KR101639657B1 - Method and server for searching similar goods - Google Patents

Abstract

A similar article search method and a similar article search server are provided. The similar product retrieval method may further include retrieving at least one attribute related to an image of a first product selected by the user, the method further comprising the steps of: The method comprising the steps of: obtaining at least one attribute relating to an image of a second product; analyzing attributes relating to the image of the plurality of goods to estimate the user's interest property and obtaining an interest property value; Retrieving an analogous product from the product database using the attribute value, and providing the retrieved similar product information to the user's terminal.

Description

METHOD AND SERVER FOR SEARCHING SIMILAR GOODS [0002]

The present invention relates to a similar article search method and a similar article search server.

A user can access a shopping site through a user terminal such as a personal computer, and can search for various items such as clothes, shoes, bags, etc., which are held at the shopping site.

The conventional shopping site mainly provides a search interface that uses text as a query, and provides information of a product matching a search keyword input by a user as a search result. Alternatively, the conventional shopping site classifies products into various categories using meta information of products, and provides information of products matching search conditions (or items) selected by the user as search results.

However, in the conventional shopping site, it is difficult for a user to search for a desired product. For example, when a user wants to find a particular garment, it is often difficult to express the style, total length, neckline, sleeve length, pattern, and the like.

Published Patent Application No. 10-2010-0044669, April 30, 2010

The object of the present invention is to provide a similar product retrieval method and a similar product retrieval method for providing an information of a product similar to a product selected by a user by comparing the image of the product selected by the user with the image of another product, .

Another object of the present invention is to provide a similar goods search method and an similar goods search server for analyzing images of a plurality of goods selected by a user continuously and quickly providing goods information suitable for a user's search intention.

The problems to be solved by the present invention are not limited to the above-mentioned problems, and other problems that are not mentioned can be clearly understood by those skilled in the art from the following description.

According to an aspect of the present invention, there is provided a similar product search method performed by a similar product search server, the method comprising: displaying at least one attribute related to an image of a first product selected by a user Obtaining at least one attribute relating to an image of a second product that the user subsequently selects for the first product, analyzing attributes related to the image of the plurality of goods to estimate the user's interest property And retrieving an item of interest from the item database using the item of interest and the item of interest; And providing the retrieved similar product information to the user's terminal.

In some embodiments, the step of retrieving the similar product from the product database using the interest property and the property of interest includes computing a similarity score using at least one attribute of each product image in the product database, , Similar goods can be searched by assigning different weights according to whether the attribute values are the same or the attribute having the same attribute value is the interest attribute.

In some embodiments, the step of retrieving an analogous product from the product database using the property of interest and the property of interest may include the step of determining that at least one property value for an image of each product in the product database is equal to the property of interest It is possible to search for similar goods.

In some embodiments, the similar product search method further comprises analyzing attributes relating to the images of the plurality of goods to estimate the user's interest attributes and obtaining the interest property values, And obtaining at least one attribute relating to the image of the selected third merchandise.

In some embodiments, at least one attribute of the image of the item may be differently defined according to the category of the item.

In some embodiments, at least one attribute relating to an image of the good may be extracted from a characteristic of at least one major area of the image of the good.

In addition, the at least one main area may be defined differently according to the categories of the goods.

In some embodiments, at least one attribute relating to an image of the article may be extracted from the name of the article or the meta information of the article including the description of the article.

In some embodiments, the similar product retrieval method further comprises, prior to the step of the user obtaining at least one attribute relating to the image of the first product, retrieving each item of the product and each item extracted from the meta information of each item And creating a goods property database that stores at least one property relating to the image of the goods.

The product property database may store at least one attribute related to an ID of each product and an image of each product, and may be reverse indexed according to the at least one property value.

Also, the step of creating the product property database may be a step of creating the product property database by removing unnecessary goods or goods information by using the attributes of each product and the meta information of each product .

According to an aspect of the present invention, there is provided a similar product search server for obtaining at least one attribute related to an image of a first product selected by a user, A user intention estimating unit for estimating an interest property of the user by analyzing the attributes of the images of the plurality of goods and obtaining an interest property value, And a goods information transmitting unit for providing information of the searched similar goods to the terminal of the user.

In some embodiments, the merchandise search unit may calculate similar scores using at least one attribute relating to an image of each product in the merchandise database, and determine whether the attribute having the same attribute value or having the same attribute value is an attribute of interest A similar product can be searched by assigning different weights to the product.

In some embodiments, the merchandise search unit may search for similar merchandise by determining whether at least one attribute value of an image of each merchandise in the merchandise database is equal to the interest attribute value.

In some embodiments, the image attribute obtaining unit may obtain at least one attribute relating to an image of a third article subsequently selected by the user subsequent to the second article.

In some embodiments, at least one attribute of the image of the item may be differently defined according to the category of the item.

In some embodiments, at least one attribute relating to an image of the good may be extracted from a characteristic of at least one major area of the image of the good.

In addition, the at least one main area may be defined differently according to the categories of the goods.

In some embodiments, at least one attribute relating to an image of the article may be extracted from the name of the article or the meta information of the article including the description of the article.

In some embodiments, the similar product search server further includes a database creation unit that creates a product property database that stores at least one attribute related to an image of each product and an image of each product extracted from the meta information of each product .

The product property database may store at least one attribute related to an ID of each product and an image of each product, and may be reverse indexed according to the at least one property value.

Also, the database generating unit may generate the product attribute database by removing unnecessary goods or goods information by using the attribute of each product image and the meta information of each product.

Other specific details of the invention are included in the detailed description and drawings.

According to the similar product retrieval method and the similar product retrieval server of the present invention, since the similar product is retrieved from the product database using the attribute related to the image of the product selected by the user, not only the text but also the image of the product is received as a query, It is possible to provide information of a product similar to the selected product.

According to the similar product retrieval method and the similar product retrieval server of the present invention, the attributes of the user's interest are analyzed by analyzing the attributes of the first product selected by the user and the images of the second product selected subsequently, Since the similar goods are searched by using the interest attribute, information of goods suitable for the user's search intention can be quickly provided.

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

1 is a block diagram illustrating an environment in which a similar goods search server according to an embodiment of the present invention is provided.
2 is a block diagram illustrating a configuration of a similar goods search server according to an embodiment of the present invention.
3 is a block diagram for explaining a configuration for generating a goods database of the similar goods search server according to an embodiment of the present invention.
4A and 4B are diagrams for explaining estimation of a user's interest property in an embodiment of the present invention.
5 is a flowchart illustrating a similar product retrieval method according to an embodiment of the present invention.

BRIEF DESCRIPTION OF THE DRAWINGS The advantages and features of the present invention and the manner of achieving them will become apparent with reference to the embodiments described in detail below with reference to the accompanying drawings. It will be understood, however, that the invention is not limited to the disclosed embodiments, but is capable of other various forms of implementation, and that these embodiments are provided so that this disclosure will be thorough and complete, and will fully convey the scope of the invention to those skilled in the art. And the present invention is only defined by the scope of the claims.

It is to be understood that one element is referred to as being "connected to" or "coupled to" another component, either directly connected or coupled to another component, Element is included. On the other hand, when an element is referred to as being "directly coupled to" or "directly coupled to " another element, it means that it does not intervene in another element.

Like reference numerals refer to like elements throughout.

As used herein, the term " and / or "includes each and every combination of one or more of the mentioned items.

Although the first, second, etc. are used to describe various components, modules and / or sections, it is to be understood that the components, modules and / or sections are not limited by these terms. These terms are only used to distinguish one element, module or section from another element, module or section. Therefore, it goes without saying that the first component, the first module or the first section mentioned below may be a second component, a second module or a second section within the technical spirit of the present invention.

The singular forms herein include plural forms unless the context clearly dictates otherwise.

The terms "comprises" and / or "comprising" as used herein mean the presence or addition of one or more other components, modules and / or sections in addition to the components, modules and / Do not exclude.

Unless defined otherwise, all terms (including technical and scientific terms) used herein may be used in a sense commonly understood by one of ordinary skill in the art. Also, predefined terms are not ideally or excessively interpreted unless explicitly defined otherwise.

Hereinafter, a similar goods search server according to an embodiment of the present invention will be described in detail with reference to FIG. 1 to FIG.

1 is a block diagram illustrating an environment in which a similar goods search server according to an embodiment of the present invention is provided.

Referring to FIG. 1, a user terminal 100, a web server 200, and a similar goods search server 300 are connected to each other through a network. The user terminal 100, the web server 200, and the similar goods search server 300 can exchange data and / or information with each other.

The network may be composed of networks of various sizes such as a local area network (LAN), a metropolitan area network (MAN), a wide area network (WAN), and the like. The network may be configured as a wired or wireless network.

The user terminal 100 may be a personal computer (PC) such as a desk top, a lap top, or the like. Alternatively, the user terminal 100 may be a portable electronic device such as a smartphone, a personal digital assistant (PDA), a tablet PC, or the like. The user terminal 100 may be a non-illustrated other computing device including a processor, input / output means, and communication means.

The Web server 200 may be a shopping site server. A user can access a shopping site provided by the web server 200 through the user terminal 100 and search for and purchase various items such as clothes, shoes, bags, etc., which are held at the shopping site.

Within a shopping site, a user can search for a desired product in various ways. For example, the user may input a search keyword and receive information of a product matching the search keyword as a search result. As another example, the user can select a predetermined search condition (or item), and receive information of a matching product with the search result. Then, the user will select and select a commodity suitable for the search intention among the plurality of commodities included in the search result list.

The web server 200 may be a portal site server that provides a variety of content services such as a café, a mail, a blog, a map, a dictionary, a news, a stock, a real estate, a movie, a music and a bulletin board in addition to shopping.

The similar product search server 300 receives an image of a product selected by the user as a query and compares the image with another product image stored in the shopping site to provide information of a product similar to the product selected by the user to the user terminal 100 do.

The similar product search server 300 may provide the similar product information directly to the user terminal 100 or may provide the similar product information to the web server 200 so that the user terminal 100 may download It is possible to receive the information of the goods.

For example, the web server 200 can update the search result list so as to include information about similar goods, thereby providing information on the similar goods to the user terminal 100. [ In this case, the search result list may be updated each time the user selects a product included in the search result list.

Alternatively, for example, the web server 200 may provide the user terminal 100 with a list of similar products different from the search result list. In this case, the similar product list may be updated each time the user selects a product included in the search result list.

The information of the similar product may include information such as an image of the similar product, name, description, price, and brand name.

The similar product search server 300 can arrange the information of the similar goods according to the order of high similarity with the product selected by the user.

Alternatively, the similar product search server 300 may arrange the information of the similar products according to the ranking of the predetermined ranking, the product price, the date and time of the registration, and the product review.

1, the present invention is not limited to this, and a plurality of Web servers 200 may be connected to the user terminal 100 and the similar product search server 300 through a network .

Although the web server 200 and the similar goods search server 300 are shown separately in Fig. 1, according to the embodiment, the web server 200 and the similar goods search server 300 may be combined into one server Will be apparent to those of ordinary skill in the art.

2 is a block diagram illustrating a configuration of a similar goods search server according to an embodiment of the present invention.

Referring to FIG. 2, the similar goods search server 300 according to an embodiment of the present invention includes a similar goods search engine 310, a goods index database 320, and a goods property database 330.

The similar product search engine 310 can search for a product similar to a product selected by the user using the product index database 320 and the product property database 330. [

The similar product search engine 310 may be software designed to retrieve information using a database. According to an embodiment, the similar product search engine 310 may be in the form of a combination of software and hardware.

The goods index database 320 may include information for providing a predetermined search condition (or item) to the user and for sorting information of a plurality of goods in the goods search result list or the similar goods list. For example, the commodity index database 320 may include fields related to each commodity ID, category ID, commodity name, commodity description, commodity price, and date and time registered in the database as shown in Table 1 below, The invention is not limited thereto.

[Table 1]

The goods property database 330 may include property information on the image of the goods. For example, the merchandise attribute database 330 may include fields related to each merchandise ID, category ID, and attribute of the merchandise image, as shown in Table 2 below, but the present invention is not limited thereto.

 [Table 2]

In the embodiment of the present invention, categories of goods can be distinguished as fashion, beauty, home appliance, computer, furniture, life, health, food, infant, travel, leisure, automobile and the like.

Also, one higher category may include a plurality of lower categories. For example, the category of fashion may include sub-categories such as women's clothing, men's clothing, unisex, sportswear, underwear, shoes, bags, jewelry, and the like. And the category of women's clothing may include sub-categories such as jacket, jumper, t-shirt, vest, blouse, skirt, cardigan, pants, coat,

Attributes related to the image of the goods may be defined differently depending on the category of the goods. As an example, a product belonging to a category of women's clothing may include attributes such as style, total length, neckline, sleeve length, pattern, color, etc. for the image. As another example, a product belonging to a bag category may include attributes such as style, ornament, main material, string, color, etc. for the image. That is, the attribute related to the image of the product can be defined according to the characteristics of the product that can describe (or describe) the image of the product.

Each attribute of an image of a product may have an attribute value indicating the shape or state of the features. For example, a neckline attribute may have attributes such as round neck, U neck, V neck, collar neck, chin neck, turtleneck, hood neck, and sleeves attributes such as sleeveless, cap sleeves, short sleeves, Long-sleeved, and the like, and the pattern attribute can have attribute values such as plain, print, floral, stripe, hop, dot, character, check,

When retrieving a similar product based on the similarity between images of a product, a method of extracting geometric features such as lines, shapes, and textures from an image may be generally used. However, extracting such features from an image and searching for similar items (ignoring the similarity of the images themselves) can ignore the similarity of the inherent properties of the product. For example, non-similar items may be searched for similar items because their geometric characteristics are similar, although the attributes such as style, total length, and neckline are different.

On the other hand, according to the present invention, attributes of an image of a product are defined according to a category of a product, and similar goods are searched using attributes related to an image of a product. Thus, .

On the other hand, in order to quickly search for information on products having similar attributes, the product property database 330 may be reverse indexed according to attribute values.

The similar product search engine 310 may include an image attribute acquisition unit 311, a user's intention estimation unit 312, a product search unit 313, and a product information transfer unit 314.

The image attribute acquisition unit 311 can acquire attributes related to images of a plurality of products selected by the user. The image attribute acquisition unit 311 may acquire a plurality of attributes related to the image of the first product selected by the user and acquire a plurality of attributes relating to the image of the second product selected by the user subsequent to the first product. The image attribute acquiring unit 311 may acquire a plurality of attributes related to the image of the third goods that the user subsequently selects to the second goods.

The image attribute acquisition unit 311 may receive the ID of the product selected by the user from the product search result list from the web server 200. [ The image attribute acquisition unit 311 can search the product attribute database 330 to see if the ID of the product exists. If the ID of the product exists in the product attribute database 330, the image attribute obtaining unit 3110 can obtain the attribute related to the image of the product.

The image attribute acquisition unit 311 can provide the acquired attribute information to the user's intention estimating unit 312. [

The user's intention estimating unit 312 may estimate attributes of the user's interest by analyzing the attributes of the images of the plurality of goods selected by the user. In the present specification, the term " attribute of interest " refers to an attribute of the user who is more careful than other attributes of the plurality of attributes related to the image of the goods to be searched.

For example, as a result of analysis of the attributes related to the images of the plurality of goods selected by the user, the retail attribute of the goods having a certain ratio or more may be short-sleeved, the pattern attribute may be plain or striped, and the neckline attribute may be round. In addition, the proportion of the products having the same color attribute as the color may be smaller than the predetermined ratio. In this case, the user intention estimating unit 312 may estimate the user's interest property based on a certain ratio. Accordingly, the user intention estimating unit 312 can estimate the retail attribute, the pattern attribute, and the neckline attribute as a user's interest attribute. The predetermined ratio may be set in advance and may be changed as needed. According to an embodiment, the user's interest attributes may have a plurality of levels.

Alternatively, a plurality of attributes relating to an image of a product (without distinguishing between an attribute of interest and an attribute of interest) have different levels of attention levels (e.g., first through nth attention levels, where n is a natural number greater than 1) . The attribute is estimated as an attribute of the first level of interest and the second rate is less than the first rate and the second rate is less than the second rate, If the attributes of the above products have the same value, then the attribute can be estimated as an attribute of the second interest level. In the above example, if the ratio of the same goods with the color attribute is less than the first rate and is equal to or greater than the second rate, the color attribute may be estimated as an attribute of the second interest level. Alternatively, if the color attributes of the plurality of products are different as a result of analysis of the attributes related to the images of the plurality of goods selected by the user, the color attribute may be estimated as an attribute of the lowest interest level.

In addition, the user intention estimating unit 312 may analyze attributes of images of a plurality of goods selected by the user for a predetermined period. The predetermined period may be configured in units of time or date.

The user intention estimating unit 312 may provide the estimated user attribute of interest to the product retrieving unit 313. [ Then, the goods search unit 313 can acquire a user's interest property value (attribute value of interest property).

The goods search unit 313 can search for similar goods from the goods database (goods index database, goods property database) by using the interest attribute and the interest property value of the user.

In some embodiments, the product retrieval unit 313 may calculate a similar score using the attributes of each product in the product database and determine whether the product is a similar product according to the similar score. In the process of calculating the similar scores, the product search unit 313 may assign different weights according to whether the attribute values are the same. For example, if the value of the retail badge attribute of the product is equal to the value of the retail badge attribute of the product selected by the user, a high weight value can be given, and a low weight value can be given if the value is different .

In addition, in the process of calculating the similar score, the product search unit 313 may add (or additionally) different weights according to whether the attribute having the same attribute value is the attribute of interest. For example, assume that the retailer attribute is the user's interest attribute, and the color attribute is not the user's interest attribute. In this case, if the value of the retail badge attribute of the product is equal to the value of the retail badge attribute of the product selected by the user, a high weight value is given. If the color attribute value of the product is equal to the color attribute value of the product selected by the user, Weighting can be given. The goods searching unit 313 may assign different weights according to the level of interest.

Alternatively, the product retrieving unit 313 may assign a negative weight. The product search unit 313 may assign a negative weight value when an attribute value of an attribute of the product is different from an attribute value of the attribute of the product selected by the user. In addition, the product search unit 313 may assign negative weights when the attribute values of interest attributes are different from each other, or when the attribute values of attributes having a certain level of interest or higher are different from each other.

In another embodiment, the product retrieval unit 31 may determine whether the similarity product is determined by determining whether the attribute value of the image of each product in the product database is the same as the property of interest, without calculating the similarity score. For example, the merchandise search unit 313 can search for a similar merchandise having the same retail merchandise attribute value as the retail merchandise attribute value of the merchandise selected by the user, without regard to other attribute values.

The product search unit 313 can provide the product information transfer unit 313 with index information such as the ID of the searched product, the category ID, the product name, the product description, the product price, and the date and time registered in the database.

The goods information transmitting unit 314 can provide the information of the searched goods to the user's terminal 100. [

3 is a block diagram for explaining a configuration for generating a goods database of the similar goods search server according to an embodiment of the present invention.

3, the similar product search server 300 according to an exemplary embodiment of the present invention further includes a database generation unit 340, an image database 350, and a shopping meta database 360 to generate a product database do.

The database generating unit 340 may generate the goods index database 320 and the goods property database 330 using the image database 350 and the shopping meta database 360. [

The image database 350 may include image information (or data) of goods.

The shopping meta database 360 may include shopping meta information of goods. For example, as shown in the following Table 3, the shopping meta database 360 stores fields related to each product ID, category ID, product name, product description, product price, date and time registered in the database, But the present invention is not limited thereto.

[Table 3]

The image database 350 and the shopping meta database 360 may be provided from a vendor server that supplies the merchandise to the shopping site, or may be provided by the web server 200.

The database generating unit 340 may extract the characteristics of each product from the shopping meta information of each product in the shopping meta database 360 and store the feature information of each product in the product index database 320. [ The database generating unit 340 may provide a predetermined search condition (or item) to the user from the shopping meta information of each product in the shopping meta database 360 or may extract information for sorting the information of the goods .

The database generating unit 340 may store the attributes related to the image of each product and the image of each product extracted from the shopping meta information of each product in the product property database 330. [

To this end, the database generation unit 340 may extract attributes related to the image of the corresponding product from the characteristics of the main regions of the product image. The database generating unit 340 may use image descriptors such as Sobel, Gabor, LBP, HOG, SIFT, and SURF well known in the art.

The main areas of the product image may be defined differently depending on the category of the product. For example, a category of a woman's clothing may be defined as a main area such as a neckline area, a retail area, a hem area, and the like in an image of a product. As another example, the category of the bag may be defined as a main area, such as a decorative area, a string area, and the like.

In addition, the database generating unit 340 can extract the attribute of the product image from the meta information of the product, that is, the product name or the product description. In this case, the product name or the product description should include keywords describing the image of the product.

The database generation unit 340 can generate a product attribute database in which unnecessary goods or information of goods is removed by using the attributes of each product image and the meta information of each product.

3, the similar goods search server 300 includes a plurality of database generators for generating the commodity index database 320 and the commodity attribution database 330, respectively, according to an embodiment of the present invention. .

4A and 4B are diagrams for explaining estimation of a user's interest property in an embodiment of the present invention. FIGS. 4A and 4B illustrate a product selected by a predetermined user on the shopping site according to the flow of time.

Referring to FIG. 4A, it can be seen that the user is looking for a product having a sleeved short sleeve attribute, a pattern attribute of plain or striped, and a neckline attribute of a round shape. On the other hand, it can be seen that the user is searching for the product irrespective of the color. Therefore, it is possible to estimate the retail badge, the pattern, and the neckline attribute as a user's interest attribute. Alternatively, it is possible to estimate the retail length, the pattern, and the neckline attributes as the attributes of the first level of interest and the color attributes as the attributes of the second level of interest.

Referring to FIG. 4B, it can be seen that the user is looking for a sleeved sleeper with a color attribute of achromatic color. On the other hand, it is also possible to know that the user is not interested in the pattern attribute and is looking for the product. Therefore, it is possible to estimate the retail position and the color attribute as the user's interest attributes. Alternatively, the retailer may estimate the color attribute as an attribute of the first interest level, and the pattern attribute as an attribute of the second interest level.

Hereinafter, a similar product search method according to an embodiment of the present invention will be described in detail with reference to FIG. For the sake of convenience of description, contents overlapping with the similar goods search server described with reference to Figs. 1 to 4 will be omitted.

5 is a flowchart illustrating a similar product retrieval method according to an embodiment of the present invention.

Referring to FIG. 5, in step S410, the similar product search method according to an embodiment of the present invention acquires at least one attribute related to an image of a first product selected by the user.

Then, in step S420, the user obtains at least one attribute relating to the image of the second product subsequently selected for the first product.

At least one attribute relating to an image of the goods may be defined differently depending on the category of the goods.

At least one attribute relating to an image of a product may be extracted from a feature of at least one major area of the product image. And, at least one main area may be defined differently depending on the category of the goods.

In addition, at least one attribute relating to an image of the goods may be extracted from the meta information of the goods including the name of the goods or the description of the goods.

Although not explicitly shown, the similar product retrieval method according to an embodiment of the present invention may further include a step of the user obtaining an attribute related to the image of the third product selected subsequently to the second product. That is, the user can acquire attributes relating to images of two or more goods selected by the user.

Then, in step S430, the attributes of the images of the plurality of goods are analyzed to estimate the user's interest property, and the value of the interest property is obtained.

Then, in step S440, similar goods are searched from the merchandise database using the user's interest attribute and the interest attribute value.

At this time, a similar score is calculated using at least one attribute related to an image of each product in the product database, and different weights are given according to the same property, interest property, interest level, can do.

Alternatively, the similar goods may be searched by judging whether at least one attribute value related to the image of each goods in the goods database is equal to the interest property value.

Then, in step S450, the information of the searched similar goods is provided to the user's terminal.

Although not explicitly shown, the similar product retrieval method according to an embodiment of the present invention stores at least one attribute related to an image of each product extracted from the image of each product and the meta information of each product in advance The goods attribute database 330 may be created.

The product property database 330 stores at least one attribute related to the ID of each product and the image of each product, and can be reverse indexed according to at least one property value.

The steps of a method or algorithm described in connection with the embodiments of the invention may be embodied directly in hardware, software modules, or a combination of the two, executed by a processor. The software modules may be stored in any form of computer readable storage medium such as a random access memory (RAM), read-only memory (ROM), EPROM, EEPROM, flash memory, registers, hard disk, removable disk, CD- Or may reside on a recording medium. An exemplary recording medium is coupled to a processor, which is capable of reading information from, and writing information to, the recording medium. Alternatively, the recording medium may be integral with the processor. The processor and the recording medium may reside in an application specific integrated circuit (ASIC). The ASIC may reside within the user terminal. Alternatively, the processor and the recording medium may reside as discrete components in a user terminal.

While the embodiments of the present invention have been described with reference to the accompanying drawings, it is to be understood by those skilled in the art that the present invention may be embodied in other specific forms without departing from the spirit or essential characteristics thereof. It is therefore to be understood that the above-described embodiments are illustrative in all aspects and not restrictive.

100: User terminal
200: Web server
300: Similar Product Search Server
310: Similar Product Search Engine
311: Image attribute acquisition unit
312: User intention estimating unit
313:
314: goods information transfer unit
320: Product index database
330: Product property database
340:
350: Image database
360: Shopping Metabase

Claims (22)

A similar product search method performed by a similar product search server,
Obtaining at least one attribute relating to an image of a first product selected by the user;
Obtaining at least one attribute of an image of a second commodity that the user subsequently selects for the first commodity;
Analyzing attributes related to an image of the plurality of goods to estimate the user's interest property and obtaining an interest property value;
Retrieving an analogous product from a product database using the attribute of interest and the property of interest; And
And providing the retrieved similar product information to the user's terminal,
Wherein the step of searching for similar goods from the goods database using the interest property and the property value of interest includes the steps of calculating a similarity score using at least one property related to an image of each product in the goods database, Or a similar product is searched by assigning different weights according to whether an attribute having the same attribute value is an interest attribute. delete A similar product search method performed by a similar product search server,
Obtaining at least one attribute relating to an image of a first product selected by the user;
Obtaining at least one attribute of an image of a second commodity that the user subsequently selects for the first commodity;
Analyzing attributes related to an image of the plurality of goods to estimate the user's interest property and obtaining an interest property value;
Retrieving an analogous product from a product database using the attribute of interest and the property of interest; And
And providing the retrieved similar product information to the user's terminal,
Wherein the step of retrieving the similar goods from the goods database using the interest property and the interest property value comprises the steps of determining whether at least one property value of the image of each product in the goods database is the same as the property of interest, To search for similar goods. The method according to claim 1 or 3,
The method of claim 1, further comprising: prior to the step of estimating a user's interest attribute and obtaining an interest property value by analyzing an attribute of the image of the plurality of goods, Further comprising the step of: acquiring the attribute. The method according to claim 1 or 3,
Wherein at least one attribute of the image of the product is defined differently according to the category of the product. The method according to claim 1 or 3,
Wherein at least one attribute relating to an image of the commodity is extracted from a feature of at least one major region of the image of the commodity. The method according to claim 6,
Wherein the at least one main area is defined differently according to the category of the goods. The method according to claim 1 or 3,
Wherein at least one attribute relating to an image of the goods is extracted from the name of the goods or the meta information of the goods including the description of the goods. The method according to claim 1 or 3,
Storing at least one attribute relating to an image of each article and an image of each article extracted from the meta information of each article, prior to the step of the user obtaining at least one attribute relating to the image of the first article Further comprising the step of creating a goods property database. 10. The method of claim 9,
Wherein the product property database stores at least one attribute related to an ID of each product and an image of each product and is reverse indexed according to the at least one property value. 10. The method of claim 9,
Wherein the step of generating the goods property database comprises the steps of creating the goods property database by removing unnecessary goods or goods information by using the attributes of each goods and the meta information of each goods, Way. An image attribute acquiring unit acquiring at least one attribute relating to an image of a first product selected by a user and acquiring at least one attribute relating to an image of a second product selected by the user subsequent to the first product;
A user intention estimating unit for estimating an interest property of the user by analyzing attributes related to images of the plurality of goods and obtaining an interest property value;
A product retrieval unit for retrieving a similar product from a product database using the interest property and the property value of interest; And
And a merchandise information transmitter for providing the searched terminal with information about the similar merchandise,
The product retrieval unit may calculate a similar score using at least one attribute related to an image of each product in the product database, and determine whether the attribute having the same attribute value or the attribute having the same attribute value has a different weight value To search for similar goods. delete An image attribute acquiring unit acquiring at least one attribute relating to an image of a first product selected by a user and acquiring at least one attribute relating to an image of a second product selected by the user subsequent to the first product;
A user intention estimating unit for estimating an interest property of the user by analyzing attributes related to images of the plurality of goods and obtaining an interest property value;
A product retrieval unit for retrieving a similar product from a product database using the interest property and the property value of interest; And
And a merchandise information transmitter for providing the searched terminal with information about the similar merchandise,
Wherein the goods search unit searches for similar goods by determining whether at least one attribute value related to an image of each of the goods in the goods database is equal to the interest property value. 15. The method according to claim 12 or 14,
Wherein the image attribute obtaining unit obtains at least one attribute related to an image of a third article that the user subsequently selects for the second article. 15. The method according to claim 12 or 14,
Wherein at least one attribute of the image of the goods is defined differently according to the category of the goods. 15. The method according to claim 12 or 14,
Wherein at least one attribute relating to an image of the product is extracted from a feature of at least one major area of the image of the product. 18. The method of claim 17,
Wherein the at least one main area is defined differently according to the category of the goods. 15. The method according to claim 12 or 14,
Wherein at least one attribute relating to an image of the goods is extracted from the name of the goods or the meta information of the goods including the description of the goods. 15. The method according to claim 12 or 14,
Further comprising a database creating unit for creating a product property database that stores at least one attribute related to an image of each product and an image of each product extracted from the meta information of each product. 21. The method of claim 20,
Wherein the product property database stores at least one attribute related to an ID of each product and an image of each product and is reverse indexed according to the at least one property value. 21. The method of claim 20,
Wherein the database generation unit generates the product property database by removing unnecessary goods or goods information by using the attributes of each product image and the meta information of each product.

Images