KR102205556B1 - Method and apparatus for creating model for categorizing cultural properties using automatic analysis of formative element - Google Patents

Abstract

The present invention relates to a method and an apparatus for creating a model for categorizing cultural assets by using automatic analysis of figurative elements produced through machine learning. The method comprises the steps of: accessing a database in which cultural asset information is stored; extracting formation information by applying a shape recognition algorithm to the image of a first cultural asset stored in the database; extracting material information by applying a material recognition algorithm to the image of the first cultural asset; extracting color information by applying a color recognition algorithm to the image of the first cultural asset; extracting pattern information by applying a pattern recognition algorithm to the image of the first cultural asset; extracting the use of the first cultural asset by applying a text analysis algorithm to the explanatory text of the first cultural asset stored in the database; repeating the above steps on each cultural asset stored in the database; and creating a cultural asset classification database by tagging shape information, material information, color information, pattern information, and use extracted by repeating the steps on the image of each cultural asset. Therefore, the cultural assets can be classified and searched for according to the characteristics thereof.

Description

Method and device for creating a model for classification of cultural properties using automatic analysis of formative elements {METHOD AND APPARATUS FOR CREATING MODEL FOR CATEGORIZING CULTURAL PROPERTIES USING AUTOMATIC ANALYSIS OF FORMATIVE ELEMENT}

The present invention relates to a method and apparatus for generating a model for classifying cultural properties using automatic analysis of formative elements, and more specifically, for extracting features including formative elements in an existing database storing cultural property information. The present invention relates to a method and apparatus for generating a cultural property classification database that is easy to search by extracting information on cultural properties by category by applying an algorithm.

As history progresses, mankind has achieved cultural achievements through various activities, and advanced cultures have discovered, preserved, and recorded these cultural achievements, and succeeded by the next generation to establish a national identity and achieve advanced cultural development on top of them. I am doing it.

In addition, with the development of electronic recording media and information and communication technology, new methods for discovering, preserving and recording cultural properties are being developed.

In this regard, Korean Registered Patent No. 1776059 relates to "a system and method for investigating cultural assets", and according to the request of a user terminal equipped with an application for investigating cultural assets, the location, cadastral, old building information, and evidence And academic data, old map information, and chronological analysis information are formed in a plurality of layers, and layers according to the user's selection are displayed by overlapping among the plurality of layers, so that useful information can be provided to users when investigating or discovering cultural properties. Provides a cultural property investigation system and its method.

Meanwhile, in order for cultural property information discovered, preserved and recorded through considerable effort to have a present meaning, it must be able to be easily searched and learned by current generations.

However, until now, there is a problem that it is difficult to search and learn because cultural property information has not been systematically arranged on a computerized basis.

Accordingly, a classified database that not only discovers, preserves, and records cultural properties, but also allows easy and systematic access to recorded cultural property information, can be created using artificial intelligence that can automatically analyze formative elements. There is a need for skills that are there.

Meanwhile, the above-described prior art is technical information possessed by the inventor for derivation of the present invention or acquired in the derivation process of the present invention, and is not necessarily known to be publicly known to the general public prior to filing the present invention. .

An embodiment of the present invention aims to solve the problem that a systematic search for cultural properties cannot be performed through a database in which information on cultural properties is recorded in an array format.

In addition, an embodiment of the present invention is to solve the problem that the characteristics of each cultural property are not arranged by category in the conventional database, and thus classification and search by characteristics of cultural property cannot be performed.

In addition, an embodiment of the present invention is to solve the problem of spending too much time and money when manually classifying and systematically organizing information from a database in which cultural property information is recorded.

In order to solve the above-described problems, an embodiment of the present invention provides a method and apparatus for generating a model for classifying cultural assets from an existing cultural heritage database through automatic analysis of formative elements using artificial intelligence.

A method for generating a model for classifying cultural assets according to an embodiment of the present invention includes extracting the shape of the cultural asset from the image of the cultural asset stored in the database through a shape recognition algorithm, and extracting the use of the cultural asset through a text analysis algorithm. It can provide a method to analyze the correlation between the shape of the spear and the use of cultural properties.

In addition, the method for generating a model for classification of cultural properties according to an embodiment of the present invention is to extract the shape of the cultural property from the cultural property image stored in the database through a shape recognition model generated through machine learning, and extract the shape and cultural property. It is possible to create an estimation model for the use according to the shape of the cultural property by correlating the use of it.

On the other hand, a method for generating a model for classifying cultural properties using automatic analysis of formative elements according to an embodiment of the present invention includes accessing a database storing cultural property information, and recognizing a shape in the image of the first cultural property stored in the database. Extracting shape information by applying an algorithm, extracting material information by applying a material recognition algorithm to the image of the first cultural property, extracting color information by applying a color recognition algorithm to the image of the first cultural property, 1 Extracting pattern information by applying a pattern recognition algorithm to the image of the cultural property, extracting the purpose of the first cultural property by applying a text analysis algorithm to the explanatory text of the first cultural property in the database, and performing the above steps in the database. Step of repeating each of the stored cultural assets, and creating a cultural asset classification database by tagging the shape information, material information, color information, pattern information, and usage extracted by the above repeating step on the image of each cultural asset It may include.

Here, the step of extracting the use of the first cultural property may include generating a signal for requesting confirmation from the user of the use extracted by the text analysis algorithm, and receiving a confirmation signal from the user regarding the use.

A method for generating a cultural asset classification model according to another embodiment of the present invention is an estimation model for the form information and use of the cultural asset by analyzing the correlation between the form information and the use of each cultural asset accumulated by the repeating step. It may further include the step of generating.

Here, in the step of generating the estimation model, the deep neural network model is trained with a data set tagged with a label indicating the use identified by the user in the form information of each cultural property accumulated by the repeating step. It may include generating a DEEP NEURAL NETWORK (DNN) model for estimating the use according to the type information.

Meanwhile, the color recognition algorithm may be configured to determine color information after correcting a color of an image in consideration of a color change due to a material.

In a method for generating a cultural property classification model according to another embodiment of the present invention, after the step of receiving a confirmation signal from a user for a use, when the steps a to f are repeated for a second cultural property, the first If the shape information extracted for the cultural property matches the shape information extracted for the second cultural property, the use identified for the first cultural property in the step of extracting the use for the second cultural property is extracted as the use for the second cultural property, and It may further include the step of receiving a confirmation signal from the user.

A method for generating a cultural property classification model according to another embodiment of the present invention is, after receiving a confirmation signal from a user for a use, performing detailed shape analysis on the first cultural property according to the identified use. Further comprising a step, the detailed shape analysis, if the purpose is a bottle, jar, or kettle, analyze the shape of the first cultural property by dividing it into a spout, a neck, a trunk, a bottom, a handle, and a lid, and the purpose is foot or In the case of cruelty, the first cultural property is analyzed by dividing the shape of the first cultural property into a frame, a body, and a handle. If the purpose is a plate, the first cultural property is analyzed by dividing it into a frame, a body, and a bottom. 1 It is possible to analyze the shape of a cultural property by dividing it into a body, a lid, and a layer.

A method for generating a cultural property classification model according to another embodiment of the present invention is to determine at least one of the production time, production place, or producer of the first cultural property by applying a text analysis algorithm to the description text for the first cultural property in the database. It may further include the step of extracting.

The computer-readable recording medium for generating a cultural property classification model according to an embodiment of the present invention may be a computer-readable recording medium storing a computer program for executing any of the above methods using a computer.

An apparatus for generating a model for classification of cultural properties using automatic analysis of formative elements according to an embodiment of the present invention includes a communication unit for accessing a database storing cultural property information, and a shape recognition algorithm applied to an image of a cultural property stored in the database. Then, shape information is extracted, material information is extracted by applying a material recognition algorithm to images of cultural properties, color information is extracted by applying a color recognition algorithm to images of cultural properties, and pattern recognition algorithms are applied to the images of cultural properties. Feature extraction unit for extracting information and extracting the use of cultural assets by applying a text analysis algorithm to the explanatory text of the cultural assets in the database, shape information, material information, and color information extracted by the extraction unit on images of each cultural asset It may include a classification DB generator for generating a cultural property classification database by tagging, pattern information, and usage.

In addition to this, another method for implementing the present invention, another system, and a computer program for executing the method may be further provided.

In addition, aspects, features, and advantages other than those described above will become apparent from the following drawings, claims, and detailed description of the invention.

According to an exemplary embodiment of the present invention, it is possible to automatically generate a database in which information on cultural assets is systematically classified by using information recorded in a database in which information on cultural assets is recorded in an array manner.

In addition, an embodiment of the present invention creates a cultural asset classification database in which the characteristics of each cultural asset are arranged by category using a database in which the characteristics of cultural assets are not arranged by category, and the cultural asset classification database is classified by characteristics of the cultural asset. And search can be made.

In addition, according to an embodiment of the present invention, it is possible to save time and cost in creating a cultural property classification database by automatically classifying and systematically organizing information from a database in which cultural property information is recorded.

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

1 shows an environment for generating a cultural property classification database and a cultural property classification model according to an embodiment of the present invention.
FIG. 2 is a diagram for explaining a method of classifying features of cultural properties by using formative elements according to an embodiment of the present invention.
3 is a view for explaining a process of extracting form information of a cultural property by applying a shape recognition algorithm to an image of a cultural property according to an embodiment of the present invention.
4 is a diagram for explaining a method of classifying features of cultural assets by category according to an embodiment of the present invention.
5 is a flowchart illustrating a method for generating a cultural property classification model according to an embodiment of the present invention.
6 is a view for explaining the classification of cultural assets by use according to an embodiment of the present invention.
FIG. 7 is a diagram for explaining a method of classifying cultural properties by use according to an embodiment of the present invention, and then classifying detailed forms additionally.
8 is a diagram for explaining a method of classifying detailed forms of cultural assets classified into bottles, jars, and kettles according to an embodiment of the present invention.
9 is a view for explaining a method of classifying patterns of cultural properties according to an embodiment of the present invention.
10 is a view for explaining a method of classifying a material of a cultural property according to an embodiment of the present invention.

Advantages and features of the present invention, and a method of achieving them will become apparent with reference to embodiments described in detail together with the accompanying drawings. However, it should be understood that the present invention is not limited to the embodiments presented below, but may be implemented in a variety of different forms, and includes all transformations, equivalents, or substitutes included in the spirit and scope of the present invention. . The embodiments presented below are provided to complete the disclosure of the present invention, and to fully inform a person of ordinary skill in the art to which the present invention belongs. In describing the present invention, when it is determined that a detailed description of a related known technology may obscure the subject matter of the present invention, a detailed description thereof will be omitted.

The terms used in the present application are used only to describe specific embodiments, and are not intended to limit the present invention. Singular expressions include plural expressions unless the context clearly indicates otherwise. In the present application, terms such as "comprise" or "have" are intended to designate the presence of features, numbers, steps, actions, components, parts, or combinations thereof described in the specification, but one or more other features. It is to be understood that the presence or addition of elements or numbers, steps, actions, components, parts, or combinations thereof, does not preclude in advance. Terms such as first and second may be used to describe various elements, but the elements should not be limited by the terms. These terms are used only for the purpose of distinguishing one component from another component.

Hereinafter, embodiments according to the present invention will be described in detail with reference to the accompanying drawings, and in the description with reference to the accompanying drawings, identical or corresponding components are assigned the same reference numbers, and redundant descriptions thereof are omitted. I will do it.

1 illustrates an environment for generating a cultural property classification database and a cultural property classification model according to an embodiment of the present invention.

A system for generating a cultural property classification model may include a cultural property database 2000, a cultural property classification estimation server 3000, and a user terminal 5000.

As shown in FIG. 1, the cultural property database 2000 may include image information on various cultural properties and a textual description of each cultural property. In the textual description, the purpose of the cultural property, color, material, pattern, production time such as the production date of the cultural property, the maker, the production place, production background, production method, excavation information, and the like may be described.

Information on cultural properties stored in the cultural properties database 2000 is only stored by linking the above text information to images of each cultural property, and information on cultural properties is not systematically classified by category.

The user terminal 5000 may be a variety of information processing devices such as a smart phone, a tablet, a computer, etc., and is an electronic device including a communication module capable of accessing the cultural asset database 2000 to search for information on cultural assets.

The user terminal 5000 may communicate with the cultural property database 2000 and the cultural property classification estimation server 3000 through various types of wired/wireless communication networks.

The user terminal 5000 searches for each cultural property by accessing the cultural property database 2000 through a communication network by a user or by a preset program. When the first cultural property searched is referred to as a first cultural property, the user terminal 5000 may extract shape information of the first cultural property by applying a shape recognition algorithm to the image of the first cultural property.

The shape recognition algorithm may use several algorithms that extract the outline of an object through image analysis. The shape recognition algorithm may be a program created using various algorithms such as OpenCV and an outline tracking algorithm.

The shape recognition algorithm may be embedded in the user terminal 5000, stored in an external server such as the cultural property classification estimation server 3000, and designed to allow the user terminal 5000 to access and utilize such a server.

In addition, the user terminal 5000 provides information on the material of the first cultural property through a material recognition algorithm from the image of the first cultural property, color information of the first cultural property through a color recognition algorithm, and the first cultural property through a pattern recognition algorithm. The pattern information of can be extracted.

In addition, the user terminal 5000 may extract information about the purpose, production time, production place, and producer of the first cultural property through a text analysis algorithm from the text description of the first cultural property.

Like the shape recognition algorithm, the material recognition algorithm, color recognition algorithm, pattern recognition algorithm, and text analysis algorithm may be embedded in the user terminal 5000, or stored in an external server such as the cultural property classification estimation server 3000, and the user The terminal 5000 may be designed to access and utilize such a server.

The user terminal 5000 may extract shape information, material information, color information, pattern information, usage, and the like by applying each of the above algorithms to each cultural property stored in the cultural property database 2000.

The user terminal 5000 may generate a cultural asset classification database by tagging information such as shape information, material information, color information, pattern information, and usage, etc. extracted from the above images of each cultural property.

The cultural property classification database may be stored in the user terminal 5000 or in the cultural property classification estimation server 3000.

The user terminal 5000 or the cultural asset classification estimation server 3000 may generate a cultural asset classification estimation model using a deep neural network of artificial intelligence based on the stored cultural asset classification database.

For example, the user terminal 5000 or the cultural property classification estimation server 3000 may analyze a correlation between the accumulated form information and usage of each cultural property to generate an estimation model for the form information and usage of the cultural property.

More specifically, the user terminal 5000 or the cultural property classification estimation server 3000 trains a deep neural network model with a data set tagged with a label for the purpose extracted from the accumulated shape information of each cultural property, A DEEP NEURAL NETWORK (DNN) model that estimates the use according to the shape information can be created.

The user terminal 5000 or the cultural property classification estimation server 3000 may perform the above operations according to a command of a processor by a preset program.

FIG. 2 is a diagram for explaining a method of classifying features of cultural properties by using formative elements according to an embodiment of the present invention.

According to the system studied by the inventors of the present invention, cultural relics can be classified into function, color, shape, material, and pattern.

For example, the celadon inlaid Gukmoranmun Jang Kyung-byeong, which is in the collection of the National Museum of Korea, is a bottle in function, green in color, has a long snout and neck, a convex body, and the material is celadon, and the pattern is It has peony pattern and chrysanthemum pattern.

As in the above example, cultural properties may be classified through five predetermined categories, and the number of categories may be determined in various ways according to exemplary embodiments.

However, for a unified database system, it may be desirable to set classification categories collectively and apply them to all cultural properties.

In the category of FIG. 2, functions may be extracted by analyzing text describing cultural properties. As another example, when deciding the use of the second cultural property, using information on the first cultural property whose purpose was previously determined, if the shape of the first cultural property and the second cultural property are similar, it is determined that the function will be similar, The use of the first cultural property similar in shape to that of the second cultural property may be adopted.

Color information can be extracted from an image of a cultural property through a color recognition algorithm, and shape or shape information can be extracted from an image of a cultural property through a shape recognition algorithm.

Meanwhile, since even the same color may be expressed differently in a photo depending on a material, a color recognition algorithm may be configured to determine color information after correcting the color of an image in consideration of color change due to the material.

The material may be extracted from the image of the cultural property through a material recognition algorithm, or may be extracted from the text description of the cultural property through a text analysis algorithm.

Patterns of cultural properties can be extracted from images of cultural properties through a pattern recognition algorithm, or from text descriptions of cultural properties through text analysis algorithms.

3 is a view for explaining a process of extracting form information of a cultural property by applying a shape recognition algorithm to an image of a cultural property according to an embodiment of the present invention.

After the user terminal 5000 accesses the cultural asset database 2000 to load the image for the first cultural asset, the outline of the first cultural asset image may be extracted through a shape recognition algorithm.

As shown in FIG. 3, the shape recognition algorithm may convert a photographic image into image information composed of simple lines by tracking the outline of the image.

Here, it can be seen that information such as color, material, and pattern of the first cultural property is omitted so that form information indicating only the form in a simpler form can be calculated as a result.

Since cultural properties performing a specific function have a similar shape to perform the function, information on the shape of the cultural property can be a means for estimating the function or purpose of the corresponding cultural property.

On the other hand, since the uses extracted from the text description of the cultural property through a text analysis algorithm or the use determined by referring to the use information of a similar type of cultural property cannot be completely accurate, the accuracy of the usage information can be increased by obtaining user confirmation.

In this case, the user terminal 5000 may generate a signal requesting the user to confirm the purpose extracted by the text analysis algorithm, and allow the user to confirm the extracted purpose.

If the extracted purpose is correct, the user may touch the confirmation button to transmit a confirmation signal for the use to the user terminal 5000, and if it is wrong, the user may touch the cancel button.

The user terminal 5000 may check whether the extracted purpose is correct or incorrect according to the user's feedback, and may store only the purpose identified by the user as a use for a corresponding cultural property.

4 is a diagram for explaining a method of classifying features of cultural assets by category according to an embodiment of the present invention.

Since cultural properties may not consist of only one color, color information may be expressed as information about a specific number of colors from the brightest to the darkest. For example, in FIG. 4, it was set to extract five colors seen in porcelain sagin.

Uses according to shape or shape can be classified into bottles, jars, dishes, feet, cups, sums, and kettles. In addition to the basic shape, detailed shape classification may be additionally performed according to these uses.

For example, additional detailed shape analysis is conducted by dividing the shape of a cultural property into a spout, neck, body, bottom, handle, and lid if the purpose is a bottle, jar, or kettle, and if the purpose is foot or cruel, it is a cultural property. The shape of the cultural property is divided into frame, body, and handle to analyze.If the purpose is a plate, the shape of the cultural property is divided into frame, body and bottom, and if the purpose is combined, the shape of the cultural property is divided into body, lid, and layer. Can be configured to

Patterns of cultural properties can be classified into plants, animals, natural landscapes, objects, characters, and people, and after the types of patterns are classified, whether the patterns are single, continuous, or mixed can be additionally analyzed.

The material of the cultural property may be classified into earthenware, celadon, buncheongsagi, white porcelain, pottery, and the like, and as shown in FIG. 4, detailed classification may be additionally performed according to the material.

5 is a flowchart illustrating a method for generating a cultural property classification model according to an embodiment of the present invention.

First, the user terminal 5000 for generating a cultural property classification model may access the cultural property database 2000 by a predetermined program (S110) and select a cultural property.

The user terminal 5000 may extract shape information from the image of the selected cultural property through a shape recognition algorithm (S120). Here, the shape recognition algorithm may be stored in the user terminal 5000 or stored in an external server may be used by the user terminal 5000.

The user terminal 5000 may also extract material information from the cultural property image through a material recognition algorithm (S130). In addition, the user terminal 5000 may extract color information from a cultural property image through a color recognition algorithm (S140). Here, since the color information may look different in the photo depending on the material, the color recognition algorithm may extract color information by correcting the color change according to the material.

In addition, the user terminal 5000 may extract pattern information of the selected cultural property through a pattern recognition algorithm (S150). The pattern recognition algorithm can detect repetitive patterns appearing on the surface of cultural properties and recognize them as patterns. The pattern can be a single pattern or a mixture of multiple patterns.

Thereafter, the user terminal 5000 may extract usage information of the selected cultural property (S160). The usage information can be extracted from the text description of the selected cultural property through a text analysis algorithm, a user's direct input can be requested, or the usage information of other cultural properties having a similar shape can be extracted.

For example, if the purpose of the first cultural property is determined to be a plate through user confirmation, etc., and it is determined that the shape of the second cultural property is similar to the first cultural property, the user terminal 5000 will indicate that the use of the second cultural property is also a plate. Can be estimated. For these estimates, the correct use can be determined through user confirmation.

With respect to the extracted usage information, a user confirmation request signal may be generated to receive confirmation whether the extracted usage information is correct (S170). Upon receiving a confirmation signal for the extracted usage information from the user, the extracted usage information may be stored in the cultural property classification database together with the cultural property image. The cultural asset classification database may be formed in the memory of the user terminal 5000 or may be formed in the cultural asset classification estimation server 3000.

If the user refuses to confirm the extracted usage information, the user terminal 5000 may request the user to input usage information for the selected cultural property and receive a response thereto (S175).

The user can accurately respond to the purpose of the selected cultural property according to his or her knowledge or the result of his search.

The user terminal 5000 may store the received usage information together with the image of the cultural asset in the cultural asset classification database.

Steps S120 to S180 may be repeatedly performed on cultural assets stored in the cultural asset database, and accordingly, shape information, material information, color information, pattern information, usage information, etc. for each cultural asset may be derived. .

In this repetitive step, use information previously extracted for other cultural properties may be used when deriving the use information. For example, if the shape information extracted for the first cultural property whose purpose was previously identified and the shape information extracted for the second cultural property match, the use of the second cultural property is extracted as the purpose identified for the first cultural property. can do.

Of course, even in this case, it may further include receiving a confirmation signal for the extracted use from the user.

In addition, although not shown in FIG. 5, the user terminal 5000 may additionally extract other information such as the production time, production place, or producer of the cultural property by applying a text analysis algorithm to the explanatory text of the cultural property in the cultural property database 2000. have.

On the other hand, the device for generating a model for classifying cultural properties using automatic analysis of formative elements for performing the steps as in FIG. 5 includes a communication unit for accessing a database storing cultural property information, and feature extraction for extracting features of cultural properties. It may include a classification DB (Database) generation unit that generates a cultural property classification database by tagging the information extracted to the sub and the cultural property image.

Here, the feature extraction unit extracts form information by applying a shape recognition algorithm to the image of cultural property stored in the database, extracts material information by applying a material recognition algorithm to the image of the cultural property, and applies a color recognition algorithm to the image of the cultural property. Color information can be extracted, pattern information can be extracted by applying a pattern recognition algorithm to images of cultural properties, and text analysis algorithms can be applied to explanatory texts of cultural properties in a database to extract the uses of cultural properties.

The feature extraction unit may probabilistically express each feature extracted for a specific cultural property. For example, the feature extraction unit may express the extraction result in a manner such as 95% probability that the upper part of the first cultural property is red, 3% probability that it is orange, and 1% probability that it is purple.

The classification DB generation unit may generate a cultural property classification database by tagging the shape information, material information, color information, pattern information, and usage extracted by the extraction unit on the image of each cultural property.

6 is a view for explaining the classification of cultural assets by use according to an embodiment of the present invention.

Pottery and earthenware can be classified into bottles, jars, dishes, and bales. Bottles, jars, dishes, and bales can be further subdivided and classified according to their shape.

The bottles are more subdivided, and are divided into Jangkyungbyeong, Maebyeong, Okhochunbyeong, Rice Cakebyeong, Tongyeongbyeong, Pyeongyeongbyeong, Square Byeong, Yubyeong, Jarabyeong, Jeongbyeong, General, Pyeonbyeong, Polygonal Byeong, Overhaul, Gwanggubyeong, Yangyi Byeong, etc. Can be.

Urns can be divided into Ipho, Wonho, Dankyungho, Jangkyungho, Gwangguho, Tahoe, Ssangyiho, Saiho, Daebuho, Yugonggwangho, Juguho, complex, Yugaeho, Taehangari, and Golho.

Dishes may be classified into an abscess dish, a direct-ball dish, a lip dish, a front dish, a fire dish, a square dish, a multi-angle dish, a bent dish, and a large dish.

Feet can be divided into a bowl, a foot, a pendulum, a head/basin, an ear head, a guard foot, and a bend.

The classification table as shown in FIG. 6 is prepared in advance, and the user terminal 5000 may classify cultural assets based on the classification table after extracting form information and usage information of each cultural asset from the cultural asset database 2000.

As another example, the user terminal 5000 may extract form information and use information from cultural assets stored in the cultural assets database 2000 and classify each cultural asset into groups based on the extracted form information and use information. The user can name each group through the user terminal 5000. Through this operation, a classification table as shown in FIG. 6 may be generated.

The above-described classification scheme is given only as an example, and the content of the present invention is not limited thereto.

7 is a view for explaining a method of classifying a cultural property by use according to an embodiment of the present invention and then further classifying detailed forms.

Cultural properties have certain morphological limitations depending on their use, and accordingly, criteria for dividing detailed forms of cultural properties for each use can be determined.

Bottles can be generally divided into spout, neck, body, and bottom, kettle can be divided into lid/neck, spout, handle, body, and bottom, and jar can be divided into spout/neck, handle, body, and bottom. I can.

In addition, a plate can be divided into a rim, a body, and a bottom, a foot can be divided into a rim, a body, and a bottom, a cup can be divided into a rim, a body, and a bottom, and the sum can be divided into a lid and a body. have.

8 is a diagram for explaining a method of classifying detailed forms of cultural assets classified into bottles, jars, and kettles according to an embodiment of the present invention.

As described above, detailed shapes may be classified according to categories classified for each cultural property for a specific purpose, but FIG. 8 describes a method of classifying detailed shapes for bottles, jars, and kettles among cultural properties for various uses.

For example, the snout can be divided into shape, intangible, cone-shaped, lip-shaped, wide-sphere, burning, trumpet-shaped, and plate-shaped, and the neck can be divided according to height, width, and position, The body can be divided according to the height and width of the top, middle, and bottom. In addition, the shape can be divided into a basic type or an additional type.

The bottom can be divided according to the width, the presence of legs, and the presence or absence of a base, the handle can be divided by the presence or absence of a handle, and the lid can be divided by the presence or absence of a lid. Of course, the handle and lid can be divided according to the shape in more detail.

The above-described classification method is given only as an example, and the content of the present invention is not limited thereto.

9 is a diagram for explaining a method of classifying patterns of cultural properties according to an embodiment of the present invention.

A pattern is a shape engraved or drawn on the surface of a cultural property, and can be largely divided into categories of nature/landscape, figures, objects, characters, and geometry.

In nature/landscape, there may be clouds, waves, waterside scenery, tenjangsaeng, and Sehansamwoo, a person may have an adult and a child, and a piece may have a bojagi, a treasure, and a Yeouido.

In the text, there may be swastika, restorative gangnyeongja, restorative character, suja, aja, hoeja, bokja, gangnyeongja, annual Iksu, Mansumugang, etc., and geometry may include lightning, tortoiseshell, line, and net.

The above-described patterns are given only as examples, and the content of the present invention is not limited thereto.

10 is a view for explaining a method of classifying a material of a cultural property according to an embodiment of the present invention.

When cultural assets are classified by material, they can be largely divided into earthenware, celadon, and buncheongsagi.

Earthenware may include soft earthenware, hard earthenware, and condensed earthenware, and celadon may include green celadon, inlaid celadon, Jinsa celadon, Yeonri character, pure celadon, hoe celadon, and iron celadon.

In addition, the Buncheongsagi can be classified into techniques used in manufacturing such as printing technique, bakjiki technique, iron painting technique, make-up technique, inlay technique, intaglio technique, and earyal technique.

The above-described classification system for cultural properties by criteria is established through data collection and research by humans, and according to these criteria, the user terminal 5000 can classify cultural properties in the cultural properties database 2000 according to the established classification system.

However, without establishing a classification system through data collection and research by humans, the user terminal 5000 extracts and classifies the features of the cultural assets database 2000 through the various feature extraction algorithms mentioned above, and classifies the cultural assets systematically. It is also possible to create a database.

By taking this method, instead of manually classifying vast amounts of cultural property data, automatic classification through artificial intelligence may be possible, and a new classification system different from the method performed by humans may be derived.

The embodiment according to the present invention described above may be implemented in the form of a computer program that can be executed through various components on a computer, and such a computer program may be recorded in a computer-readable medium. In this case, the medium is a magnetic medium such as a hard disk, a floppy disk, and a magnetic tape, an optical recording medium such as a CD-ROM and a DVD, a magnetic-optical medium such as a floptical disk, and a ROM. , A hardware device specially configured to store and execute program instructions, such as RAM, flash memory, and the like.

Meanwhile, the computer program may be specially designed and configured for the present invention, or may be known and usable to those skilled in the computer software field. Examples of the computer program may include not only machine language codes such as those produced by a compiler, but also high-level language codes that can be executed by a computer using an interpreter or the like.

In the specification of the present invention (especially in the claims), the use of the term "above" and a reference term similar thereto may correspond to both the singular and the plural. In addition, when a range is described in the present invention, the invention to which an individual value falling within the range is applied (unless otherwise stated), and each individual value constituting the range is described in the detailed description of the invention. Same as

If there is no explicit order or contradictory description of the steps constituting the method according to the present invention, the steps may be performed in a suitable order. The present invention is not necessarily limited according to the order of description of the steps. The use of all examples or illustrative terms (for example, etc.) in the present invention is merely for describing the present invention in detail, and the scope of the present invention is limited by the above examples or illustrative terms unless limited by the claims. It does not become. In addition, those skilled in the art can recognize that various modifications, combinations, and changes may be configured according to design conditions and factors within the scope of the appended claims or their equivalents.

Therefore, the spirit of the present invention is limited to the above-described embodiments and should not be determined, and all ranges equivalent to or equivalently changed from the claims to be described later as well as the claims to be described later are the scope of the spirit of the present invention. It will be said to belong to.

Claims (10)

As a method for generating a model for classification of cultural properties using automatic analysis of formative elements,
a) accessing a database in which cultural property information is stored;
b) extracting shape information by applying a shape recognition algorithm to the image of the first cultural property stored in the database;
c) extracting material information by applying a material recognition algorithm to the image of the first cultural property;
d) extracting color information by applying a color recognition algorithm to the image of the first cultural property;
e) extracting pattern information by applying a pattern recognition algorithm to the image of the first cultural property;
f) extracting a purpose of the first cultural property by applying a text analysis algorithm to the description text of the first cultural property in the database;
g) repeating the steps for each cultural property stored in the database; And
h) including the step of creating a cultural property classification database by tagging the image of each cultural property with shape information, material information, color information, pattern information, and usage extracted by the repeating step,
The step of extracting the use of the first cultural property,
Generating a signal for requesting confirmation from the user for the purpose extracted by the text analysis algorithm, and receiving a confirmation signal from the user for the purpose,
The method for generating a model for classification of the cultural property,
After receiving a confirmation signal from the user for the use,
Further comprising the step of performing a detailed shape analysis of the first cultural property according to the identified use,
The detailed shape analysis,
When the use is a bottle, jar, or kettle, the shape of the first cultural property is analyzed by dividing it into a spout, a neck, a body, a bottom, a handle and a lid,
If the purpose is feet or cruel, the first cultural property is divided into a frame, a body, and a handle, and analyzed,
If the use is a plate, the shape of the first cultural property is analyzed by dividing it into a frame, a body, and a bottom,
If the above uses are summed, analyzing the shape of the first cultural property by dividing it into a body, a lid, and a layer,
A method for creating a model for classification of cultural properties.
delete The method of claim 1,
The method for generating a model for classification of the cultural property,
Analyzing the correlation between the type information and the use of each cultural property accumulated by the repeating step, further comprising the step of generating an estimation model for the type information and use of the cultural property,
A method for creating a model for classifying cultural properties.
The method of claim 3,
Generating the estimation model,
Training a deep neural network model with a data set in which the use identified by the user is tagged with a label in the form information of each cultural property accumulated by the iterative step to estimate the use according to the form information of the cultural property. Including the step of generating a DNN model,
A method for creating a model for classifying cultural properties.
The method of claim 1,
The color recognition algorithm is configured to determine color information after correcting the color of the image in consideration of color change due to material,
A method for creating a model for classifying cultural properties.
delete As a device for generating a model for classification of cultural properties using automatic analysis of formative elements,
A communication unit for accessing a database in which cultural property information is stored;
Shape information is extracted by applying a shape recognition algorithm to the image of the cultural property stored in the database, material information is extracted by applying a material recognition algorithm to the image of the cultural property, and color information by applying a color recognition algorithm to the image of the cultural property A feature extraction unit configured to extract pattern information by applying a pattern recognition algorithm to the image of the cultural property, and extracting the purpose of the cultural property by applying a text analysis algorithm to the explanatory text of the cultural property in the database;
Includes a classification DB generation unit for generating a cultural property classification database by tagging the shape information, material information, color information, pattern information, and usage extracted by the extraction unit on the image of each cultural property,
The feature extraction unit,
After extracting the use of the cultural property, it is further configured to generate a signal requesting confirmation from the user for the purpose extracted by the text analysis algorithm, and to receive a confirmation signal from the user for the use,
The feature extraction unit,
When receiving a confirmation signal from the user for the use, it is additionally configured to perform detailed shape analysis of the cultural property according to the identified use,
The detailed shape analysis,
If the use is a bottle, jar, or kettle, the shape of the cultural property is analyzed by dividing it into a spout, a neck, a body, a bottom, a handle, and a lid,
If the above uses are feet or cruel, analyze the shape of the cultural property by dividing it into a frame, a body, and a handle,
If the use is a plate, the shape of the cultural property is divided into a frame, a body and a bottom, and analyzed,
When the above uses are summed, the shape of the cultural property is analyzed by dividing it into a body, a lid, and a layer,
A device for creating a model for classifying cultural properties.
delete The method of claim 7,
The feature extraction unit,
After receiving a confirmation signal from the user regarding the use of the first cultural property, if the shape information extracted for the second cultural property matches the shape information extracted for the first cultural property, the confirmed use for the first cultural property Is further configured to extract for use for the second cultural property,
A device for creating a model for classifying cultural properties.
delete

Images