KR102106462B1 - Method for filtering similar problem based on weight - Google Patents

Abstract

According to an embodiment of the present invention, a weight-based similar problem filtering method includes receiving an image from a learner terminal, performing OCR processing on the received image, and as the OCR processing is performed, the image Identifying a search request problem from and separating characters included in the identified search request problem into component units, and extracting keywords in a formula included in the identified search request problem and in the component unit Based on the separated characters and the extracted keyword, to filter the pre-stored search target problem and to determine the similarity to the filtered search target problem, the components of the filtered search target problem are separated into categories. Character separated by step and the component unit and search target separated by the category Determining a first similarity of the character portion of the component of the problem and determining a second similarity of the formula portion of the extracted keyword and the component of the search target problem separated by the category, and the first similarity and the Determining a final similarity by assigning a predetermined weight to each of the second similarities and extracting a problem having a similarity level greater than or equal to a preset similarity criterion based on the determined final similarity level, and extracting similarities greater than or equal to the preset similarity criterion As the problem having is extracted, it may include the step of transmitting commentary information corresponding to the extracted problem to the terminal of the learner.

Description

Weight-based similarity filtering method {METHOD FOR FILTERING SIMILAR PROBLEM BASED ON WEIGHT}

The present invention relates to a method for filtering similar problems based on weight. More specifically, it relates to a weight-based similarity problem filtering method using OCR processing results for a question image.

When a learner registers a question on the Q & A bulletin board on the web, an online service is provided in which the answerer registers an answer to the question and the registered answer is provided to the learner again. In order to provide answers to questions registered in such an online service, keyword-based classification and search techniques of texts composing a question are generally used. Such a keyword-based classification and search technique has an advantage of providing a high number of answers and providing an accurate answer by allowing a registered question to be exposed to a large number of respondents, especially in the case of a question composed of text-based general-purpose keywords.

On the other hand, the keyword-based classification and search technique is not effective in the case of a problem in which it is difficult for the learner to clearly express the content of the question in text, such as a problem in which the question includes a mathematical expression or an image. This is because, in the case of a problem including a mathematical expression or an image in addition to text, the registration of the question is not easy, and keyword-based classification and search are difficult. As a result, since the question is not searched to the answerer by keyword search, a problem arises in that the learner cannot be provided with the answer.

Nevertheless, there is no method to easily register, effectively classify and search questions for problems involving equations or images. In particular, a learning service capable of immediately providing matched commentary information by analyzing a question in the form of an image is not provided.

Korean Patent Publication No. 2012-0131480

The technical problem to be solved by the present invention is to provide a method and apparatus for analyzing a question in the form of an image and providing commentary information searched for based on the analysis result.

Specifically, the technical problem to be solved by the present invention is to perform an optical character reading (Optical Character Recognition, hereinafter, OCR) process on an image provided from a learner terminal to search for the same or similar problem as a question included in the image. Is to provide a way. In addition, a method of extracting commentary information matched to the searched problem and providing it to a learner terminal is provided.

Another technical problem to be solved by the present invention is to provide a method and apparatus for constructing a problem and corresponding commentary information as a database and providing a learning service using the constructed database.

Specifically, another technical problem to be solved by the present invention is to provide a method for storing a question as a new problem when a problem matching the learner's question does not exist in the database. In addition, it provides a method for receiving commentary information on a new problem from the answerer's terminal and transmitting it to the learner's terminal.

Another technical problem to be solved by the present invention is to provide a method and apparatus for providing a Q & A messaging interface between a learner terminal and an answerer terminal.

Another technical problem to be solved by the present invention is to provide a method and apparatus for providing a learning service based on a learning service access location of a learner terminal.

The technical problems of the present invention are not limited to the technical problems mentioned above, and other technical problems not mentioned will be clearly understood by those skilled in the art from the following description.

The method for providing a learning service for solving the technical problem includes receiving an image from a learner terminal, performing OCR processing on the received image, and performing a search request from the image as the OCR processing is performed. In the step of identifying and searching for a problem having a similarity of the search request problem and a predetermined similarity criterion or higher, among the previously stored search target problems, and as a result of the search, when the problem having the similarity is extracted, to the extracted problem And transmitting corresponding commentary information to the learner's terminal.

The apparatus for providing a learning service for solving the technical problems includes: a network interface communicating with at least one processor and a learner terminal, a memory for loading a computer program performed by the processor, the computer program, a search target problem, and the And storage for storing commentary information matched to a search target problem, wherein the computer program is operated to receive an image from the learner terminal and to perform OCR processing on the received image and the OCR processing is performed. Accordingly, an operation of identifying a search request problem from the image and an operation of searching for a problem having a similarity of the search request problem and a predetermined similarity criterion or higher among the search target problem and the result of the search, the preset similarity criterion is If the problem with the extracted degree of similarity, and may include the operation to be transmitted to the learner terminal the description information corresponding to the extracted problem.

According to an embodiment of the present invention, the learner has the advantage of being able to easily register a question in the form of an image file and receive an answer to the registered question. In particular, according to an embodiment of the present invention, since the learning service providing apparatus can identify a learner's question including not only text, but also a mathematical expression, a figure, a picture, and a photograph, the learner asks a question about a problem of various subjects And commentary information can be provided.

In addition, according to another embodiment of the present invention, since a messaging interface between a learner terminal and an answerer terminal is provided, there is an effect that a learner can ask a question in real time and receive an answer to it.

In addition, according to another embodiment of the present invention, the learner can be recommended various problems extracted from the learning service providing apparatus based on the questioned problem as well as the questioned problem. Accordingly, a learner's learning experience can be maximized.

In addition, according to another embodiment of the present invention, since a differential learning service is provided according to the learning service access location of the learner terminal, the learner's learning immersion at a particular access location increases.

1 is a block diagram of a system for providing a learning service according to an embodiment of the present invention.
2 is a block diagram of an apparatus for providing a learning service according to another embodiment of the present invention.
3 is a flowchart of a method for providing a learning service according to another embodiment of the present invention.
4 is a flow chart for explaining a problem search method of a learning service providing apparatus according to another embodiment of the present invention.
5 is an exemplary diagram for explaining problems and commentary information stored in a learning service providing apparatus, referred to in some embodiments of the present invention.
6 and 7 are exemplary diagrams for explaining the components and weights in question, which are referred to in some embodiments of the present invention.
8 is a flowchart of a method for providing commentary information to a learner terminal according to another embodiment of the present invention.
9 is an illustration of a messaging interface between a learner terminal and an answerer terminal, referenced in some embodiments of the invention.
10 is an exemplary diagram for explaining a recommendation problem, which is referred to in some embodiments of the present invention.
11 is an illustration of learner question-based learning content, referenced in some embodiments of the present invention.
12 is a block diagram of a system for providing a learner-based learning service according to another embodiment of the present invention.
13 is a flowchart of a method for providing a learner-based learning service according to another embodiment of the present invention.

Hereinafter, preferred embodiments of the present invention will be described in detail with reference to the accompanying drawings. Advantages and features of the present invention, and methods for achieving them will be clarified with reference to embodiments described below in detail together with the accompanying drawings. However, the present invention is not limited to the embodiments disclosed below, but may be implemented in various different forms, and only the embodiments allow the publication of the present invention to be complete, and general knowledge in the technical field to which the present invention pertains. It is provided to fully inform the holder of the scope of the invention, and the invention is only defined by the scope of the claims. The same reference numerals refer to the same components throughout the specification.

Unless otherwise defined, all terms (including technical and scientific terms) used in the present specification may be used as meanings commonly understood by those skilled in the art to which the present invention pertains. In addition, terms defined in the commonly used dictionary are not ideally or excessively interpreted unless specifically defined. The terminology used herein is for describing the embodiments and is not intended to limit the present invention. In this specification, the singular form also includes the plural form unless otherwise specified in the phrase.

In the present specification, the problem is designed to measure a learner's learning achievement and increase a learning level, and may be a work including at least one of letters, figures, graphs, pictures, photos, and equations. In addition, in the present specification, the problem may be documented and printed on paper, but the embodiment of the present invention is not limited thereto, and the problem may mean data that can be stored in an individual file format.

The method for providing a learning service according to an embodiment of the present invention may include a weight-based similarity problem filtering method as an embodiment. The apparatus for providing a learning service according to an embodiment of the present invention may include a weight-based similar problem recommendation apparatus as an embodiment.

Accordingly, in this specification, the weight-based similar problem filtering method and the weight-based similar problem filtering apparatus according to an embodiment of the present invention may be referred to as a learning service providing method and a learning service providing device, respectively.

1 is a block diagram of a system for providing a learning service according to an embodiment of the present invention.

Referring to FIG. 1, the learning service providing system may include a learner terminal 50, a learning service providing device 100, and an answerer terminal 200. The illustrated learner terminal 50, the learning service providing device 100, and the responder terminal 200 are computing devices that can communicate with each other through a network. In particular, the problem 10 photographed by the learner terminal 50 in FIG. 1 is illustrated as an example in which a character and a formula are included.

The learner terminal 50 may acquire an image of the problem 10 and transmit it to the learning service providing apparatus 100. In FIG. 1, as a method in which the learner terminal 50 acquires an image for the problem 10, a case in which the learner terminal 50 photographs the problem 10 through a camera provided is illustrated as an example, The embodiment is not limited to this. As another example, the learner terminal 50 downloads an image including the problem 10 from a web page on the Internet, or generates a problem by generating content input from the learner as a problem through an input unit (not shown). ).

According to an embodiment of the present invention, the learner terminal 50 may perform a preprocessing operation on the acquired image before transmitting the image of the problem 10 to the learning service providing apparatus 100. have. As an example of a pre-processing operation performed by the learner terminal 50, the learner terminal 50 may automatically correct the direction of the image in which the problem 10 is photographed. For example, depending on the photographing angle of the problem 10, such as when the learner terminal 50 photographs the problem 10 upside down, the problem 10 may be included in the image while being tilted or upside down. At this time, the learner terminal 50 may correct the direction of the photographed image. Specifically, the learner terminal 50 may correct and store the image by rotating the direction of the image so that the direction of the problem 10 coincides with the direction displayed by the learner terminal 50.

As another example of the pre-processing operation for the image, the learner terminal 50 may identify only the problem 10 portion on the image and store only the problem 10 portion. That is, the image photographed through the learner terminal 50 may include an area of the problem 10 and a background area, and the learner terminal 50 identifies only the area of the problem 10 on the image, and the problem Areas other than (10) may be treated as noise and filtered. This may be transmitted to the learning service providing apparatus 100.

As another example, the learner terminal 50 identifies the location of the problem 10 on the image, and when the problem 10 is photographed to be biased to a specific location on the image, the problem 10 is placed in the middle position of the image The image may be corrected so that the image may be corrected in consideration of the size of the problem 10 on the image.

In order to perform the above-described preprocessing task, the learner terminal 50 may store in advance a preprocessing program for the acquired image.

According to another embodiment of the present invention, the above-described pre-processing operation may be performed by the learning service providing apparatus 100 after the image acquired by the learner terminal 50 is transmitted to the learning service providing apparatus 100. have.

The learning service providing apparatus 100 may perform OCR processing on an image including the problem 10 received from the learner terminal 50. According to an embodiment of the present invention, as OCR processing is performed, the learning service providing apparatus 100 may identify the problem 10 on the image. For example, when the problem 10 is a math problem including a formula, the learning service providing apparatus 100 identifies characters and formulas constituting the problem 10 included in the image through OCR processing on the image. Then, the image of the area in which the problem 10 is displayed can be converted into identified characters and equations. The learning service providing apparatus 100 may also build a database of problems by storing the converted characters and formulas in one problem unit.

According to another embodiment of the present invention, when the learning service providing apparatus 100 receives an image for a problem from the learner terminal 50, it searches for a problem that is the same as or similar to a problem on the image received in the built database. You can. To this end, the learning service providing apparatus 100 may be received from the learner terminal 100, and before the image including the problem is OCR processed, at least one problem may be stored in advance. Specifically, the learning service providing apparatus 100 may collect and store the at least one problem by receiving from the terminal of another learner or crawling on the web. Alternatively, the learning service providing apparatus 100 may receive and store the at least one problem directly from an administrator of the learning service providing apparatus 100.

According to another embodiment of the present invention, the learning service providing apparatus 100 may store not only the at least one problem, but also commentary information about the problem in advance. The learning service providing apparatus 100 may store problem and commentary information, respectively, and may store matching relationships for both. To this end, the database built on the learning service providing apparatus 100 may be a relational database, but the embodiment of the present invention is not limited thereto, and the database built on the learning service providing apparatus 100 is a graph database (Graph Database). In this case, each problem to be stored and commentary information therefor are composed of a node, and a similarity relationship or recommendation relationship between each question and a matching relationship between each question and commentary information may be configured as an edge. .

Meanwhile, commentary information about the problem may be received from the responder terminal 200.

The responder terminal 200 may receive a problem from the learning service providing apparatus 100 and receive commentary information about the problem from the answerer. The responder terminal 200 may transmit commentary information input from the responder to the learning service providing apparatus 100.

According to another embodiment of the present invention, the responder terminal 200 receives a question about the problem from the learner terminal 50 through the Q & A messaging interface provided by the learning service providing apparatus 100, and therefor Commentary information may be generated based on the answer of the respondent, and provided to the learner terminal 50.

The learner terminal 50 may be any one of a camera equipped with a smart phone, a tablet PC, and a communication module, but is not limited thereto, and the desktop PC, a laptop PC, a PDA VR (Virtual Reality) imaging device, AR ( Augmented Reality). The responder terminal 200 may also be any one of the computing devices illustrated for the learner terminal 50.

The learning service providing device 100 may be either a mobile or fixed computing device. In particular, according to an embodiment of the present invention, the problem is identified by analyzing the image received from the learner terminal 50, and in response, the commentary information collected from the answerer terminal 200 is retrieved, and the search result is obtained from the learner terminal ( 50).

In addition to the above-described configuration, the learning service providing system of FIG. 1 may further include a relay server that relays data communication between components, one or more other learner terminals, and one or more other answerer terminals.

Next, a function and operation of the learning service providing apparatus 100 will be described in more detail with reference to FIG. 2. 2 is a block diagram of an apparatus 100 for providing a learning service according to another embodiment of the present invention.

The learning service providing apparatus 100 loads a computer program performed by the processor 101, the network interface 102 communicating with the one or more processors 101, the learner terminal 50, and the responder terminal 200. It may include a memory 103 and a storage 104 for storing computer programs.

The processor 101 controls the overall operation of each component of the learning service providing apparatus 100. The processor 101 is a CPU (Central Processing Unit), MPU (Micro Processor Unit), MCU (Micro Controller Unit), AP (Application Processor), AP (Application Processor) or any form well known in the art of the present invention It can be configured to include a processor. Further, the processor 101 may perform operations on at least one application or program for executing the method according to embodiments of the present invention. The learning service providing apparatus 100 may include one or more processors.

The network interface 102 supports wired and wireless Internet communication of the learning service providing device 100. In addition, the network interface 102 may support various communication methods in addition to the Internet, which is a public communication network. To this end, the network interface 102 may comprise a communication module well known in the art.

In particular, the network interface 102 provides various user interfaces (UIs) to the learner terminal 50 and the responder terminal 200 in the process of providing the learning service of the learning service providing apparatus 100 according to an embodiment of the present invention. You can.

The memory 103 stores various data, commands and / or information. Memory 103 may load one or more programs 105 from storage 104 to execute methods in accordance with embodiments of the present invention. The memory 103 can be composed of volatile memory that can be read and written and has a high read or write speed. As one example, any one of RAM, DRAM, or SRAM may be provided in the memory 103.

The storage 104 may store the one or more programs 105 and Q & A data 107 non-temporarily. In FIG. 2, as an example of the one or more programs 105, a learning service providing S / W (hereinafter, learning service providing software) 105 is illustrated.

The storage 104 is a non-volatile memory such as a read only memory (ROM), an erasable programmable ROM (EPROM), an electrically erasable programmable ROM (EPMROM), a flash memory, a hard disk, a removable disk, or a technical field to which the present invention pertains. And any known form of computer-readable recording media.

According to another embodiment of the present invention, the learning service providing software 105 may be installed not only in the learning service providing apparatus 100, but also in the learner terminal 50 and the responder terminal 200. That is, the learning service providing software 105 may be an agent that performs various functions using data transmitted and received between each installed device.

The Q & A data 107 is based on a problem received by the learning service providing apparatus 100 or a problem crawled on the web by the learning service providing apparatus 100 by the learner terminal 50 shooting or directly generated by the learner's input. Data. In addition, the Q & A data 107 may include commentary information received by the learning service providing apparatus 100 crawled on the web or directly input or commentary information received from the responder terminal 200. The Q & A data 107 may be stored in the form of a database illustrated in the description of FIG. 1.

In addition, although not shown, the learning service providing apparatus 100 may include an output unit for outputting a UI according to an embodiment of the present invention, for example, a display unit, and input for directly receiving problems and commentary information It may further include wealth.

Hereinafter, functions and operations of the apparatus 100 for providing a learning service according to an embodiment of the present invention will be described in more detail based on the description of FIGS. 1 and 2.

3 is a flowchart of a method for providing a learning service according to another embodiment of the present invention. Each of the following steps is performed by the learning service providing apparatus 100, and in particular, each of the following steps is performed as the processor 101 performs an operation according to the learning service providing software 105.

Referring to FIG. 3, the apparatus 100 for providing a learning service may receive an image of a problem (S10). As described above, the image for the problem may be an image obtained from the learner terminal 50, or may be received by the learning service providing apparatus 100 by performing web crawling.

The learning service providing apparatus 100 may perform OCR processing on the received image (S20). At this time, as the algorithm used for the OCR processing, an algorithm widely known in the technical field to which the present invention belongs may be used, but according to an embodiment of the present invention, a machine learning algorithm is applied as an algorithm for the OCR processing, which is included in the problem. It can improve the recognition accuracy of the characters and formulas. For example, as a machine learning algorithm, a convolutional neural network (CNN) model may be used.

Hereinafter, steps of performing OCR processing of the apparatus 100 for providing a learning service using the convolutional neural network model according to an embodiment of the present invention will be described.

In the general OCR processing method, it is necessary to identify a type of characters on an image, and then segmentation (recognition of characters into component units such as characters) is essential.

On the other hand, according to an embodiment of the present invention, the learning service providing apparatus 100 may not perform a separate recognition process for character type recognition and component units of characters by using a convolutional neural network model. Specifically, the learning service providing apparatus 100 may extract characteristic information for each pixel by analyzing all pixels on the image without separating recognition from the received image without prior recognition of the type of the character. Next, the learning service providing apparatus 100 may predict the type of the character on the image from the feature information of the extracted pixel, based on information previously learned through the convolutional neural network model. At this time, the prediction may be applied to a stochastic prediction technique, and the more the amount of learning through the convolutional neural network model is, the higher the accuracy of the stochastic prediction.

 In particular, according to an embodiment of the present invention, the learning service providing apparatus 100 may recognize not only letters but also expressions by using the learning result through the convolutional neural network model. The formula may include numbers and arithmetic symbols. Specifically, the learning service providing apparatus 100 may predict a character or a formula constituted by each pixel from the extracted feature information based on previously learned information. The learning service providing apparatus 100 may cluster a plurality of pixels constituting any one of predicted characters or equations, and identify at least one of letters and equations based on the clustered pixels.

The learning service providing apparatus 100 may recognize a problem on an image by recognizing each character even when a problem on the image is composed of only letters, for example, a Korean problem, an English problem, or a problem in which Korean and English are mixed. Can be identified. Also, the learning service providing apparatus 100 may identify a problem on the image by recognizing the character and the formula, respectively, when the problem on the image is a problem composed of letters and formulas, for example, a math problem.

Next, the learning service providing apparatus 100 may search for a pre-stored problem based on the problem identified as the result of the OCR processing (S30). As described above, the learning service providing apparatus 100 may previously store Q & A data 107 constructed as a database built in the storage 104. The learning service providing apparatus 100 may search for the Q & A data 107 stored in advance, and determine whether a problem identical or similar to the problem identified as the result of the OCR processing is extracted (S40). In addition, in step S40, the learning service providing apparatus 100 may extract the same or similar problems as the identified problems as a result of the search of the Q & A data 107. At this time, the learning service providing apparatus 100 may determine the problem having the similarity of the identified problem and a preset criterion as the same or similar problem. The similarity determination method will be described later in the description of FIG. 4.

As a result of the determination in step S40, when the same or similar problem is extracted, the learning service providing apparatus 100 may extract commentary information about the extracted problem and transmit it to the learner terminal 50 (S50). . To this end, the learning service providing apparatus 100 may match and store a problem and commentary information about the problem on the Q & A data 107.

As a result of the determination in step S40, if the same or similar problem is not extracted, the learning service providing apparatus 100 displays the problem identified as the result of the OCR processing performed in step S20 as Q & A data 107 It can be stored (S60). That is, if the same or similar problem is not extracted in step S40, the identified problem is treated as a new problem that does not exist in the Q & A data 107, and the Q & A data 107 is stored in the storage 104. Is stored in. According to another embodiment, the learning service providing apparatus 100 may store a received image and a problem identified as a result of OCR processing thereof.

According to another embodiment, the learning service providing apparatus 100 may generate metadata about the identified problem when storing the identified problem. In addition, the learning service providing apparatus 100 may perform an indexing operation for improving the search speed when searching for the identified problem later. For example, indexing can be performed based on special keywords or symbols that make up the identified problem. That is, when a mathematical symbol or keyword serving as a standard for classifying the identified problem is identified, the learning service providing apparatus 100 may automatically allocate a preset key value to the identified problem.

According to another embodiment of the present invention, the learning service providing apparatus 100 may determine whether to store the identified problem based on a preset quality criterion in step S60.

For example, the learning service providing apparatus 100 may identify the degree of noise inclusion or resolution of the received image, and if the quality is equal to or higher than a preset quality criterion, the identified problem may be stored.

As another example, the learning service providing apparatus 100 identifies, when the quality of Hangul, equations, figures, graphs, images, etc., which are components included in the identified problem, is greater than or equal to a preset quality standard as a result of OCR processing of the received image You can also save the problem. At this time, whether or not to store the identified problem may be determined policy by the administrator of the learning service providing apparatus 100. That is, when a preset essential component exists in the learning service providing apparatus 100 by the administrator, the learning service providing apparatus 100 identifies when the OCR processing result for the essential component is higher than a preset quality standard You can also save the problem.

As another example, the learning service providing apparatus 100 may determine whether or not the quality is higher than a preset quality standard for each component of the identified problem as a result of OCR processing. Accordingly, the learning service providing apparatus 100 according to an embodiment of the present invention determines the quality score for each component of the identified problem as a result of OCR processing, and identifies the problem only when the sum total is greater than or equal to the preset score You can also save

In the above-described embodiment, the preset quality criterion may be determined by the manager of the learning service providing apparatus 100.

According to the above embodiment, the learning service providing apparatus 100 may not be stored in the storage 104 when the quality of the received image is low or the quality of the OCR processing result is low. Accordingly, it is possible to save space in the storage 104 of the learning service providing apparatus 100, and there is an advantage of minimizing unnecessary search targets.

4 is a flowchart illustrating a problem search method of the apparatus 100 for providing a learning service according to another embodiment of the present invention. In particular, the problem search method of the learning service providing apparatus 100, which is performed between the step of searching for a problem based on the OCR processing result described in the description of FIG. 3 (S30) and the problem extraction step of similarity criteria (S40). 4 will be described with reference to each step of FIG.

Referring to FIG. 4, the learning service providing apparatus 100 may separate the identified characters into component units as a result of OCR processing (S31). For example, if the identified character is Hangul, the learning service providing apparatus 100 may identify it, and separate the Hangul identified by consonants and vowels that are components of the character.

In addition, the learning service providing apparatus 100 may extract keywords in the identified formula as a result of the OCR process (S32). For example, when the identified formula includes a specific formula symbol such as an integral symbol, the learning service providing apparatus 100 may identify it.

Prior to steps S31 and S32, the learning service providing apparatus 100 may separately identify the components of the identified text in units of respective components such as text, equations, images, figures, and graphs.

The learning service providing apparatus 100 may filter a problem stored in advance in the learning service providing apparatus 100 based on the components of the separated text and the extracted keyword (S33). To this end, the learning service providing apparatus 100 may store Q & A data 107 in the storage 104 as described above in the description of FIG. 2. In addition, the Q & A data 107 may include the same or similar problems as at least some of the separated character components and the extracted keywords.

Hereinafter, for convenience of description, a problem identified by the learning service providing apparatus 100 is referred to as a search request problem, and among the problems previously included in the Q & A data 107, a problem identical or similar to at least a part of the above Let's call it a search target problem. The filtering step S33 will be described in detail.

Specifically, the learning service providing apparatus 100 may separate characters, which are included in the search request problem, for each component, and extract keywords in formulas. In addition, the learning service providing apparatus 100 may perform hashing on the search request problem based on the components of the separated characters and keywords in the extracted formula. When the hashing of the search request problem is performed, the learning service providing apparatus 100 may filter the search target problem indexed and stored in advance based on the hashing. That is, the learning service providing apparatus 100 filters at least some of the same or similar problems among the components included in the search request problem and the extracted keywords in the formula among the problems previously included in the Q & A data 107. can do. At this time, at least one search target problem may be filtered.

Next, the learning service providing apparatus 100 may separate each filtered search target problem into categories of components (S34). For example, the learning service providing apparatus 100 may be divided into a character part and a formula part among the components of the search target problem.

The learning service providing apparatus 100 may assign a similarity weight to the separated character parts (S35). To this end, the learning service providing apparatus 100 may measure the similarity of the character parts separated in step S34 based on the component of the character of the search request problem in step S31. Specifically, the learning service providing apparatus 100 includes each component of the separated characters included in the search request problem, the distance between each component, the arrangement order of each component, and words combined with each component, all Based on at least one of the lengths of the text area, the text part of the problem to be searched can be analyzed, and similarity weights can be assigned to the analyzed text part.

The case where the character part of the search request problem is "Please obtain the x value" and the character part of the problem to be searched is "x value" is described as an example. The learning service providing apparatus 100 may identify the English letter "x" of the search request problem and the Korean "value", "e", "nine", "bottom", "hour", and "o". At this time, the learning service providing apparatus 100 analyzes the character parts of the problem to be searched, based on the existence of the English letter "x" of the search request problem and the "value" in Korean, and the order of the two, and the search object You can weight the problem. In the above example, in the case of Hangul, the learning service providing apparatus 100 may also identify components of each word, and based on this, may analyze the character part of the search target problem.

The learning service providing apparatus 100 may assign weights to each of the problems filtered in the filtering step S33, and determine the similarity of each filtered search target problem using the assigned weights ( S36). The learning service providing apparatus 100 may determine whether a problem equal to or greater than a preset similarity criterion is extracted based on the similarity (S40). That is, the learning service providing apparatus 100 may determine whether there is a problem with a similarity criterion or higher based on the similarity determined in step S36. Referring to FIGS. 3 and 4, when a problem with a similarity criterion or higher exists, the learning service providing apparatus 100 may extract it and transmit commentary information about the extracted problem to the learner terminal 50.

5 is an exemplary diagram for explaining problems and commentary information stored in the apparatus 100 for providing a learning service, referred to in some embodiments of the present invention. Hereinafter, the relationship between the problem and the commentary information will be described in detail with reference to FIG. 5. In FIG. 5, the Q & A data 107 is shown as an example, and in particular, a case in which the problem and commentary information about it are matched in advance is shown as an example.

The Q & A data 107 includes a problem 10, a problem 11, and a problem 12, and may include commentary information 500, commentary information 501, and commentary information 502 matched to each. You can. 3 and 4, the learning service providing apparatus 100 may identify the received image as a search request problem as a result of OCR processing on the received image. At this time, in the learning service providing apparatus 100 step S36, on the Q & A data 107, a search request problem and a search target problem having a similarity level higher than a preset similarity criterion may be extracted. For example, when the problem 10 in FIG. 5 is an extracted problem, the learning service providing apparatus 100 may extract the commentary information 501 and transmit it to the learner terminal 50.

For example, the learning service providing apparatus 100 may transmit commentary information of a problem having the highest similarity to the learner terminal 50 when there are a plurality of search target problems having similarities of more than a predetermined similarity criterion.

As another example, the learning service providing apparatus 100, if the problem to be searched having a similarity level equal to or higher than a preset similarity criterion is a problem 10 or a problem 11, both the commentary information 501 and the commentary information 502 are learners It can also be transmitted to the terminal 50.

In the above examples, the learning service providing apparatus 100 may transmit not only the commentary information to the learner terminal 50, but also the extracted problem. This is because the search request problem identified as the result of the OCR processing of the received image and the extracted problem are similar, but may not be the same. In this case, the learner terminal 50 may receive a problem similar to that of the first image transmitted, but a similar problem and commentary information. Accordingly, the learner can indirectly learn the problem to be searched by learning the received problem and commentary information.

Next, with reference to FIGS. 6 and 7, a method of searching for a problem in the learning service providing apparatus 100 referred to in the description of FIG. 4 will be described in more detail.

6 and 7 are exemplary diagrams for explaining the components and weights in question, which are referred to in some embodiments of the present invention. In particular, in FIG. 6, it is assumed that the Q & A data 107 includes a problem 10, a problem 11, and a problem 12, and the problem 10 is a search request problem.

Referring to FIG. 6, the apparatus 100 for providing a learning service may separately identify the problem 10 in units of components. Accordingly, the learning service providing apparatus 100 may separately identify the equation 610 of the problem 10 and the character 620. 4 and 6, the learning service providing apparatus 100 filters the problem 11 and the problem 12 among the Q & A data 107 (S33), the problem 11, and the problem 12 ) Can be separated into the formula 611 and the character 621 and the formula 612 and the character 622, respectively.

The learning service providing apparatus 100 may determine the similarity between the formula 610 of the search request problem 10 and the formula 611 and the formula 612 of the filtered problems 11 and 12. To this end, the learning service providing apparatus 100 may extract keywords of the formula 610 and analyze the formula 611 and the formula 621 based on the keywords. As a result of the analysis, the learning service providing apparatus 100 may assign similarity weights to the equations 611 and 621. The weight given to the equation will be referred to as the first weight.

The learning service providing apparatus 100 may determine the similarity between the characters 620 of the search request problem 10 and the characters 621 and 622 of the filtered problems 11 and 12. To this end, the learning service providing apparatus 100 separates the characters 620 into component units, the components of the separated characters, the distances between the components, words in which the components are combined, the arrangement order of the components, and the whole characters Character 621 and character 622 may be analyzed based on at least one of the lengths of the regions. As a result of the analysis, the learning service providing apparatus 100 may assign similarity weights to the characters 621 and 622. The weight assigned to the characters will be referred to as a second weight hereinafter.

The learning service providing apparatus 100 may determine the similarity between the problems 11 and 12 based on the first weight and the second weight.

In FIG. 7, a problem 710 in which the graph 630 is further included in the problem 10 in FIG. 6 is illustrated as a search request problem. Referring to FIG. 7, the learning service providing apparatus 100 may separately identify the search request problem 710 by using the formula 610, the text 620, and the graph 630. In FIG. 7, in particular, the case where the problem 710 includes the graph 630 is illustrated, but the embodiment of the present invention is not limited thereto, and the problem 710 may include a figure or an image. In addition, it is assumed that at least some of the equations 611, characters 720, and graphs 730 may be included in one filtered problem, and may be included in different problems, respectively.

4 and 7, the learning service providing apparatus 100 is included in the Q & A data 107 based on the formula 610, the character 620, and the graph 630 of the problem 710, respectively. Filtered problems can be analyzed.

For example, the learning service providing apparatus 100 may include a graph 630 of the formula 610 and the formula 611 of the problem 710, the characters 620 and 720 of the problem 710, and the problem 710. ) And the graph 730, respectively, and similarity weights can be assigned to the equations 611, the characters 720, and the graphs 730, respectively.

At this time, the learning service providing apparatus 100 is based on the importance of the first weight, which is the weight for the formula, the second weight, which is the weight for the characters, and the weight for the graph (hereinafter referred to as the third weight). Similarity can be determined. For example, the importance of weights may be preset in the order of the first weight, the second weight, and the third weight in the learning service providing apparatus 100. That is, weights may be added to the first weight, the second weight, and the third weight again. To this end, a specific factor is multiplied by the first weight, the second weight, and the third weight, so that an inferiority relationship between the first, second, and third weights may be previously determined.

According to another embodiment of the present invention, after the learning service providing apparatus 100 separately identifies the equations 610, characters 620, and graphs 630 that are components of the problem 710, each component is a problem The area occupied at 710 may be analyzed, and the importance level between the first weight, the second weight, and the third weight may be determined. For example, when the graph 630 occupies the largest area, the learning service providing apparatus 100 may determine the third weight as the most important factor in determining the similarity to the filtered problem. Alternatively, when the horizontal length of the character 620 is the longest, the learning service providing apparatus 100 may determine the second weight as the most important factor in determining the similarity to the filtered problem. In other words, the learning service providing apparatus 100 may dynamically determine the importance of the weight of each component based on the result of the OCR processing for the problem 710.

According to another embodiment of the present invention, the learning service providing apparatus 100 may assign an additional weight to a value having the highest similarity weight. For example, the equations 610 and 611 of the problem 710, the characters 620 and 720 of the problem 710, and the graphs 630 and 730 of the problem 710, respectively, are shown. As a result of comparison, the similarity between the equations 610 and 611 is highest, the similarity between the graphs 630 and 730 is the second highest, and the similarity between the letters 620 and 720 is the lowest. In this case, the magnitude of the similarity weight may be determined in the order of the first weight, the third weight, and the second weight. In this case, the learning service providing self-governing apparatus 100 may multiply, for example, only the first weight by one or more preset factors, so that the first weight is reflected most in the similarity determination.

So far, referring to FIGS. 3 to 7, when the search target problem is the same or similar to the search request problem based on the similarity, transmitting commentary information to the learner terminal 50 (S50) or similarity When a problem above the standard has not been extracted, the step of saving the identified search request problem (S60) has been mainly described.

Next, after the step (S60) of FIG. 3, when the learner terminal 50 does not acquire the commentary information about the search request problem, the embodiment performed by the learning service providing apparatus 100 is shown in FIGS. This will be described with reference to FIG. 9.

However, in the embodiment described herein, after step S60 of FIG. 3, the learner terminal 50 does not acquire commentary information about the search request problem, as well as the learning service after step S50 of FIG. 3 It may be performed by the providing device 100. For example, even if the learner terminal 50 acquires commentary information about a search request problem, if the learner requests additional commentary, this embodiment may be performed after step S50. Alternatively, the learner terminal 50 acquires commentary information about a search request problem and similar problems, and the same applies when a learner requests a commentary of a search request problem.

Hereinafter, after the step S60 of FIG. 3, a case in which the learner terminal 50 has not obtained commentary information about the search request problem will be described as an example.

8 is a flowchart of a method for providing commentary information to a learner terminal 50 according to another embodiment of the present invention. Referring to FIG. 8, the learning service providing apparatus 100 may receive a comment request message for a problem from a learner terminal (S61). The learner may operate the learner terminal 50 to transmit a comment request message for the problem to the learning service providing apparatus 100. To this end, the learning service providing apparatus 100 may provide the learner terminal 50 with a user interface for generating and transmitting the comment request message.

The learning service providing apparatus 100 may transmit a comment request message to at least one of a plurality of responder terminals in response to the comment request message received from the learner terminal 50 (S63). The comment request message received from the learner terminal 50 and the comment request message transmitted to at least one terminal may be the same message, but embodiments of the present invention are not limited thereto. For example, the learning service providing apparatus 100 may generate a comment request message for transmission to at least one UE based on the comment request message received from the learner terminal 50. Alternatively, if the problem identified in step S60 is stored as a new problem, the learning service providing apparatus 100 automatically generates a comment request message for the new problem even if a comment request is not received from the learner terminal 50. In addition, the automatically generated commentary request message may be transmitted to at least one of the plurality of responder terminals.

The number of responder terminals 200 shown in FIG. 1 may be multiple, and the plurality of responder terminals may be terminal devices registered in advance in the learning service providing apparatus 100. The learning service providing apparatus 100 may store contact information of a plurality of responder terminals in advance.

The learning service providing apparatus 100 may receive commentary information about a problem from at least one responder terminal (S65). It is assumed that at least one answerer terminal is the answerer terminal 200 of FIG. 1. The responder terminal 200 may receive a comment request message from the learning service providing apparatus 100, and may also receive at least one of an image of a problem and a search request problem.

The responder may input an image of the received problem and commentary information on at least one of a search request problem to the responder terminal 200. At this time, the commentary information may be in the form of an image file. Alternatively, by receiving input from the responder through the commentary information generation interface provided through the answerer terminal 200, the answerer terminal 200 may generate commentary information.

When commentary information is received from the answerer terminal 200, the learning service providing apparatus 100 may transmit the received commentary information to the learner terminal 50 (S67). In addition, the learning service providing apparatus 100 may match and store the received commentary information and a search request problem. In this case, the learning service providing apparatus 100 may determine whether to store the commentary information received according to the above-described quality criteria in the description of step S60 of FIG. 3. The learning service providing apparatus 100 may match and store the received commentary information and the answerer information of the answerer terminal 200.

Next, the learning service providing apparatus 100 may determine whether a request for generating a messaging interface is received from the learner terminal 50 (S68). For example, the learner receives commentary information through the learner terminal 50 in step S67, but if there are additional questions or inquiries about related problems, the learning service providing apparatus through the learner terminal 50 A request for generating a messaging interface may be transmitted to (100).

In step S68, as a result of the determination, the learning service providing apparatus 100 may generate a messaging interface between the learner terminal 50 and the responder terminal 200 when a request for generating a messaging interface is received (S69). In this case, prior to the creation of the messaging interface, the learning service providing apparatus 100 may transmit an invitation message for the messaging interface to the answerer terminal 200. The learning service providing apparatus 100 may generate a messaging interface between the learner terminal 50 and the answerer terminal 200 when a message corresponding to the invitation message is received from the answerer terminal 200. The learning service providing apparatus 100 may control a session between the learner terminal 50 and the answerer terminal 200 to be formed, and provide an instant messaging interface with learners and answerers as participants.

9 is an illustration of a messaging interface between a learner terminal and an answerer terminal, referenced in some embodiments of the invention. In FIG. 9, as an example of a messaging interface, a chat window 900 is shown with an answerer and a learner as participants.

Referring to FIG. 9, the learning service providing apparatus 100 may support a message exchange 910 for problem and commentary information between a learner and an answerer through the chat window 900. To do this, you can do things like message text, symbols and formulas. In addition, the learning service providing apparatus 100 may search for and extract an image of a problem received from the learner terminal 50 or an identified search request problem and a problem that exceeds a predetermined similarity criterion. Problems above a predetermined similarity criterion may include the same problem as the search request problem and similar problems.

The learning service providing apparatus 100 may analyze metadata of the same problem or a similar problem with the search request problem. In addition, the learning service providing apparatus 100 may search for and extract related problems related to the same problem or similar problems extracted based on the analyzed metadata.

For example, the similarity problem may be a problem in which some characters differ from the search request problem or different numbers. That is, when a mathematical problem is described as an example, the principle or formula required to solve the problem is the same, but the given constant value is different or the expression of the variable is different.

In the case of the similarity problem, the similarity is higher than the preset similarity criterion, but it may be a problem that cannot be classified into the identity area with the search request problem. For example, if the similarity preset in the learning service providing apparatus 100 is 90% or more, it is determined to be the same or similar problem to the search request problem, and if more than 95% identity is recognized, it is determined to be the same problem as the search request problem. I assume. At this time, the similarity problem may be a problem in which 90% or more and less than 95% identity are recognized.

For example, the related problem may also be a problem related to a problem, such as a category of the same problem as the unit to be searched, or a problem solving method, if the problem of mathematics is described as an example. That is, when the search request problem is a problem of a set unit of a math unit, a related problem is a problem of calculating an equation using a number obtained by obtaining an intersection of two sets.

In the case of the related problem, the similarity of the search request problem and the preset similarity criterion is low, but the information on the metadata may be related. As illustrated above, it is assumed that the similarity preset in the learning service providing apparatus 100 requires equality of 90% or more. At this time, when the learning service providing apparatus 100 determines that similarity with the search request problem is found and 80% of the identity is recognized, it may be determined that the searched problem is not the same or similar problem to the search request problem.

In one embodiment, the learning service providing apparatus 100 may analyze metadata about the searched problem. Metadata may include, for example, unit information of a problem when the problem is a math problem, keyword information for describing and classifying the problem, and the like. The learning service providing apparatus 100 may analyze metadata and determine a relationship between a searched problem and a search request problem. That is, the learning service providing apparatus 100 determines the relevance to the search request problem through metadata analysis of the searched problem, as well as when the searched problem is the same or similar problem to the search request problem, but not the same or similar problem can do.

To this end, the learning service providing apparatus 100 may be previously set with respect to similarity criteria.

Here, the relevance criterion refers to the similarity in the metadata of the search request problem and the searched problem, not the criteria for determining the similarity between the search request problem and the letters, numbers, and formulas constituting the search problem when the search request problem is a math problem. It may be a criterion for judging. That is, the relevance criterion is a criterion for judging the relationship between a problem and a problem, and can be determined policy.

For example, a search request problem has a quadratic equation as a unit category, a root formula as a solution method, and metadata with level 3 as difficulty information, and a searched problem is a quadratic equation as a unit category, a root formula as a solution method, and It is assumed that metadata having level 3 is used as factorization and difficulty information. In addition, it is assumed that the unit category, the information of the solution method, and the difficulty information are parameters having a numerical value as parameters of respective relevance criteria. In this case, the learning service providing apparatus 100 may compare the search request problem and the parameter values of the searched problem to determine relevance. When the difference between the search request problem and the parameter value of the searched problem is equal to or less than a preset value, the learning service providing apparatus 100 may determine that the searched problem is related to the search request problem. At this time, the learning service providing apparatus 100 may compare a plurality of parameters to determine the relationship between the searched problem and the search request problem.

The learning service providing apparatus 100 may determine a searched problem as a related problem when relevance of a predetermined relevance criterion or higher is recognized.

The learning service providing apparatus 100 may search for and extract the related problem in step S30 of searching for a problem to be searched based on the identified problem of FIG. 3 and step S40 of extracting a problem of similarity criteria. . The learning service providing apparatus 100 may extract a problem identical or similar to a search request problem, and then search and extract a related problem with respect to the extracted problem, but embodiments of the present invention are not limited thereto. That is, the learning service providing apparatus 100 may search for and extract related problems in a process separate from the search and extraction operation of the same or similar problem to the search request problem.

The learning service providing apparatus 100 may provide a related problem to at least one of the learner terminal 50 and the responder terminal 200. For example, when a related problem is provided to the responder terminal 200, the responder may recommend the related problem 920 to the learner through the chat window 900. To this end, the learning service providing apparatus 100 may select the related problem provided and provide an interface for providing to the learner terminal 50 through the chat window 900 to the answerer terminal 200. In addition, the learning service providing apparatus 100 may provide an interface for searching for a related problem in the respondent's terminal 200.

Meanwhile, according to another embodiment of the present invention, a search request is automatically made by the learning service providing apparatus 100, by a request of the learner terminal 50, or by manipulation of the responder terminal 200 of the responder It is not the same as the problem, but similar or related problems can be recommended. As described above, the same or similar problem to the search request problem or a related problem is different from the problem requested by the learner terminal 50, but a problem provided to the learner terminal 50 may be referred to as a recommendation problem.

10 is an exemplary diagram for explaining a recommendation problem, which is referred to in some embodiments of the present invention.

Referring to FIG. 10, the learning service providing apparatus 100 may classify a problem belonging to the Q & A data 107 of FIG. 2 according to at least one of similarity criteria and relevance criteria. Problem group 0, problem group 1, problem group 2, problem group 3, and problem group 10 of FIG. 10 are groups of problems classified according to the relevance criteria, and problems belonging to each problem group are similar problems grouped according to the similarity criteria to be.

Problems A1 and B1 are the same problem as the search request problem or the search request problem. In FIG. 10, problem group 0 is a group of problems corresponding to difficulty 1, problem group 1, problem group 2 and problem group 3 are groups of problems corresponding to difficulty 2, and problem group 10 is problems corresponding to difficulty 3 The group case was illustrated. Difficulty level 1, difficulty level 2, and difficulty level 3 can be further classified into detailed difficulty levels, respectively. That is, even a problem belonging to the same difficulty 2 may have a differential difficulty in the horizontal direction. For example, problems A1 and A2 are problems corresponding to difficulty 2, but problem A2 may have a higher level of detail than problem A1.

In addition, problem group 1, problem group 2, and problem group 3 in FIG. 10 may be groups of problems corresponding to difficulty 2 having the same difficulty level, but may have different unit categories in metadata. That is, the vertical height of the problem group in FIG. 10 may indicate a difference between unit categories.

In one embodiment, the learning service providing apparatus 100 may identify the problem group 1 as a result of similarity determination when the problem A1 is searched, that is, when the learner queries the problem A1. The learning service providing apparatus 100 may extract a problem similar to the problem the learner asked from the identified problem group 1 and provide it to the learner terminal 50 as a recommended problem. A problem belonging to the same problem group may be a similar problem having only a difference between a search request problem and a constant, and the learning service providing apparatus 100 may, for example, recommend problem A2 of problem group 1 as a recommendation problem.

In another embodiment, the learning service providing apparatus 100 when the problem A2 is searched, the result of the similarity determination, identifies the problem group 2 as a group of similar problems and extracts the similar problem from the problem group 2 to learner terminal 50 Can be provided as a recommendation issue. In addition, the learning service providing apparatus 100 may extract a problem that is higher than a predetermined relevance criterion based on the problem A2. At this time, the learning service providing apparatus 100 may first identify a problem group having an adjacent difficulty based on the difficulty information in the metadata, and then extract and recommend a problem belonging to the identified problem group. In FIG. 10, problem group 0 and problem group 10 are illustrated as identified problem groups. Problem group 0 has the same unit category as problem B1, but the group for basic learning is low, and problem group 10 has the same unit category as problem B1, but it can be the group that the problem for advanced learning belongs to. have.

The learning service providing apparatus 100 may extract, for example, the problem a2 belonging to the problem group 0 and / or the problem D1 belonging to the problem group 10 as a related problem and provide it to the learner terminal 50 as a recommended problem.

In another embodiment, the learning service providing apparatus 100 may store a learner's question problem history 1000. In FIG. 10, the case where the problem A1 and the problem B1 are stored in the question problem history is illustrated. The apparatus 100 for providing a learning service may identify a problem group having the same difficulty as a question problem to a learner, but having different unit categories, based on the question problem history 1000. That is, the learning service providing apparatus 100 may identify the problem group 3 and provide the learner terminal 100 with the problem C1 belonging to the problem group 3 as a recommended problem, for example.

In addition, the learning service providing apparatus 100 may identify a problem group having a different difficulty from the question problem based on the question problem history 1000. For example, the learning service providing apparatus 100 may identify the problem group 0 and / or the problem group 10, and may recommend a problem belonging to the identified problem group to the learner.

11 is an illustration of learner question-based learning content, referenced in some embodiments of the present invention.

Referring to FIG. 11, the learning service providing apparatus 100 may generate learner customized learning content 1100 based on the question problem history 1000 of FIG. 10. The learning service providing apparatus 100 may determine the learner's learning level based on the question problem history 1000. In addition, the learning service providing apparatus 100 may determine the learning achievement of the learner for a certain period based on the problems added to the question problem history 1000 during a preset period.

The learning service providing apparatus 100 may extract a search request problem, the same problem, a similar problem, and a related problem as a recommendation problem based on the question problem history 1000, and the learning content 1100 consisting of the extracted problem Can generate In FIG. 11, the case where the learning content 1100 includes a search request problem or a problem 1110 determined to be the same problem as a search request problem in the first region and includes a recommendation problem 1120 in the second region Became. The search request problem 1110 and the recommendation problem 1120 may be sequentially arranged in the learning content 1100 based on at least one parameter of metadata, but embodiments of the present invention are not limited thereto. The search request problem or the problem 1110 determined as the same problem and the recommendation problem 1120 may be arranged in the learning content 1100 in any order.

The apparatus 100 for providing a learning service may generate individual learning contents based on parameters on metadata, such as a unit category, a difficulty level, and a solution method of a problem. For example, the learning service providing apparatus 100 may generate learning content composed only of the same unit, learning content composed only of problems of deeper difficulty, and learning content composed only of problems to which a similar solution is repeatedly applied.

The learning service providing apparatus 100 may transmit the generated learning content 1100 to the learner terminal 50.

12 is a block diagram of a system for providing a learner-based learning service according to another embodiment of the present invention. Particularly, as a learning service providing system based on the learner class of FIG. 12, a case in which the learner authentication device 1200 is added to the learning service providing system of FIG. 1 is illustrated as an example.

Referring to FIG. 12, the learner authentication device 1200 may detect that the learner terminal 50 is located in a preset area 1210. When the learner terminal 50 is connected to the learning service provided by the learning service providing apparatus 100, the learner authentication device 1200 may display unique information of the learner terminal 50 connected to the learning service in the area 1210. Can be identified. When the identified unique information is transmitted to the learning service providing apparatus 100, the learning service providing apparatus 100 may determine a learner's learner class as a user of the learner terminal 50. The learning service providing apparatus 100 may provide a learning service corresponding to the determined learner class to the learner terminal 50.

For example, the unique information of the learner terminal 50 may be Mac address information, and the learner authentication device 1200 may include an access point (AP) that provides Wi-Fi Internet in a preset area 1210. Can be The learning service providing apparatus 100 starts accessing the learning service within the area 1210 preset in the learner terminal 50 based on the ip information assigned to the AP and / or the MAC address information of the learner terminal 50. You can judge

The learning service providing apparatus 100 may provide different learning services for each learner class. The learning service providing apparatus 100 may determine a learner class based on at least one of a learning service type, a learning service charge amount, a service access location, and a learning service participation degree that the learner has applied for.

For example, the learning service providing apparatus 100 may provide a basic learning service when a learner having a basic level is connected to the learning service. The basic learning service may be a service that, when an image containing a problem is received from the learner terminal 50, analyzes it and provides commentary information on the identified search request problem. At this time, the service cost may be charged to the learner who is the basic class whenever the commentary information is transmitted to the learner terminal 50 for the search request problem.

As another example, the learning service providing apparatus 100 may provide a learning service for each grade according to each learner's grade when a learner who is higher than the basic grade is connected to the learning service. The learning service for each class may include a basic learning service described above, and may additionally provide a recommendation problem to the learner terminal 50. In addition, the learning service for each grade may include a service for periodically or irregularly providing the learning content 1100 illustrated in FIG. 11 to the learner terminal 50 according to the learner's grade. In this case, for learners of a higher level of learners, commentary information on a search request problem is provided free of charge, and learning content 1100 may be provided for each content for a fee.

As another example, the learning service providing apparatus 100 determines whether the access location of the learner terminal 50 is within a preset area 1210 when a learner of a basic grade or a learner whose learner grade is not determined is connected to the learning service. You can. When the access position of the learner terminal 50 is within the area 1210, the learning service providing apparatus 100 may determine the learner's grade as a higher-grade learner. The learning service providing apparatus 100 may provide a learning service for each class to the learner terminal 50 of the learner determined as the upper class learner.

The content and billing policy of the basic learning service and the learning service by grade are only examples, and embodiments of the present invention are not limited thereto.

13 is a flowchart of a method for providing a learner-based learning service according to another embodiment of the present invention. Hereinafter, a method of providing a learner-based learning service will be described again with reference to FIG. 13. Each step is performed by the learning service providing apparatus 100.

Referring to FIG. 13, the learning service providing apparatus 100 may identify a learning service access location of the learner terminal 50 (S1301). To this end, the learning service providing system may include a learner authentication device 1200 as illustrated in FIG. 12. The learning service providing apparatus 100 may identify the access location of the learner terminal 50 by receiving the unique information of the learner terminal 50 from the learner authentication device 1200.

The learning service providing apparatus 100 may determine whether the access position of the learner terminal 50 is a preset area (S1303). For example, the preset area is a place where a learning service is provided, and may be an educational facility such as a reading room, school, or academy. Place information of the educational facility may be preset in the learning service providing apparatus 100.

As a result of the determination, when the access position of the learner terminal 50 is not a preset area, the learning service providing apparatus 100 determines whether the learner's rating of the learner terminal 50 accessing the learning service is a basic level or a higher level. can do. That is, the learning service providing apparatus 100 may determine whether the learner is the target of the learning service for each grade (S1305).

The learning service providing apparatus 100 may provide a basic learning service to the learner terminal 50 when the learner is determined to be a basic class learner (S1307).

On the other hand, if it is determined in step S1303 that the access point of the learner terminal 50 is within a preset area, or if the learner's grade is determined to be a higher grade in step S1305, the learning service providing apparatus 100 is a learner According to the learner's grade, the terminal 50 may provide a learning service for each grade (S1309).

The methods according to embodiments of the present invention described so far with reference to the accompanying drawings may be performed by execution of a computer program embodied in computer readable code. The computer program may be transmitted from a first computing device to a second computing device through a network such as the Internet and installed in the second computing device, and thus used in the second computing device. The first computing device and the second computing device include a server device, a fixed computing device such as a desktop PC, and a mobile computing device such as a notebook computer, a smart phone, and a tablet PC.

Although the embodiments of the present invention have been described with reference to the accompanying drawings, a person skilled in the art to which the present invention pertains may be implemented in other specific forms without changing the technical concept or essential features of the present invention. You will understand. Therefore, it should be understood that the above-described embodiments are illustrative in all respects and not restrictive.

Claims (5)

As a method performed by the learning service providing device,
Receiving an image from a learner terminal;
Performing OCR processing on the received image;
As the OCR process is performed, identifying a search request problem from the image;
Separating characters included in the identified search request problem into component units;
Extracting keywords in the formula included in the identified search request problem;
Filtering a pre-stored search target problem based on the text separated by the component unit and the extracted keyword;
Separating components of the filtered search target problem into categories to determine similarity to the filtered search target problem;
Determining a first degree of similarity between a character separated by the component unit and a component of a search target problem separated by category;
Determining a second similarity of the formula part among the extracted keywords and components of the search target problem separated by the categories;
Determining a final similarity by assigning predetermined weights to the first similarity and the second similarity;
Extracting a problem having a similarity level higher than a preset similarity criterion based on the determined final similarity level; And
And as the problem having a similarity level equal to or greater than the preset similarity criterion is extracted, transmitting commentary information corresponding to the extracted problem to the learner's terminal,
Weight-based similarity filtering method. The method of claim 1, wherein the search request problem includes characters and formulas as components,
The step of performing the OCR process,
Analyzing pixels on the image based on pre-trained information using a convolutional neural network (CNN) model; And
Identifying, by the analysis, at least one of a character and a formula on the image,
Weight-based similarity filtering method. The method of claim 1, wherein the search request problem includes characters, formulas, and graphs as components,
Extracting a keyword in the formula included in the identified search request problem further includes extracting a graph image included in the identified search request problem,
The determining of the second similarity may include determining a third similarity of the graph portion among the extracted graph images and components of the search target problem separated by the categories,
The step of determining the final similarity,
And determining a final similarity by assigning predetermined weights to the first similarity, the second similarity, and the third similarity, respectively.
Weight-based similarity filtering method. According to claim 1, The step of determining the final similarity by assigning the predetermined weight,
A first weight is assigned to any one of the first and second similarities, and a second weight is assigned to the other of the first and second similarities, which has a low similarity. Including the step of giving,
Weight-based similarity filtering method. According to claim 2,
Analyzing the pixels on the image,
Extracting feature information for each pixel on the image; And
And predicting a character or a formula constituting each pixel based on the previously learned information and the extracted feature information.
The step of identifying at least one of letters and formulas on the image may include:
Clustering a plurality of pixels constituting any one of the predicted character or equation; And
And identifying any one of the characters or expressions based on the clustered pixels.
Weight-based similarity filtering method.

Images