Connection question correction method, storage medium and electronic equipment

Abstract

The invention discloses a connection question correction method, a storage medium and electronic equipment. The method comprises the following steps: constructing a connection question bank; and identifying a connection structure, a connection object and a connection relationship in the to-be-corrected connection question image, obtaining a student answer, matching a standard answer corresponding to the connection question in the connection question bank with the student answer, if the standard answer is completely matched with the student answer, determining that the correction result is correct, otherwise, determining that the correction result is wrong. According to the method, through recognition of the connection structures and the number of the levels, in combination with connection object naming and link representation of the connection relation between the connection objects, the method can be suitable for correcting connection questions of various connection structures, and is also suitable for correcting connection questions containing multiple link relations in answers; the automation and the accuracy of correcting the connection questions can be effectively improved.

Description

Technical Field

[0001] The present invention relates to the technical field of intelligent marking, and in particular to a connecting question marking method, a storage medium and an electronic device. Background Art

[0002] At present, photo-taking and judging questions has been widely used in various electronic educational devices. Users can take photos of test questions through terminal devices and upload them to the application. The application uses the question type judgment model to perform layout analysis and text analysis on the image to obtain the overall layout frame and text frame. At the same time, according to the characteristics of the question type, the layout frame and text frame corresponding to each test question are analyzed, and then the layout frame and text frame of each test question are cropped. The cropped image is used to recognize the content of the question through the recognition model, and the question is judged according to the content of the question.

[0003] However, the above method can only handle limited question types, such as arithmetic questions (horizontal, vertical, and freehand), application questions, etc., but there are still quite a few question types that cannot be automatically graded using the above method, such as connection questions. Due to the variety of connection questions, it is generally difficult to automatically grade them using conventional logical rules. Therefore, how to accurately and efficiently grade connection questions has become an urgent problem to be solved in the process of automated marking. Summary of the invention

[0004] In view of the deficiencies in the prior art, the present invention aims to provide a connecting-line question grading method, a storage medium and an electronic device.

[0005] In order to achieve the above object, the present invention adopts the following technical solution:

[0006] A method for correcting a connecting question comprises the following steps:

[0007] S1. Construct a connection question database to store each connection question and its question data information, wherein the question data information includes connection structure and connection relationship;

[0008] The connection structure includes the layout structure of the connection objects in the connection question and the corresponding number of levels; the construction process of the connection relationship is: first, according to the number of levels of the connection structure, each connection object of each layer is named in sequence according to the set naming rules; after naming the connection objects of each layer, all the connection relationships included in the standard answer are represented by links, and each link includes the names of two connection objects that have a connection relationship;

[0009] S2, retrieve the connecting-line question from the connecting-line question database for the user to answer. After the user completes the answer, obtain the connecting-line question image to be corrected, use the first target retrieval model to identify the connection structure of the connecting-line question image to be corrected, and identify and locate each layer in the connection structure; then use the second target detection model to identify each connecting line object in each layer of the connection structure and obtain the regional coordinates of each connecting line object, and then name the connecting line object of each layer according to the same naming rules as the connecting-line question database; then use the third target detection model to detect each connecting line in the connecting-line question image to be corrected, and obtain the coordinates of both ends of each connecting line;

[0010] S3, according to the area coordinates of each connection object in the connection problem image to be corrected and the coordinates of the two ends of each connection line, the two connection objects connected by each connection line are matched, so as to obtain the connection relationship between each connection object in the connection problem image to be corrected, and then the names of each connection object obtained by naming in step S2 are used to construct links to represent the connection relationship existing in the connection problem image to be corrected, and all the constructed links constitute the student's answer;

[0011] S4. Match the standard answer of the corresponding connect-the-difference question in the connect-the-difference question database with the student answer identified in step S3. If they are completely matched, the grading result is correct, otherwise the grading result is wrong.

[0012] Further, in step S2, the target to be detected by the third target detection model is a straight line whose two ends are respectively connected to two connection objects.

[0013] Furthermore, in step S4, before matching, it is first determined whether the number of links included in the standard answer is consistent with the number of links included in the student answer. If they are inconsistent, the grading is terminated and a detection error exception is prompted.

[0014] The present invention also provides a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, and the computer program implements the above method when executed by a processor.

[0015] The present invention also provides an electronic device, comprising a processor and a memory, wherein the memory is used to store a computer program; and when the processor is used to execute the computer program, the above method is implemented.

[0016] The beneficial effects of the present invention are as follows: the present invention can be applied to the correction of connection questions with various connection structures, and also to the correction of connection questions whose answers contain multiple link relationships, through the identification of connection structure and number of levels, combined with the naming of connection objects and the use of links to represent the connection relationship between connection objects, and can effectively improve the automation and accuracy of the correction of connection questions. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 12 is an example diagram of a connection structure in an embodiment of the present invention. DETAILED DESCRIPTION

[0018] The present invention will be further described below in conjunction with the accompanying drawings. It should be noted that this embodiment is based on the technical solution and provides a detailed implementation method and specific operation process, but the protection scope of the present invention is not limited to this embodiment.

[0019] This embodiment provides a method for correcting a connecting question, comprising the following steps:

[0020] S1. Construct a connection question database to store the question data information of each connection question, wherein the question data information includes the connection structure and the connection relationship;

[0021] The connection structure includes the layout structure of the connection objects in the connection problem and the corresponding number of levels. The layout structure is quite diverse, such as Figure 1 As shown, it can be a left-right structure (such as Figure 1 (a)), upper and lower structures (such as Figure 1 (b)), left-center-right structure (such as Figure 1 (c)), upper-middle-lower structure (such as Figure 1 (d)) etc., accordingly, the left-right structure and the top-bottom structure are two-layer structures, and the left-middle-right structure and the top-middle-bottom structure are three-layer structures. In addition, there can be more layout structures and more levels, which are only exemplified here.

[0022] The process of building a connection relationship is as follows: first, according to the number of levels of the connection structure and the set naming rules, name each connection object in each layer in sequence (such as Figure 1 As shown, for a two-layer structure, the first layer is named with capital letters A, B, C, etc., and the second layer is named with numbers 1, 2, 3, etc. For a three-layer structure, the first layer is named with capital letters A, B, C, etc., the second layer is named with numbers 1, 2, 3, etc., and the third layer is named with lowercase letters a, b, c. There may be other naming rules, which are only used as examples here). After naming the connection objects of each layer, all connection relationships contained in the standard answer are represented by links, and each link contains the names of two connection objects with a connection relationship. For example, between the first layer and the second layer, A and 1 are connected, then the link can be represented as A-1, and between the second layer and the third layer, 1 and a are connected, then the link is represented as 1-a.

[0023] S2. Retrieve a connect-the-wire question from the connect-the-wire question database for the user to answer. After the user has answered, obtain the connect-the-wire question image to be corrected, use the first target retrieval model to identify the connection structure of the connect-the-wire question image to be corrected, and identify and locate each layer in the connection structure; then use the second target detection model to identify each connect-the-wire object in each layer of the connection structure and obtain the regional coordinates of each connect-the-wire object, and then name the connect-the-wire objects in each layer according to the same naming rules as the connect-the-wire question database; then use the third target detection model to detect each connecting line in the connect-the-wire question image to be corrected, and obtain the coordinates of both ends of each connecting line.

[0024] It should be noted that the target to be detected by the third target detection model is a straight line whose two ends are respectively connected to two connection objects.

[0025] S3. According to the area coordinates of each connection object in the connection question image to be corrected and the coordinates of the two ends of each connection line, the two connection objects connected by each connection line are matched to obtain the connection relationship between the various connection objects in the connection question image to be corrected, and then the names of the various connection objects obtained by naming in step S2 are used to construct links to represent the connection relationship existing in the connection question image to be corrected. All the constructed links constitute the student's answer.

[0026] S4. Match the standard answer of the corresponding connect-the-difference question in the connect-the-difference question database with the student answer identified in step S3. If they are completely matched, the grading result is correct, otherwise the grading result is wrong.

[0027] Furthermore, in step S4, before matching, it is first determined whether the number of links included in the standard answer is consistent with the number of links included in the student answer. If they are inconsistent, the grading is terminated and a detection error exception is prompted.

[0028] It should be noted that the first target detection model, the second target detection model and the third target detection model can all adopt the YOLOv model based on deep learning and be pre-trained and optimized.

[0029] The method of this embodiment is suitable for various electronic devices such as tablets, mobile phones, and PCs.

[0030] For those skilled in the art, various corresponding changes and modifications can be made according to the above technical solutions and concepts, and all of these changes and modifications should be included in the protection scope of the claims of the present invention.

Claims

1. A method for correcting connecting questions, characterized in that: The steps include: S1. Construct a connection question database to store each connection question and its question data information, wherein the question data information includes connection structure and connection relationship; The connection structure includes the layout structure of the connection objects in the connection question and the corresponding number of levels; the construction process of the connection relationship is: first, according to the number of levels of the connection structure, each connection object of each layer is named in sequence according to the set naming rules; after naming the connection objects of each layer, all the connection relationships included in the standard answer are represented by links, and each link includes the names of two connection objects that have a connection relationship; S2, retrieving a connecting question from the connecting question database for the user to answer, and after the user completes the answer, obtaining an image of the connecting question to be corrected, identifying a connection structure of the image of the connecting question to be corrected using the first target retrieval model, and identifying and locating each layer in the connection structure; Then, the second object detection model is used to identify each line object in each layer of the connection structure and obtain the regional coordinates of each line object, and then the line objects in each layer are named according to the same naming rules as the line-connecting question database; then, the third object detection model is used to detect each line in the line-connecting question image to be corrected, and the coordinates of the two ends of each line are obtained; S3, according to the area coordinates of each connection object in the connection problem image to be corrected and the coordinates of the two ends of each connection line, the two connection objects connected by each connection line are matched, so as to obtain the connection relationship between each connection object in the connection problem image to be corrected, and then the names of each connection object obtained by naming in step S2 are used to construct links to represent the connection relationship existing in the connection problem image to be corrected, and all the constructed links constitute the student's answer; S4. Match the standard answer of the corresponding connect-the-difference question in the connect-the-difference question database with the student answer identified in step S3. If they are completely matched, the grading result is correct, otherwise the grading result is wrong.

2. The method according to claim 1, characterized in that In step S2, the target to be detected by the third target detection model is a straight line whose two ends are respectively connected to two connection objects.

3. The method according to claim 1, characterized in that In step S4, before matching, it is first determined whether the number of links contained in the standard answer is consistent with the number of links contained in the student answer. If they are inconsistent, the grading is terminated and a detection error exception is prompted.

4. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the method according to any one of claims 1 to 3 is implemented.

5. An electronic device, characterized in that: The method comprises a processor and a memory, wherein the memory is used to store a computer program; when the processor is used to execute the computer program, the method described in any one of claims 1 to 3 is implemented.

Images