CN117173721A

CN117173721A - Question correcting method, device, equipment and medium

Info

Publication number: CN117173721A
Application number: CN202311140268.0A
Authority: CN
Inventors: 刘军
Original assignee: Shenzhen Xingtong Technology Co ltd
Current assignee: Shenzhen Xingtong Technology Co ltd
Priority date: 2023-09-05
Filing date: 2023-09-05
Publication date: 2023-12-05

Abstract

The present disclosure relates to a method, apparatus, device and medium for correcting a title, including: acquiring a theme page image; identifying a printed text in the theme page image; according to the printing text, searching a standard page image corresponding to the same topic page as the topic page image in a preset first database; searching a target reference answer corresponding to the target question in a preset second database based on the standard page image aiming at the target question in the question page image, wherein the second database comprises the reference answer marked with the keyword; and correcting the answer content corresponding to the target questions according to the target reference answers. The method and the device can improve the accuracy of question correction.

Description

Question correcting method, device, equipment and medium

Technical Field

The disclosure relates to the technical field of image processing, and in particular relates to a method, a device, equipment and a medium for correcting a title.

Background

Photographing correction is a correction mode commonly used in education and teaching. At present, general questions with fixed answers can be accurately and efficiently photographed and corrected, however, some questions have the same answers but have various answer modes, such as common application questions, and a plurality of different questions solving processes can be adopted to obtain the same answers. The examination questions are very complicated to modify, and the accuracy of the current photographing modification mode is low.

Disclosure of Invention

To solve the above technical problems or at least partially solve the above technical problems, the present disclosure provides a topic modification method, apparatus, device, and medium.

According to an aspect of the present disclosure, there is provided a topic modification method including:

acquiring a theme page image;

identifying a printed text in the topic page image;

according to the printing text, searching a standard page image corresponding to the same title page as the title page image in a preset first database;

searching a target reference answer corresponding to the target question in a preset second database based on the standard page image aiming at the target question in the question page image, wherein the second database comprises reference answers marked with keywords;

and correcting the answer content corresponding to the target question according to the target reference answer.

According to another aspect of the present disclosure, there is provided a topic modification apparatus,

the image acquisition module is used for acquiring a question page image;

the text recognition module is used for recognizing the printed text in the theme page image;

the image retrieval module is used for retrieving standard page images corresponding to the same topic page with the topic page images in a preset first database according to the printing text;

The answer retrieval module is used for retrieving target reference answers corresponding to the target questions in the question page images based on the standard page images in a preset second database, wherein the second database comprises reference answers marked with keywords;

and the correcting module is used for correcting the answer content corresponding to the target question according to the target reference answer.

According to another aspect of the present disclosure, there is provided an electronic device including:

a processor;

a memory for storing the processor-executable instructions;

the processor is used for reading the executable instructions from the memory and executing the instructions to realize the topic correction method.

According to another aspect of the present disclosure, there is provided a non-transitory computer-readable storage medium storing computer instructions that, when executed on a terminal device, cause the terminal device to implement a topic modification method.

Compared with the prior art, the technical scheme provided by the embodiment of the disclosure has the following advantages:

the title correction method, device, equipment and medium provided by the embodiment of the disclosure comprise the following steps: acquiring a theme page image; identifying a printed text in the theme page image; according to the printing text, searching a standard page image corresponding to the same topic page as the topic page image in a preset first database; searching a target reference answer corresponding to the target question in a preset second database based on the standard page image aiming at the target question in the question page image, wherein the second database comprises the reference answer marked with the keyword; and correcting the answer content corresponding to the target questions according to the target reference answers. The method and the device can improve the accuracy of question correction.

Drawings

The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the disclosure and together with the description, serve to explain the principles of the disclosure.

In order to more clearly illustrate the embodiments of the present disclosure or the solutions in the prior art, the drawings that are required for the description of the embodiments or the prior art will be briefly described below, and it will be obvious to those skilled in the art that other drawings can be obtained from these drawings without inventive effort.

FIG. 1 is a flow chart of a method for topic modification provided in an embodiment of the present disclosure;

FIG. 2 is a schematic diagram of a topic page image and a standard page image provided by an embodiment of the present disclosure;

FIG. 3 is a schematic diagram illustrating the modification of solution content of different answer types provided by an embodiment of the present disclosure;

FIG. 4 is a schematic structural diagram of a question model according to an embodiment of the disclosure;

fig. 5 is a schematic structural diagram of a topic modification apparatus according to an embodiment of the present disclosure;

fig. 6 is a schematic structural diagram of an electronic device according to an embodiment of the disclosure.

Detailed Description

In order that the above objects, features and advantages of the present disclosure may be more clearly understood, embodiments of the present disclosure will be described in more detail below with reference to the accompanying drawings. While certain embodiments of the present disclosure have been shown in the accompanying drawings, it is to be understood that the present disclosure may be embodied in various forms and should not be construed as limited to the embodiments set forth herein, but are provided to provide a more thorough and complete understanding of the present disclosure. It should be understood that the drawings and embodiments of the present disclosure are for illustration purposes only and are not intended to limit the scope of the present disclosure.

It should be understood that the various steps recited in the method embodiments of the present disclosure may be performed in a different order and/or performed in parallel. Furthermore, method embodiments may include additional steps and/or omit performing the illustrated steps. The scope of the present disclosure is not limited in this respect.

The term "including" and variations thereof as used herein are intended to be open-ended, i.e., including, but not limited to. The term "based on" is based at least in part on. The term "one embodiment" means "at least one embodiment"; the term "another embodiment" means "at least one additional embodiment"; the term "some embodiments" means "at least some embodiments. Related definitions of other terms will be given in the description below. It should be noted that the terms "first," "second," and the like in this disclosure are merely used to distinguish between different devices, modules, or units and are not used to define an order or interdependence of functions performed by the devices, modules, or units.

It should be noted that references to "one", "a plurality" and "a plurality" in this disclosure are intended to be illustrative rather than limiting, and those of ordinary skill in the art will appreciate that "one or more" is intended to be understood as "one or more" unless the context clearly indicates otherwise.

The names of messages or information interacted between the various devices in the embodiments of the present disclosure are for illustrative purposes only and are not intended to limit the scope of such messages or information.

For some questions such as application questions and the like with the same answers but various answer modes, the current photographing correction mode is very complicated to correct, and the accuracy is low. Based on the above, in order to support photographing correction of topic types such as application topics and improve correction accuracy, embodiments of the present disclosure provide a topic correction method, device, apparatus and medium. For ease of understanding, embodiments of the present disclosure are described in detail below.

Fig. 1 provides a flow chart of a method of topic modification, which may be performed by an electronic device or a server. The electronic device or the server are both topic modification ends in the embodiments of the present disclosure. The electronic device may include a smart phone, a tablet computer, a desktop computer, a notebook computer, and other devices with communication functions. The server may be a cloud server or a server cluster, or other devices with storage and computing functions.

As shown in fig. 1, the topic modification method may include the steps of:

S102, acquiring a title page image.

In this embodiment, the theme page image may be obtained through an image selection operation, an image capturing operation, or an image uploading operation, where the theme page image includes at least one theme, such as a blank-filling theme, a judgment theme, a drawing theme, and an application theme, and each theme includes a theme content of a print body and a solution content of a handwriting body.

In a scenario taking homework as an example, electronic devices such as a mobile phone and a tablet computer can be used to photograph pages of a student who finishes homework, and then image processing such as correction, noise reduction and optical compensation is performed on photographed page images, so that high-quality theme page images are obtained. The topic page image can be stored in the electronic device and subjected to subsequent topic correction operation by the electronic device, or the topic page image can be uploaded to a server by the electronic device and subjected to topic correction operation by the server.

The embodiment utilizes the theme page image to carry out subsequent test questions correction, can shorten correction time, improve correction efficiency, avoid the influence of human factors on correction results and improve correction accuracy. In addition, in practical application, the method is very convenient to photograph or select the theme page image from the gallery, is not limited by time and place, so that theme correction can be performed anytime and anywhere, and correction convenience can be improved.

S104, identifying the printed text in the theme page image.

In one embodiment, fonts in the topic page image can be classified through a pre-trained font classification model to obtain a print class; determining each printing area image belonging to the printing body category in the theme page image; and performing text recognition on the images of each printing area through a preset text recognition model to obtain a printing text.

S106, searching a standard page image corresponding to the same title page as the title page image in a preset first database according to the printing text.

In the searching process, the standard page image corresponding to the same topic page with the topic page image can be searched in the first database for storing the standard page image according to the printing text in the topic page image and/or typesetting information among the topics; the standard page image is an electronic version image corresponding to the same title page as the title page image.

As shown in fig. 2, the topic page image displayed in the left graph is an image for truly answering the topic page, the standard page image in the right graph is an electronic version image for the same topic page, the standard page image and the topic page image comprise the same topic, and the page typesetting is the same.

Since the standard page image greatly reduces the retrieval range of the questions, the retrieval of the subsequent reference answers based on the standard page image is much easier.

S108, searching a target reference answer corresponding to the target question in a preset second database based on the standard page image aiming at the target question in the question page image, wherein the second database comprises the reference answer marked with the keyword.

The target title in this embodiment may be any title in the title page image. It can be understood that the standard page image corresponds to the questions in the question page image one by one, so that the current standard question corresponding to the target question in the question page image is determined in the standard page image, then the current reference answer corresponding to the current standard question is obtained from the second database storing the reference answer, and the target reference answer corresponding to the target question is determined.

The second database is used for storing the reference answers marked with the keywords, and an association relation is established between the reference answers and the standard questions, so that the reference answers associated with the current standard questions can be quickly retrieved from the second database according to the association relation. Keywords marked by reference answers are generally numbers, subjects and the like in the reference answers; taking the reference answer of "3 apples for Xiaoming and Xiaoang" as an example, the keyword may be "3". The keyword can express the key and invariable answer information in the reference answer, and meanwhile, the order, format and other information of the answer content are not limited in the form of word segmentation, based on the keyword, the keyword can be used as effective information for question correction, and when the keyword is used for correcting the questions (such as application questions) with fixed answer but flexible answer modes, correction can be made only by comparing part of vocabularies in the answer content with the keyword, and the correction mode can reduce correction difficulty and improve correction accuracy.

In the prior art, a current standard question corresponding to a target question is generally searched in a database of questions directly in a mode that a single question corresponds to a single question, and then a reference answer of the current standard question is used as a target reference answer. However, this way of finding questions alone can be slow to retrieve if the number of questions is large; moreover, for questions with fewer characters, such as drawing questions, accurate current standard questions can be difficult to retrieve. Compared with the method of searching the questions, the scheme provided in the steps S106 and S108 in the embodiment is that the standard page image is searched from the whole page image of the questions, and the matched standard page image can be quickly and accurately searched because the page is more in characteristics (such as text information, typesetting information and similar graphic characteristics) suitable for matching; and for the drawing questions with few characters, the image is utilized to search the image, so that the searching difficulty can be reduced, and the searching accuracy can be improved. After the standard page image is searched, the searching range of the questions is greatly reduced, so that the current standard questions of the topic titles are determined based on the standard page image, and the reference answers of the current standard questions are given to the target questions, and the target reference answers can be obtained rapidly and accurately. Therefore, the above-mentioned method of searching the image and then determining the reference answer can improve the answer searching efficiency and accuracy.

S110, correcting the answer content corresponding to the target subject according to the target reference answer.

In this embodiment, the topic area of the target topic in the topic page image may be detected first, and then the key text content in the topic area may be detected, where the key text content includes: characters such as "answer" or "solution", underlines, brackets; determining an area where the key text content in the question area is located as a solution area; solution content in the solution area is identified.

In a possible embodiment, it may be compared whether the solution content is consistent with the target reference answer; if the answers are consistent, the answer content can be determined to be correct. If the answers are inconsistent, the answer content is not necessarily wrong, but the answer content is possibly another answer mode different from the target reference answer; because the keywords in the answers are fixed in any answer mode, the embodiment can continuously compare whether the keywords in the answer content are consistent with the keywords marked by the target reference answers; if the words are consistent, the correct answer content can be determined; if the words are inconsistent, a solution content error may be determined.

The title correction method provided by the embodiment of the disclosure obtains a title page image; identifying a printed text in the theme page image; according to the printing text, searching a standard page image corresponding to the same topic page as the topic page image in a preset first database; searching a target reference answer corresponding to the target question in a preset second database based on the standard page image aiming at the target question in the question page image, wherein the second database comprises the reference answer marked with the keyword; and correcting the answer content corresponding to the target questions according to the target reference answers. In the technical scheme, firstly, the topic page image is acquired and topic modification is carried out based on the topic page image, so that modification time can be greatly shortened, and modification efficiency is improved; secondly, searching a standard page image according to the printing text in the topic page image, and searching a target reference answer based on the standard page image, so that the searching efficiency and the accuracy can be improved; then, correcting the answer content according to the target reference answer marked with the keyword, wherein the keyword can express the fixed answer information of the keyword in the target reference answer, and the word segmentation form is not limited to the fixed answer form; therefore, the target reference answer and the keywords thereof are utilized for correction, so that correction difficulty can be reduced, and correction accuracy can be improved.

With respect to the above step S108, the present embodiment provides a specific process of retrieving a target reference answer corresponding to a target question in a preset second database based on a standard page image, which includes the following.

A topic area of a target topic in the topic page image is detected. In one approach, question number information and text line information in a question page image may be detected; determining boundary coordinate information of each text line according to the text line information and the question number information; determining the area range of each topic based on the boundary coordinate information; the topic area of the current target topic is determined based on the area range of each topic. Alternatively, the topic page image may be input to a pre-trained segmentation model, such that the segmentation model outputs a bounding box bounding the topic area, thereby obtaining the topic area of the target topic. The above is merely an example of detecting each topic area, and other detection methods may be adopted in practical applications.

Determining a standard question corresponding to the question area in the standard page image; it will be appreciated that the standard page images are in one-to-one correspondence with the topics in the topic page images, and that standard topics corresponding to the topic areas in the topic page images are then determined in the standard page images.

Acquiring a second database associated with the standard page image; searching a reference answer of the standard question in a second database; and determining the retrieved reference answer as a target reference answer of the target question.

In this embodiment, a corresponding reference answer may be generated for each standard question in the standard page image, and the keywords of the reference answer may be marked. And storing the reference answers marked with the keywords in a second database, establishing an association relationship between the standard page images and the second database, and establishing an association relationship between each standard question and the reference answer. In this case, a reference answer to the standard question may be retrieved from the second database and determined as a target reference answer to the target question.

The reference answer in the present embodiment can be obtained as follows. The reference answer for each question may be provided in a standard page image, such as a teacher's book; alternatively, the answer position may be a blank state to be filled in, as with the question page. For standard page images providing reference answers, the reference answer of the text format of each question can be directly extracted, and the reference answer is associated with the corresponding standard question. For a standard page image with blank answer positions, referring to fig. 2, a reference answer is filled in at the answer positions of the questions in a manual writing manner, and for the convenience of extracting the reference answer, a complete reference answer is also written in a manual writing manner for the answer of the blank filling type; then, a reference answer in the text format of each question is extracted, and the reference answer is associated with the corresponding standard question.

The embodiment can mark the keywords in the reference answers manually; it is also possible to extract keywords in the reference answers and tag the reference answers with their keywords according to some keyword extraction algorithms, such as TF-IDF (Term Frequency-inverse text Frequency index) algorithm, textRank algorithm, RAKE (Rapid Automatic Keyword Extraction, fast automatic keyword extraction) algorithm.

According to the above manner, when the standard question corresponding to the target question is determined, the reference answer of the keyword marked corresponding to the standard question can be obtained, and then the reference answer is determined as the target reference answer corresponding to the target question.

In this embodiment, the answer content corresponding to the target question is modified according to the target reference answer marked with the keyword, and the implementation process of this step may be shown in fig. 3.

The embodiment comprises the following steps: obtaining answer types of answer contents in the target questions; and correcting the answer content according to the answer type.

The standard questions or the target reference answers corresponding to the target questions can be marked with answer types in advance, namely answer types of answer contents. Alternatively, the answer type may be obtained by classifying the answer content.

The types of responses described above can be generally classified into: including mixed types of alphanumeric and numeric, pictorial, purely numeric, and purely textual. Different ways of modifying the different answer types can be adopted, and are described below.

In this embodiment, when the answer type is determined to be a mixed answer type including text and number, the answer content is modified according to the target reference answer and the keywords marked by the target reference answer through a preset question judgment model.

The answer type is a mixed type containing an alphanumeric character, and for example, the following answer content "4 m long per segment". For the mixed type answer content, the embodiment can correct the answer content according to the target reference answer and the marked keywords thereof through the judgment question model.

In one example, the decision problem model is, for example, a BERT (Bidirectional Encoder Representations from Transformers, transformer-based bi-directional encoder representation) model. BERT is a pre-trained language model based on a transducer architecture, the main goal of which is to learn a generic language representation through pre-training so that fine tuning can be done in a variety of natural language processing tasks.

The pretraining process of BERT includes two phases: masked Language Model (MLM, mask language model) and Next Sentence Prediction (NSP, next sentence prediction). In the MLM phase, BERT will randomly replace some words in the input sentence with "[ MASK ]" tags, and then train a model to predict these replaced words. In the NSP phase, BERT will input two sentences and train a model to determine whether the two sentences are consecutive.

By the pre-training of the two stages, the BERT can learn the context information in the sentence, thereby better understanding the meaning of the sentence. In the fine tuning stage, BERT may be used for various natural language processing tasks, such as text classification, question-answering systems, named entity recognition, and the like. BERT can achieve very good performance in natural language processing tasks.

As shown in fig. 4, the question model in the present embodiment may include: the embedding module, the coding module and the decoding module are based on the above, and the process of correcting the answer content according to the target reference answer and the marked keywords thereof by using the preset question judging model can be referred to as follows.

And mapping the solution content into a solution matrix, mapping the target reference answers into a reference answer matrix, and mapping the keywords of the target reference answers into a word matrix through an embedding module.

And determining a first correlation of the solution matrix with the reference answer matrix at the semantic level and a second correlation of the solution matrix with the word matrix at the semantic level through the coding module.

Identifying, by the decoding module, semantic relativity of semantics of the answer content and semantics of the target reference answer according to the first relativity, and identifying whether keywords of the target reference answer are contained in the answer content according to the second relativity; and then, correcting the answer content according to the semantic relevance and the keywords.

According to the embodiment, according to the semantic correlation between the semantic meaning of the first correlation identification solution content and the semantic meaning of the target reference answer, the semantic correlation is judged to be higher than a preset correlation threshold. If the semantic relevance is not higher than the preset relevance threshold, the meaning of the answer content is different from the meaning expressed by the target reference answer, and then the answer content error can be determined. If the semantic relevance is higher than the preset relevance threshold, the answer content is indicated to be the same as the semantic expressed by the target reference answer, and the second relevance can be combined to improve the correction accuracy. Then, whether the keyword of the target reference answer is contained in the answer content is identified based on the second correlation. When the keyword is included in the answer content, it is determined that the answer content is correct.

In order to avoid misjudgment, the embodiment can also generate audit prompt information for prompting users such as teachers, parents and the like to manually judge whether the answer content is correct or not under the condition that the semantic relevance is not higher than a relevance threshold value and/or the answer content contains keywords. After that, the embodiment may receive the judgment result about the answer content fed back by the user, and label the judgment result for the answer content.

According to the embodiment, the question judging model is used, and the answer content is corrected based on the target reference answer and the keywords thereof, so that correction accuracy can be improved, support types for complex questions such as application questions are increased, and therefore, the questions can be corrected for students more efficiently.

In the present embodiment, in response to determining that the answer type is the drawing type, image similarity between the image in the target reference answer and the image in the answer content is determined; and correcting the answer content according to the image similarity.

Wherein, the answer type is a drawing type, for example, the answer content is drawing an equilateral triangle. For the answer content of the mapping type, the embodiment can perform topic correction according to the similarity between images. Specifically, a first image in the answer content and a second image in the target reference answer are extracted; for comparison, the first image and the second image may be scaled to two images of smaller size difference. The first image and the second image are respectively processed into 256-dimensional matrixes through the ResNet18 network model, the Euclidean distance between the two matrixes is calculated, and the Euclidean distance is used for representing the image similarity between the first image and the second image. The smaller the Euclidean distance is, the higher the image similarity is, and the correct answer content is determined under the condition that the Euclidean distance is smaller than a preset distance threshold, namely the image similarity is higher than a preset similarity value; otherwise, determining that the answer content is wrong under the condition that the image similarity is not higher than a preset similarity value.

In the embodiment, determining the text matching degree of the target reference answer and the answer content in response to the fact that the answer type is determined to be the pure number type; and correcting the answer content according to the text matching degree.

In the embodiment, in response to determining that the answer type is a plain text type, identifying answer content; when the answer content is identified as the word, correcting the answer content according to the semantic similarity between the target reference answer and the answer content; when the answer content is identified as a sentence, correcting the answer content according to the target reference answer and the marked keywords thereof.

The embodiment can identify the information such as text length, text semantics, part of speech, named entity and the like of the answering content, and determine whether the answering content is a word or a sentence according to the information. If the answer content is identified as a word, the target reference answer is correspondingly a word, and based on the word, the answer content can be modified according to the semantic similarity between the target reference answer and the answer content. It will be appreciated that the semantic similarity between the target reference answer and the answer content is high, indicating that there may be a hyponym, such as "same" and "equal", between the answer content and the target reference answer, in which case it may be determined that the answer content is correct.

If the answer content is recognized as a sentence, the answer content may be modified according to the target reference answer and the keywords marked by the target reference answer with reference to the above-described answer content of a mixed type including the words and numerals. For example, identifying whether the semantic similarity of the answer content to the target reference answer is above a preset relevance threshold; and if so, identifying whether the word meaning between the word segmentation in the answer content and the key word of the target reference answer is the same, and if so, determining that the answer content is correct. And, in the event that at least one of the above items is not, determining that the answer content is wrong.

The above embodiment adopts different correction modes aiming at the answer contents of different answer types, can increase the supporting effect on more question correction and improve the accuracy of question correction.

After each topic is modified according to the above embodiment, the present embodiment may further calculate a total score of all topics in the whole topic page image according to the preset score of each topic. Further, the method may further include: and receiving comments, suggestions and other information submitted by the user. The title page image, the title correction result, the total score, the comment and other information can be stored in the electronic equipment or uploaded to the cloud end, so that the user can conveniently check and trace.

In summary, in the topic modification method provided by the embodiment of the disclosure, the topic page image is utilized to modify the topic, so that modification time can be shortened, and modification efficiency, modification accuracy and modification convenience are improved. When the standard page image is searched according to the printed text in the question page image, when the image is searched from the whole page, the matched features are more in the whole page, so that the matched standard page image can be quickly and accurately searched, and for the drawing questions with fewer characters, the searching difficulty can be reduced and the searching accuracy can be improved by using the image to search the image. After the standard page image is searched, the search range of the questions is greatly reduced, so that the target reference answer is searched based on the standard page image, and the search efficiency and the accuracy can be improved. When the target questions are corrected, correcting the answer content according to the target reference answers marked with the keywords, wherein the keywords of the target reference answers can express the key and unchanged answer information in the answers, and meanwhile, the form of word segmentation is not limited to a fixed answer form; therefore, when the keyword is used for correcting the questions (such as application questions) with fixed answer and flexible answer modes, correction can be made by only comparing part of vocabularies in the answer contents with the keyword, and the correction mode can reduce correction difficulty and improve correction accuracy.

The embodiment of the disclosure further provides a device for implementing the topic modification method, and the device is described below with reference to fig. 5. In the embodiment of the disclosure, the topic modification apparatus may be an electronic device or a server. The electronic device may include a mobile phone, a tablet computer, a desktop computer, a notebook computer, and other devices with communication functions. The server may be a cloud server or a server cluster, or other devices with storage and computing functions.

Fig. 5 shows a schematic structural diagram of a topic modification apparatus according to an embodiment of the present disclosure.

As shown in fig. 5, the topic modification apparatus 200 may include:

an image acquisition module 210, configured to acquire a topic page image;

a text recognition module 220 for recognizing a printed text in the theme page image;

the image retrieving module 230 is configured to retrieve, according to the printed text, a standard page image corresponding to the same topic page as the topic page image in a preset first database;

the answer retrieval module 240 is configured to, for a target question in the question page image, retrieve a target reference answer corresponding to the target question in a preset second database based on the standard page image, where the second database includes reference answers marked with keywords;

And the correcting module 250 is used for correcting the answer content corresponding to the target question according to the target reference answer.

In one embodiment, the answer retrieval module 240 is further configured to:

detecting a topic area of the target topic in the topic page image;

determining a standard question corresponding to the question area in the standard page image;

acquiring a second database associated with the standard page image;

retrieving a reference answer to the standard question in the second database;

and determining the retrieved reference answer as a target reference answer of the target question.

In one embodiment, the modification module 250 is further configured to:

obtaining answer types of answer contents in the target questions;

and correcting the answer content according to the answer type.

In one embodiment, the modification module 250 is further configured to:

determining image similarity between the image in the target reference answer and the image in the answer content in response to determining that the answer type is a mapping type;

correcting the answer content according to the image similarity,

wherein, the correction module 250 is further configured to:

Determining a text matching degree of the target reference answer and the answer content in response to determining that the answer type is a pure number type;

and correcting the answer content according to the text matching degree.

In one embodiment, the modification module 250 is further configured to:

and when the answer type is determined to be a mixed answer type containing words and numbers, correcting the answer content according to the target reference answer and the marked keywords thereof through a preset judgment model.

In one embodiment, the question model includes: an embedding module, an encoding module and a decoding module, the modifying module 250 is further configured to:

mapping the answer content into an answer matrix, mapping the target reference answer into a reference answer matrix, and mapping the key words of the target reference answer into a word matrix through the embedding module;

determining, by the encoding module, a first correlation of the solution matrix with the reference answer matrix at a semantic level and a second correlation of the solution matrix with the word matrix at the semantic level;

identifying, by the decoding module, a semantic correlation of the semantic meaning of the answer content with the semantic meaning of the target reference answer according to the first correlation, and identifying whether the answer content contains a keyword of the target reference answer according to the second correlation;

And correcting the answer content according to the semantic relativity and the keywords.

In one embodiment, the modification module 250 is further configured to:

identifying the answer content in response to determining that the answer type is a plain text type;

when the answer content is identified as a word, correcting the answer content according to the semantic similarity between the target reference answer and the answer content;

when the answer content is identified as a sentence, correcting the answer content according to the target reference answer and the marked keywords thereof.

The device provided in this embodiment has the same implementation principle and technical effects as those of the foregoing method embodiment, and for brevity, reference may be made to the corresponding content of the foregoing method embodiment where the device embodiment is not mentioned.

The exemplary embodiments of the present disclosure also provide an electronic device including: at least one processor; and a memory communicatively coupled to the at least one processor. The memory stores a computer program executable by the at least one processor for causing the electronic device to perform a method according to embodiments of the present disclosure when executed by the at least one processor.

The present disclosure also provides a computer program product comprising a computer program, wherein the computer program, when executed by a processor of a computer, is for causing the computer to perform a method according to embodiments of the disclosure.

Referring to fig. 6, a block diagram of an electronic device 300 that may be a server or a client of the present disclosure, which is an example of a hardware device that may be applied to aspects of the present disclosure, will now be described. Electronic devices are intended to represent various forms of digital electronic computer devices, such as laptops, desktops, workstations, personal digital assistants, servers, blade servers, mainframes, and other suitable computers. The electronic device may also represent various forms of mobile devices, such as personal digital processing, cellular telephones, smartphones, wearable devices, and other similar computing devices. The components shown herein, their connections and relationships, and their functions, are meant to be exemplary only, and are not meant to limit implementations of the disclosure described and/or claimed herein.

As shown in fig. 6, the electronic device 300 includes a computing unit 301 that can perform various appropriate actions and processes according to a computer program stored in a Read Only Memory (ROM) 302 or a computer program loaded from a storage unit 308 into a Random Access Memory (RAM) 303. In the RAM 303, various programs and data required for the operation of the device 300 may also be stored. The computing unit 301, the ROM 302, and the RAM 303 are connected to each other by a bus 304. An input/output (I/O) interface 305 is also connected to bus 304.

Various components in the electronic device 300 are connected to the I/O interface 305, including: an input unit 306, an output unit 307, a storage unit 308, and a communication unit 309. The input unit 306 may be any type of device capable of inputting information to the electronic device 300, and the input unit 306 may receive input numeric or character information and generate key signal inputs related to user settings and/or function controls of the electronic device. The output unit 307 may be any type of device capable of presenting information and may include, but is not limited to, a display, speakers, video/audio output terminals, vibrators, and/or printers. Storage unit 308 may include, but is not limited to, magnetic disks, optical disks. The communication unit 309 allows the electronic device 300 to exchange information/data with other devices through a computer network, such as the internet, and/or various telecommunications networks, and may include, but is not limited to, modems, network cards, infrared communication devices, wireless communication transceivers and/or chipsets, such as bluetooth (TM) devices, wiFi devices, wiMax devices, cellular communication devices, and/or the like.

The computing unit 301 may be a variety of general and/or special purpose processing components having processing and computing capabilities. Some examples of computing unit 301 include, but are not limited to, a Central Processing Unit (CPU), a Graphics Processing Unit (GPU), various specialized Artificial Intelligence (AI) computing chips, various computing units running machine learning model algorithms, a Digital Signal Processor (DSP), and any suitable processor, controller, microcontroller, etc. The computing unit 301 performs the respective methods and processes described above. For example, in some embodiments, the topic modification method may be implemented as a computer software program tangibly embodied on a machine-readable medium, such as storage unit 308. In some embodiments, part or all of the computer program may be loaded and/or installed onto the electronic device 300 via the ROM 302 and/or the communication unit 309. In some embodiments, the computing unit 301 may be configured to perform the topic modification method by any other suitable means (e.g., by means of firmware).

Program code for carrying out methods of the present disclosure may be written in any combination of one or more programming languages. These program code may be provided to a processor or controller of a general purpose computer, special purpose computer, or other programmable data processing apparatus such that the program code, when executed by the processor or controller, causes the functions/operations specified in the flowchart and/or block diagram to be implemented. The program code may execute entirely on the machine, partly on the machine, as a stand-alone software package, partly on the machine and partly on a remote machine or entirely on the remote machine or server.

In the context of this disclosure, a machine-readable medium may be a tangible medium that can contain, or store a program for use by or in connection with an instruction execution system, apparatus, or device. The machine-readable medium may be a machine-readable signal medium or a machine-readable storage medium. The machine-readable medium may include, but is not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any suitable combination of the foregoing. More specific examples of a machine-readable storage medium would include an electrical connection based on one or more wires, a portable computer diskette, a hard disk, a Random Access Memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

As used in this disclosure, the terms "machine-readable medium" and "computer-readable medium" refer to any computer program product, apparatus, and/or device (e.g., magnetic discs, optical disks, memory, programmable Logic Devices (PLDs)) used to provide machine instructions and/or data to a programmable processor, including a machine-readable medium that receives machine instructions as a machine-readable signal. The term "machine-readable signal" refers to any signal used to provide machine instructions and/or data to a programmable processor.

To provide for interaction with a user, the systems and techniques described here can be implemented on a computer having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to a user; and a keyboard and pointing device (e.g., a mouse or trackball) by which a user can provide input to the computer. Other kinds of devices may also be used to provide for interaction with a user; for example, feedback provided to the user may be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user may be received in any form, including acoustic input, speech input, or tactile input.

The systems and techniques described here can be implemented in a computing system that includes a background component (e.g., as a data server), or that includes a middleware component (e.g., an application server), or that includes a front-end component (e.g., a user computer having a graphical user interface or a web browser through which a user can interact with an implementation of the systems and techniques described here), or any combination of such background, middleware, or front-end components. The components of the system can be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include: local Area Networks (LANs), wide Area Networks (WANs), and the internet.

The computer system may include a client and a server. The client and server are typically remote from each other and typically interact through a communication network. The relationship of client and server arises by virtue of computer programs running on the respective computers and having a client-server relationship to each other.

The foregoing is merely a specific embodiment of the disclosure to enable one skilled in the art to understand or practice the disclosure. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the generic principles defined herein may be applied to other embodiments without departing from the spirit or scope of the disclosure. Thus, the present disclosure is not intended to be limited to the embodiments shown and described herein but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A method of topic modification comprising:

acquiring a theme page image;

identifying a printed text in the topic page image;

2. The method of claim 1, wherein the retrieving, based on the standard page image, a target reference answer corresponding to the target question in a preset second database comprises:

detecting a topic area of the target topic in the topic page image;

acquiring a second database associated with the standard page image;

retrieving a reference answer to the standard question in the second database;

3. The method according to claim 1 or 2, wherein said modifying the answer content corresponding to the target question according to the target reference answer comprises:

obtaining answer types of answer contents in the target questions;

and correcting the answer content according to the answer type.

4. A method according to claim 3, wherein said modifying said answer content according to said answer type comprises:

correcting the answer content according to the image similarity,

wherein, the correcting the answer content according to the answer type further comprises:

and correcting the answer content according to the text matching degree.

5. A method according to claim 3, wherein said modifying said answer content according to said answer type comprises:

6. The method of claim 5, wherein the question model comprises: an embedding module, an encoding module and a decoding module,

and correcting the answer content according to the target reference answer and the marked keywords thereof through a preset judgment model, wherein the correction comprises the following steps:

7. A method according to claim 3, wherein said modifying said answer content according to said answer type comprises:

8. A device for modifying a subject, comprising:

the image acquisition module is used for acquiring a question page image;

9. An electronic device, the electronic device comprising:

a processor;

a memory for storing the processor-executable instructions;

the processor is configured to read the executable instructions from the memory and execute the instructions to implement the method of any of the preceding claims 1-7.

10. A non-transitory computer readable storage medium storing computer instructions which, when executed on a terminal device, cause the terminal device to implement the method of any of claims 1-7.