The processing backend can identify which teacher wrote the received correction trajectory by using the identification information of the dot matrix pen that is bound to the teacher's identity, such as the MAC information of the dot matrix pen.

[0076] According to an embodiment, the writing information received by the processing backend includes at least one of writing trajectory information and code point coordinate information. As described above, the processing backend can identify the corresponding student and / or teacher based on the identification information of the received dot matrix pen, and then associate subsequently received writing information with the identified student / teacher in the database.

[0077] As an example, the processing backend can generate a reconstructed writing trajectory based on the processing of the received writing information. Specifically, the processing backend converts the received writing trajectory information into an image and performs image processing to generate the reconstructed writing trajectory.

[0078] Optionally, the image can first be processed into grayscale and then binarized for subsequent image segmentation. Image binarization facilitates further image processing, simplifies the image, reduces data volume, and highlights the outline of the target of interest. Therefore, to process and analyze binary images, the grayscale image must first be binarized to obtain a binarized image. Binarization can convert a grayscale image into a binary image. For example, pixels with grayscale values ​​greater than a certain threshold, such as a critical grayscale value, are set as grayscale maxima, and pixels with grayscale values ​​less than this value are set as grayscale minima, thus achieving binarization. For example, the pixels of an image can be binarized to 0 or 255. This is beneficial because when further processing the image, the set properties of the image only depend on the position of the pixels with a value of 0 or 255, no longer involving multi-level pixel values, making the processing simpler and reducing the amount of data processing and compression. Depending on the threshold selected, the binarization algorithm can be based on a fixed threshold or an adaptive threshold.

[0079] According to the embodiments, after binarizing an image, image segmentation can be performed. Image segmentation refers to the process of subdividing a digital image into multiple sub-regions (sets of pixels). The purpose of image segmentation is to simplify or change the representation of an image, making it easier to understand and analyze. Image segmentation is commonly used to locate objects and boundaries (lines, curves, etc.) in an image. More precisely, image segmentation is a process of labeling each pixel in an image, a process that makes pixels with the same label share some common visual characteristics. The result of image segmentation is a set of sub-regions on the image (these sub-regions collectively cover the entire image), or a set of contour lines extracted from the image (e.g., edge detection). Each pixel in a sub-region is similar under some metric or calculated characteristic, such as color, brightness, and texture. Adjacent regions are very different under some metric.

[0080] Here, taking a single page of an assignment as an example, we will explain in detail the process of generating a reconstructed writing trajectory based on the received writing trajectory and corresponding position coordinate information according to the embodiments of this disclosure. As an example, the processing backend can generate a digital image including the reconstructed handwriting based on the received student's handwriting and corresponding position coordinate information. The position of the reconstructed writing trajectory on the digital image is determined based on the corresponding position coordinate information, thereby generating a digital image including the student's reconstructed handwriting. The position of the reconstructed handwriting on the digital image corresponds to the position of the handwriting on the paper assignment sheet.

[0081] Similarly, the processing backend can generate a digital image including the reconstructed correction handwriting based on the received teacher's correction handwriting and corresponding location coordinate information. The position of the reconstructed correction trajectory on the digital image is determined based on the corresponding location coordinate information. Thus, a digital image including the teacher's reconstructed correction handwriting is generated, wherein the position of the reconstructed correction handwriting on the digital image corresponds to the position of the correction handwriting on the paper homework sheet.

[0082] Then, as described above, the aforementioned digital image, including the reconstructed student handwriting, can be processed in grayscale and then binarized to obtain a pre-processed image. Next, the pre-processed image undergoes structuring processing. Specifically, image segmentation can be performed on the image. As an example, OpenCV projection can be used to segment the image, dividing it into sub-regions containing different content, such as assignment number sub-regions, question number sub-regions, and handwriting sub-regions.

[0083] As an example, a pre-set trajectory recognition model can be used to perform AI trajectory recognition on the trajectory content contained in each of the plurality of regions; wherein, the trajectory recognition model may include at least one of the following: a number trajectory recognition model, an letter trajectory recognition model, a character trajectory recognition model, and a special marker trajectory recognition model.

[0084] For example, handwriting trajectory recognition can be performed on the handwriting within the assignment number sub-region. Considering that assignment numbers are generally a combination of letters and numbers—for example, letters correspond to their subject parameters (e.g., A = Chinese, B = Math, C = English, etc.) and numbers define unique assignment numbers, such as 001, 002, etc.—the assignment is assigned a subject and a unique number. Therefore, the handwriting within the assignment number sub-region can be fed into a number trajectory recognition model and a letter trajectory recognition model, respectively, for comparison with pre-trained number and letter trajectories, thus enabling handwriting recognition within the assignment number sub-region.

[0085] Similarly, handwriting in the "Question Number" sub-region can be tracked. Considering that question numbers are usually a combination of numbers and special characters (e.g., dots, commas, parentheses, hyphens, etc.), a unique number is assigned to each question in this assignment. For example, 2. can represent question 2, while 3-1-(2) can represent the second sub-question under the first sub-question of question 3, and so on. Therefore, the handwriting in the "Question Number" sub-region can be put into the digital trajectory recognition model and the special marker trajectory recognition model, so as to compare it with the pre-trained digital trajectory and special marker trajectory respectively, thereby recognizing the handwriting in the "Question Number" sub-region.

[0086] Furthermore, trajectory recognition can be performed on the handwriting within the "response handwriting" sub-region to obtain specific response information. As an example, a pre-trained character recognition model can be used for recognition. For instance, the CRNN (Convolutional Recurrent Neural Network) + CTC (Connectionist Temporal Classification) algorithm can be considered for handwriting recognition. Optionally, the positions of each response trajectory in the response handwriting sub-region image can be located first, and then recognition can be performed. Thus, the positions of each response handwriting contained in the response handwriting sub-region image can first be marked with boxes, and then the CRNN + CTC algorithm can be used for handwriting recognition. Specifically, CNN can be used to extract image convolutional features, then a Long Short-Term Memory (LSTM) network can be used to further extract sequence features from the image convolutional features, and finally CTC is introduced to solve the problem of character misalignment during training, thereby generating reconstructed response handwriting.

[0087] Similarly, the aforementioned digital image, including the reconstructed teacher's handwriting corrections, can be processed in grayscale and then binarized to obtain a pre-processed image. Then, the pre-processed image undergoes structuring processing. Specifically, image segmentation can be performed. As an example, OpenCV projection can be used to segment the image, dividing it into sub-regions containing different content, such as a student information recognition region and a correction handwriting sub-region.

[0088] For example, when a teacher marks the student information recognition area with a dot matrix pen, the processing backend can recognize the teacher's handwriting after the image is preprocessed and reconstructed. It can also recognize the teacher's handwriting for the sub-region image of the "student information recognition area".

[0089] Similarly, trajectory recognition can be performed on the handwriting in the "correction handwriting" sub-region to obtain specific correction information. For example, for simple correction handwriting, such as checkmarks for correct answers and crosses for incorrect answers, a pre-trained checkmark / cross recognition model can be used for identification, such as comparing it with pre-trained checkmark / cross handwriting samples for rapid recognition. Alternatively, for handwriting indicating deductions, such as "-2" indicating a deduction of 2 points, a pre-trained digit recognition model and special marker recognition model can be used for identification. For general teacher comments, a pre-trained character recognition model can be used. For example, the CRNN (Convolutional Recurrent Neural Network) + CTC (Connectionist Temporal Classification) algorithm can be considered for handwriting recognition. Optionally, the position of each correction trajectory in the "correction handwriting" sub-region image can be located first, and then recognition can be performed. Therefore, we can first use boxes to mark the positions of each correction handwriting contained in the "correction handwriting" sub-region image, and then use the CRNN+CTC algorithm to perform handwriting recognition. Specifically, we can use CNN to extract image convolution features, then use LSTM to further extract sequence features from the image convolution features, and finally introduce CTC to solve the problem of character misalignment during training, thereby generating reconstructed correction handwriting.

[0090] Optionally, the identified marking information reflecting the teacher's marking can be associated with specific questions in the original electronic version of the assignment in the database to generate marking statistics for the assignment, such as the score for each question, the total score for the assignment, and which question was answered incorrectly.

[0091] Optionally, the recognition results of the student's "answer handwriting" sub-region image can be associated and superimposed with the recognition results of the teacher's "correction handwriting" sub-region image. That is, the reconstructed answer handwriting and the reconstructed correction handwriting are associated with specific questions in the original electronic version of the assignment in the database, thereby generating the corrected electronic assignment.

[0092] As described above, the teacher marks the student information identification area, for example, in... Figure 3A or Figure 5A Marking the student assignment label area as shown allows for the association of the teacher's handwriting with the specific assignment completed by the student. Alternatively, the association can be achieved by the teacher writing the corresponding student information (e.g., class-student ID) in a specific area of ​​the submitted paper assignment (e.g., the student information confirmation area). This allows the teacher's handwriting to be linked to the assigned student's work at the processing backend. In fact, both the student assignment label area and the student information confirmation area can be associated with a dot matrix pattern coded with the relevant information, enabling the processing backend to recognize the markings written in these areas as being associated with a specific student's specific assignment.

[0093] Furthermore, based on the location coordinates of the student's handwriting and the teacher's corrections, these coordinates can be associated with specific questions. For example, the association between a student's handwriting and the teacher's corrections can be established by recognizing that the teacher's corrections are located near the student's handwriting. In other words, the teacher's corrections are linked to the location coordinates of the student's answer. For instance, if a student's handwriting is located in the area after the question number and before another question number, the teacher's corrections will also fall within this area. Therefore, the teacher's corrections can be identified as corrections made to the student's answer for that specific question.

[0094] According to another embodiment of this disclosure, such as Figure 2BAs shown, a method for digitizing homework content is disclosed. It includes: S2105, acquiring answer handwriting and correction handwriting, and placing them into a message queue; S2110, converting the answer handwriting and correction handwriting into images based on the message queue; S2115, performing image processing on the images and structuring them, dividing them into sub-regions corresponding to identity information, question number information, answer information, and correction information, respectively; S2120, performing trajectory recognition on the handwriting in each sub-region based on an AI model; S2125, associating the identified identity information, question information, answer information, and correction information with the corresponding regions of an electronic template file.

[0095] Optionally, the system uses a student's dot-matrix pen to acquire the student's handwriting and corresponding positional coordinates on a paper dot-matrix worksheet. This handwriting and coordinates are then transmitted to a smart pen box via a data cable, and subsequently to a processing backend (e.g., a server) via WiFi. Similarly, the system uses a teacher's dot-matrix pen to acquire the teacher's corrections and corresponding positional coordinates on the paper dot-matrix worksheet. This corrections and coordinates are then transmitted to a smart terminal (e.g., a mobile phone, router, gateway) via Bluetooth, and subsequently to the processing backend (e.g., a server). The handwriting / corrections and their coordinates can be placed in a message queue, such as an MQTT message queue, and the handwriting / corrections can be converted into an image based on this message queue.

[0096] Optionally, as described above, the image can be processed, for example, by grayscale processing and then binarization, to obtain a preprocessed image.

[0097] Optionally, as described above, the preprocessed image can be structured, for example, by using OpenCV projection to segment the image into multiple regions. These regions can be segmented based on position coordinate information into sub-images that include at least one of the following regions: an assignment number region containing the student's handwriting of the assignment number; a question number region containing the student's handwriting of the subjective question number; an answer region containing the student's handwriting of their answers to the assignment questions; and a grading region containing the teacher's handwriting of their grading of the student's answers.

[0098] Optionally, handwriting recognition in the aforementioned sub-images can be performed using AI recognition models, such as number recognition models, letter recognition models, and checkmark / cross recognition models. For example, recognizing the homework number sub-image can yield a combination of letters and numbers, where letters correspond to their subject parameters (e.g., A = Chinese, B = Math, C = English, etc.), and numbers define a unique homework number, thus assigning a unique identifier to the subject and number of the homework assignment. Similarly, recognizing the question number sub-image yields the specific question number; recognizing the answer area sub-image identifies the student's handwriting for each question, thus obtaining information on the student's answers; and recognizing the grading area sub-image, for example, identifying checkmarks, crosses, deductions, and other grading handwriting, provides information on the teacher's grading of each question.

[0099] The above example illustrates the process of processing a single page of an assignment to obtain the set of questions, student answers, and teacher corrections for that single page. The process of processing other pages of the same assignment is similar and will not be repeated here.

[0100] After completing the analysis and processing of a single assignment page, we can obtain all the answer and grading information for that page. By performing similar processing on other pages of the same assignment, we can obtain the complete answer and grading information for that assignment. This allows us to associate it with the various questions of the corresponding assignment in the database, facilitating subsequent statistics, analysis, and summarization.

[0101] After processing all pages of the assignment, the data can be integrated. For example, based on the teacher's grading data, the student's answer data can be analyzed and statistically summarized to form a summary of the student's assignment completion status. Based on this, data statistics and analysis can also be performed for the entire class or the entire grade to generate corresponding learning diagnosis reports.

[0102] Figure 6 A schematic example of a job processing system according to this disclosure is shown. Figure 6 As shown, the assignment processing system includes: a smart terminal 610 on the teacher's / student's side, which acquires relevant information written by the teacher / student on the paper and transmits it to the processing backend via a communication network 620; and a processing backend 630, which generates a reconstructed writing trajectory based on the processing of the information; and associates the reconstructed writing trajectory with the corresponding area of ​​the electronic template file corresponding to the paper.

[0103] According to embodiments of this disclosure, an electronic device for electronically processing work content is also proposed. For example... Figure 7As shown, the device includes a processor 710, a memory 720 coupled to the processor, and a communication interface 730. The memory 720 stores program instructions, and the communication interface 730 receives information about handwriting and corresponding position coordinates from various smart terminals. The processor is configured to cause the device to implement the method of electronically processing the work content as described above when executing the program instructions stored in the memory.

[0104] According to embodiments of this disclosure, a computer-readable storage medium storing program instructions is also provided, which, when executed by a processor in an electronic device, causes the electronic device to implement the method for electronically processing work content as described above.

[0105] The methods, apparatus, systems, and computer-readable storage media for electronic processing of assignments disclosed herein can maintain students' habits of writing assignments and teachers' habits of grading assignments. At the same time, it is conducive to the analysis, diagnosis, statistics, and summarization of students' answers to assignments and teachers' grading, thereby facilitating the diagnosis of students' learning situation and enabling teachers to grasp students' current learning status in a timely manner so as to provide more targeted teaching guidance and make further lesson preparation arrangements.

[0106] It should be understood that the various embodiments and terminology used in this disclosure are not intended to limit the technical features disclosed herein to the specific embodiments disclosed herein; rather, this disclosure should be construed as covering various modifications, equivalents, or substitutions of the embodiments of this disclosure. Regarding the description of figures, similar or related components may be represented by similar figure references. As used herein, the singular form of a noun corresponding to an item may include one or more items unless the context clearly indicates otherwise. In the content of this disclosure, each expression used herein, such as “A or B,” “at least one of A and B,” “at least one of A or B,” “A, B, or C,” “one or more of A, B, and C,” or “one or more of A, B, or C,” may include any and all combinations of one or more of the associated listed items. Expressions such as “first,” “second,” “first,” or “second” may be used only for the purpose of distinguishing one component from other components, and not to limit the respective component in other respects (e.g., importance or order). It should be understood that if an element (e.g., the first element) is referred to as “coupled to,” “coupled to,” “connected to,” or “connected to,” regardless of whether the terms “operationally” or “communically” are used, it means that the element can be coupled to another element directly (e.g., wired), wirelessly, or via a third element.

[0107] Various embodiments of this disclosure can be implemented by software (e.g., a program) including instructions stored in a machine-readable storage medium (e.g., internal or external memory). For example, a processor of the machine (e.g., an electronic device) can invoke the instructions from the machine-readable storage medium and execute the invoked instructions. This means that the machine can perform at least one function based on at least one invoked instruction. One or more instructions may include code generated by a compiler or executable by an interpreter. The machine-readable storage medium may be provided in the form of a non-transitory storage medium. Here, the term "non-transitory" as used herein means that the storage medium is tangible but does not include signals (e.g., electromagnetic waves). The term "non-transitory" does not distinguish between cases where data is permanently stored in the storage medium and cases where data is temporarily stored in the storage medium. For example, a non-transitory storage medium may include a buffer in which data is temporarily stored.

[0108] According to embodiments, the methods according to the various embodiments disclosed in this disclosure can be provided as part of a computer program product. The computer program product can be traded as a product between a seller and a buyer. The computer program product can be distributed in the form of a machine-readable storage medium (e.g., an optical disc read-only memory (CD-ROM)), or distributed directly online (e.g., downloaded or uploaded) through an app store (e.g., the Play Store), or distributed between two user devices (e.g., smartphones). In the case of online distribution, at least a portion of the computer program product (e.g., a downloadable app) can be temporarily stored or generated in a machine-readable storage medium, such as the memory of a manufacturer's server, an app store's server, or a relay server.

[0109] According to various embodiments, each component (e.g., a module or program) described above may include one or more entities. According to various embodiments, at least one or more of the aforementioned components or operations may be omitted, or one or more components or operations may be added. Alternatively or additionally, some components (e.g., modules or programs) may be integrated into a single component. In this case, the integrated component may perform the same or similar functions performed by each corresponding component prior to integration. According to various embodiments, operations performed by modules, programs, or other components may be performed sequentially, in parallel, repeatedly, or heuristically, or at least some operations may be performed in a different order, omitted, or additional operations may be added.

[0110] The above description represents only optional embodiments of this disclosure. Any structural modifications, improvements, and refinements made by any person without departing from the principles of this disclosure are considered to be within the protection scope of this disclosure.

[0111] Several implementations have been described. However, it should be understood that various modifications can be made. For example, elements of different implementations can be combined, supplemented, modified, or removed to produce other implementations. Furthermore, those skilled in the art will understand that other structures and processes can replace those disclosed, and the resulting implementations will perform at least substantially the same functions as the disclosed implementations in at least substantially the same manner, to achieve at least substantially the same effects as the disclosed implementations. Accordingly, these and other implementations are conceived in this application.

Claims

1. A method for digitizing work content, comprising: The system receives first information transmitted from a first smart terminal, which is associated with a first trajectory written by a first user on a piece of paper that does not have a specific format designed for the question information itself, wherein an invisible dot matrix pattern with position coordinates is pre-laid on the paper. Receive second information transmitted from the second smart terminal that is associated with the second trajectory written by the second user on the paper; Based on the first information, the first trajectory is used to perform trajectory shape recognition to determine the question information associated with the first trajectory, wherein the question information includes a question number, and wherein the first information includes at least one of information about the first trajectory and information about the position coordinates of the first trajectory on the paper; Based on the second information, trajectory position identification is performed on the second trajectory to determine the question information associated with the second trajectory, wherein the second information includes at least one of information about the second trajectory and information about the position coordinates of the second trajectory on the paper; and The determined question information associated with the first trajectory and the question information associated with the second trajectory are associated with the corresponding questions in the electronic template file corresponding to the paper, wherein the electronic template file is an electronic original corresponding to the question stored in the database.

2. The method according to claim 1, wherein, The first smart terminal and the second smart terminal are dot matrix pens; the first trajectory is the handwriting of the first user on the paper when answering questions using the first smart terminal; The second trajectory is the correction handwriting made by the second user on the paper using the second smart terminal to correct the first user's answer.

3. The method according to claim 2, further comprising: Based on the first information, a reconstructed handwriting is generated on the first image, wherein the position of the reconstructed handwriting on the first image is determined based on information about the position coordinates of the first trajectory on the paper; as well as Based on the second information, a reconstructed correction handwriting is generated on the second image, wherein the position of the reconstructed correction handwriting on the second image is determined based on information about the position coordinates of the second trajectory on the paper. The first image and the second image may be the same image or different images.

4. The method according to claim 3, further comprising: Perform image grayscale processing on the first image / second image to obtain the corresponding grayscale image; The grayscale image is binarized to obtain a pre-processed image. The image after image preprocessing is subjected to structured analysis and segmented into multiple regions. The segmentation of the multiple regions is based on the positions of the reconstructed answer handwriting / the reconstructed correction handwriting on the first image and the second image, respectively. The reconstructed handwriting is subjected to trajectory shape recognition to determine the question information associated with the reconstructed handwriting or the answer information associated with the reconstructed handwriting. The reconstructed correction handwriting is subjected to trajectory position recognition to determine the question information associated with the reconstructed correction handwriting, and the reconstructed correction handwriting is subjected to trajectory shape recognition to determine the correction information associated with the reconstructed correction handwriting.

5. The method according to claim 4, wherein, The plurality of areas includes at least one of the following: an assignment number area, a question number area, a first user identity information identification area, an answer area, an objective question grading area, and a subjective question grading area. The answer information includes the answers written by the first user for objective questions, and the grading information includes information on the second user's grading of the first user's answers for subjective questions.

6. The method according to claim 5, further comprising: The reconstructed correction handwriting is subjected to trajectory location identification to determine at least one of the following associated with the reconstructed correction handwriting: the job number, the identity information of the first user.

7. The method according to claim 5, further comprising: The reconstructed handwriting is subjected to trajectory shape recognition to determine at least one of the following associated with the reconstructed handwriting: the assignment number and the question number.

8. The method according to claim 5, further comprising: A pre-trained trajectory shape recognition model is used to perform trajectory shape recognition on the reconstructed handwriting in at least one of the assignment number region, the question number region, and the answer region. as well as The pre-trained trajectory shape recognition model is used to perform trajectory shape recognition on the reconstructed correction handwriting in the subjective question correction area; The trajectory shape recognition model includes at least one of the following: a digital trajectory shape recognition model, a letter trajectory shape recognition model, a hook-and-cross trajectory shape recognition model, and a special marker trajectory shape recognition model.

9. The method according to any one of claims 4-8, further comprising: Based on the analysis and statistics of at least one of the question information, answer information, and grading information, statistical data on the first user's answers to the question is generated.

10. The method according to any one of claims 4-8, wherein, The first information further includes identification information about a first smart terminal, which is associated with the first user; the second information further includes identification information about a second smart terminal, which is associated with the second user; the method further includes: Based on the identification information about the first smart terminal, the reconstructed handwriting is associated with the first user; and Based on the identification information about the second smart terminal, the reconstructed handwriting is associated with the second user.

11. An apparatus for electronically processing work content, the apparatus comprising a processor and a memory coupled to the processor, wherein, The processor is configured to cause the apparatus to perform the method according to any one of claims 1-10 when executing program instructions stored in the memory.

12. A computer-readable storage medium storing program instructions that, when executed by a processor in an electronic device, cause the electronic device to perform the method according to any one of claims 1-10.

13. A system for electronically processing work content, comprising: The first intelligent terminal acquires first information associated with the first trajectory written by the first user on paper; The second intelligent terminal acquires second information associated with the second trajectory written by the second user on the paper; as well as The apparatus according to claim 11.

Images