WO2020251174A1

WO2020251174A1 - Method for advertising user-customized fashion item and server executing same

Info

Publication number: WO2020251174A1
Application number: PCT/KR2020/006167
Authority: WO
Inventors: 전승엽
Original assignee: 오드컨셉 주식회사
Priority date: 2019-06-10
Filing date: 2020-05-11
Publication date: 2020-12-17
Also published as: KR20200141251A

Abstract

The present invention relates to a fashion item advertising server including: a user database in which style information for each user is stored; a style database in which a style image is stored; a product database in which product information for each product sold in an on-line shopping mall is stored; a fashion item extraction unit which extracts a possessed fashion item image from the user database, extracts a recommended style image, matching the possessed fashion item image, from the style database, and then extracts an image of a product, matching the recommended style image, from the product database, and selects the extracted product image as a recommended fashion item image; and a fashion item advertisement providing unit which maps the recommended fashion item image and the possessed fashion item image onto a user image to generate a user-customized advertisement image, and provides the user-customized advertisement image to a user terminal.

Description

User-customized fashion item advertisement method and server executing it

The present invention relates to a user-customized fashion item advertisement method and a server that executes the same, and more particularly, a user who advertises a product to a user by mapping it to a user image using other fashion items that match the fashion item held by the user. It relates to a customized fashion item advertisement method and a server executing the same.

In recent years, with the development of computer communication technology and the widespread use of high-speed communication networks, the number of Internet users is explosively increasing, and due to the increase in the number of Internet users, Internet shopping is greatly highlighted.

Accordingly, advertisers who need to advertise a specific product not only advertise on offline media such as existing newspapers and TV, but also advertise on portal sites or shopping malls visited by many users on the Internet to enhance their product or corporate image. You are trying to make money through

As a method of advertising on the Internet online sites of the above companies, conventionally, advertisements made with text, animation, and video are displayed on the initial screen of each Internet site. Internet sites have been generating profits from the above corporate advertisements as a major source of income.

However, Internet advertisements that display advertisement messages of their products on the initial screens of popular Internet sites are unilaterally displayed to unspecified people who visit the site, so that advertisements are exposed to the purchasing class who are interested in specific products. It wasn't controlled.

In other words, it was not a method of analyzing information such as age group, gender, or regional connection of visitors accessing the site and advertising the product that is judged to have the greatest advertising effect on the site.

Therefore, not only does not maximize the effect compared to the cost of advertising, but also has a problem in that the advertisement on the initial screen of a popular site is limited in screen size, so that a considerable cost is required for good advertisement effect.

The present invention is a user-customized fashion item advertisement method in which a user can expect to generate revenue by increasing the purchase rate of the product by advertising a product to the user by mapping it to a user image by using other fashion items matching the fashion item possessed by the user, and the same. It aims to provide a running server.

In addition, after defining a number of styles for a fashion item in advance, the present invention searches for a product corresponding to another fashion item that matches the fashion item held by the user based on the fashion item held by the user, and advertises the product to the user as a user. An object of the present invention is to provide a user-customized fashion item advertisement method and a server that executes the method for increasing the probability of clicking an advertisement.

The objects of the present invention are not limited to the above-mentioned objects, and other objects and advantages of the present invention that are not mentioned can be understood by the following description, and will be more clearly understood by examples of the present invention. In addition, it will be easily understood that the objects and advantages of the present invention can be realized by the means shown in the claims and combinations thereof.

The fashion item advertisement server for achieving this purpose includes a user database in which style information for each user is stored, a style database in which style images are stored, a product database in which product information for each product sold in an online shopping mall is stored, and the user database. Extracting the fashion item image from the database, extracting the recommended style image matching the fashion item image from the style database, extracting the image of the product matching the recommended style image from the product database, and determining the recommended fashion item image And a fashion item advertisement providing unit configured to generate a customized advertisement image by mapping the recommended fashion item image and the holding fashion item image to a user image, and provide the user customized advertisement image to a user terminal.

In addition, the user-customized fashion item advertisement method executed in the fashion item advertisement server to achieve this purpose is the step of extracting the possessed fashion item image from the style database in which style information for each user is stored, and the style database in which the style image is stored. Extracting a recommended style image matching the holding fashion item image, extracting an image of a product matching the recommended style image from a product database in which product information for each product sold in an online shopping mall is stored, as a recommended fashion item image Determining, generating a customized advertisement image by mapping the recommended fashion item image and the holding fashion item image to a user image, and providing the customized advertisement image to a user terminal.

According to the present invention as described above, there is an advantage that it is possible to increase the purchase rate of the product by increasing the purchase rate of the product by mapping it to the user image by using another fashion item matching the fashion item held by the user to generate a profit. .

In addition, according to the present invention, after defining a number of styles for a fashion item in advance, a user searches for a product corresponding to another fashion item that matches the fashion item held by the user based on the fashion item held by the user, and advertises the product to the user. There is an advantage that it can increase the probability of clicking an advertisement.

1 is a block diagram illustrating a fashion item advertisement server according to an embodiment of the present invention.

2 is a flowchart illustrating an embodiment of a method for advertising a user-customized fashion item according to the present invention.

3 is a flowchart illustrating another embodiment of a method for advertising a user-customized fashion item according to the present invention.

The above-described objects, features, and advantages will be described later in detail with reference to the accompanying drawings, and accordingly, one of ordinary skill in the art to which the present invention pertains will be able to easily implement the technical idea of the present invention. In describing the present invention, if it is determined that a detailed description of known technologies related to the present invention may unnecessarily obscure the subject matter of the present invention, a detailed description will be omitted. Hereinafter, preferred embodiments of the present invention will be described in detail with reference to the accompanying drawings. In the drawings, the same reference numerals are used to indicate the same or similar elements.

In terms of the terms used in the present specification, “own fashion item image” refers to an image of a product purchased by a user in an online shopping mall, an image of a product clicked on, and an image of a product contained in a shopping cart. Accordingly, according to the present invention, the user image is mapped using other fashion items that match the fashion items possessed by the user, and the product is advertised to the user, thereby increasing the purchase rate of the product, and thus, profit can be expected.

Referring to FIG. 1, the fashion item advertisement server 100 includes a fashion item extraction unit 110, a fashion item advertisement providing unit 120, a user database 130, a user database management unit 140, a style database 150, and A style database management unit 160, a product database 170, and a product database management unit 180 are included.

The fashion item extraction unit 110 uses the user database 130, the style database 150, and the product database 170 to determine a recommended fashion item image that matches the user's own fashion item image.

In one embodiment, the fashion item extraction unit 110 extracts the fashion item image held by the user database 130, and extracts a style image matching the fashion item image held by the style database 150, and then the product database 170 The image of the product matching the style image is extracted from and determined as a recommended fashion item image.

First, the fashion item extracting unit 110 extracts a style image matching the holding fashion item image from the style database 150.

To this end, the fashion item extracting unit 110 extracts style label and category information by analyzing the held fashion item image.

In an embodiment, the fashion item extracting unit 110 may extract a style label from the held fashion item image and cluster the held fashion item image based on the style label. It is appropriate that the style label is extracted as about the look and feel and trend of the fashion item's appearance and feel.

According to an embodiment of the present invention, a style label for a feeling that a person can feel from the appearance of a single fashion item included in a style image, a combination of a plurality of items, etc. may be extracted and used as a style label. For example, a celebrity look, a magazine look, a summer look, a feminine look, a sexy look, an office look, a drama look, and a Chanel look may be exemplified as a style label.

Meanwhile, the fashion item extraction unit 110 may use style label information as category information.

For example, the fashion item extraction unit 110 may extract style labels for women, dresses, sleeveless, linen, white, and casual look from the owned fashion item image. In this case, the fashion item extraction unit 110 uses style labels for women, dresses, tops, and bottoms as category information of the image of the fashion item, and style labels for sleeveless, linen, white, casual look, and formal look are categories. It can be used as style label information that describes the characteristics of an image of an externally held fashion item.

Then, the fashion item extraction unit 110 calculates a feature value of the fashion item image, selects at least one style image group having a distance value within a certain range from the feature value of the fashion item image, and selects a style image A style image belonging to the group may be determined as a recommended style image.

In this case, the fashion item extraction unit 110 may use style label and category information extracted from the image in order to increase the accuracy of image search. In one embodiment, the fashion item extracting unit 110 calculates the similarity of the characteristic values of the fashion item image and the style image stored in the style database 150, and the style label and/or among products having a similarity of a predetermined range or more. The recommended style image may be determined by excluding products whose category information does not match the style label and/or category information of the owned fashion item image.

For example, when the holding fashion item is a “pink floral blouse”, the fashion item extracting unit 110 may extract the style label “A”. Thereafter, the fashion item extracting unit 110 may extract, as a recommended style image, a style image having a similarity of a preset range with a “pink floral blouse” among the style images clustered with the style label “A”.

After extracting the recommended style image matching the image of the fashion item possessed through the above process, the image of the product is extracted from the product database 170 based on the recommended style image and determined as a recommended fashion item image.

That is, the fashion item extraction unit 110 extracts a matching fashion item in a category different from the holding fashion item from among fashion items included in the recommended style image, and a product corresponding to a matching fashion item among products stored in the product database 170 It is possible to extract the image of the product and determine it as a recommended fashion item image.

In this way, the reason that fashion items of other categories that are matched with the fashion items in the recommended style image can be extracted as matching fashion items is that other items that are matched with the fashion items in the recommended style image match well. Because it is learned.

As described above, when a matching fashion item is determined using the style database 150, the fashion item extracting unit 110 may search for a matching fashion item based on the similarity of image contents in the product database 170.

Then, the fashion item extraction unit 110 determines an item corresponding to a preferred color, a preferred price range, and a brand extracted from the user database management unit 140 among matching fashion items as a recommended fashion item image.

The fashion item advertisement providing unit 120 generates a customized advertisement image by mapping the holding fashion item image and the recommended fashion item image extracted by the fashion item extracting unit 110 matching the holding fashion item image to the user image. Then, the fashion item advertisement providing unit 120 provides a user-customized advertisement image to the user terminal.

As described above, according to the present invention, a product may be advertised to a user by mapping it to a user image by using other fashion items that match the fashion item possessed by the user, thereby increasing the purchase rate of the product, and thus generating profits.

The user database 130 stores style information for each user (eg, size information, preferred color, preferred price range, brand, and information on the image of the fashion item possessed).

The user database manager 140 generates style information for each user and stores it in the user database 130. The user database management unit 140 provides a user-specific search history request message to the online shopping mall server 200, and upon receiving the user search history result message from the online shopping mall server 200, the user-specific style based on the user search history result message Information is generated and stored in the user database 130.

More specifically, the user database management unit 140 sends a search history request message for each user including user access information (eg, identification information of a user terminal, IP information of a user terminal, online shopping mall login information, etc.) Based on a user search history result message provided to the server 200 and including user purchase information and user search information from the online shopping mall server 200 (eg, products clicked on the online shopping mall and products contained in the shopping cart). Style information for each user is generated and stored in the user database 130. At this time, the user database management unit 140 may generate the user's size information, preferred color, preferred price range, and brand based on the user purchase information and the user search information, and then store it in the user database 130, based on the user search information. The user's search product may be extracted and stored in the user database 130.

In general, when a user purchases a product through an online shopping mall or puts a product of interest in the shopping cart, the user selects the size of the product as an option and then proceeds to purchase the product. There is color information and brand. Accordingly, the user database management unit 140 may determine the size information, preferred color, preferred price range, and brand of the user based on the user purchase information.

First, the user database management unit 140 generates user size information based on user purchase information and user search information, and then stores it in the user database 130.

In one embodiment, the user database management unit 140 analyzes user purchase information and user search information to extract size information for each fashion item category, then averages the size information to calculate average size information, and calculates the average size information as the size of the user. Information may be determined and stored in the user database 130.

In another embodiment, the user database management unit 140 analyzes user purchase information and user search information, extracts size information for each fashion item category, and then counts each size information to obtain the largest number of size information. It may be determined and stored in the user database 130.

In addition, the user database manager 140 may determine a user's preferred color based on user purchase information and user search information, and then store it in the user database 130.

In general, when a user purchases a product through an online shopping mall, a user selects a color of a specific product as an option and then proceeds to purchase a corresponding product. Accordingly, the purchase color information selected by the user exists in the user purchase information. Accordingly, the user database manager 140 may determine a user's preferred color based on user purchase information.

In one embodiment, the user database management unit 140 extracts purchase color information for each fashion item category based on user purchase information and user search information, and then uses the purchase color information to use the user's preference color. ).

In the above embodiment, after counting each purchase color information, the user database manager 140 may determine a priority for a user's preferred color based on the count value and store it in the user database 130.

In addition, the user database management unit 140 may determine the fashion item image to be held based on the user purchase information and the user search information and then store it in the user database 130. The reason for this is to recommend to the user an item that matches the fashion item possessed by the user.

In addition, the user database management unit 140 may extract a user's search product (eg, a clicked product and an image of a product contained in the shopping cart) based on the user search information and store it in the user database 130. The reason for this is that the user has not purchased, but clicked or put in a shopping cart is a fashion item that the user is interested in.

For the above reasons, the present invention can increase advertising efficiency by defining not only the product purchased by the user but also the product of interest as a holding fashion item, and advertising other fashion items matching the holding fashion item together.

The style database 150 may include information on a style image that can be referred to for a fashion style and coordination of a plurality of items among images collected on the web. The style database manager 160 creates a style image and stores it in the style database 150.

First, the style database management unit 160 collects a list of web addresses such as fashion magazines, fashion brands, drama production companies, celebrity agencies, SNS, online stores, etc., and includes them in the website by checking the website and tracking the link. Collected image information.

On the other hand, the style database management unit 160 may collect and index images from websites such as fashion magazines, fashion brands, drama production companies, celebrity agencies, SNS, online stores, etc., but collect image information together with index information from affiliated companies. It can also be provided separately.

In this case, the style database management unit 160 may filter out images unsuitable for style recommendation among the collected images. For example, the style database management unit 160 may filter the remaining images, leaving only images including a person-shaped object and a plurality of fashion items among the collected images.

Since the style image is used to determine other items that can be coordinated with the holding fashion item, it is appropriate to filter the image for a single fashion item. Furthermore, configuring a database with images of a person wearing a plurality of fashion items directly may be more useful than an image of the fashion item itself.

Therefore, the style database management unit 160 according to the embodiment of the present invention determines the style image included in the style database 150 by filtering the remaining images while leaving only the image including the human-shaped object and a plurality of fashion items. I can.

Thereafter, the style database manager 160 may process features of the fashion item object image included in the style image.

More specifically, the style database management unit 160 extracts the image feature information of the fashion item object included in the style image, and expresses the feature information as a vector value to generate a feature value of the fashion item object and structure the feature information of the images. can do.

Furthermore, the style database management unit 160 according to an embodiment of the present invention may extract a style label from the style image and cluster the style images based on the style label. It is appropriate that the style label is extracted as about the look and feel and trend of the fashion item's appearance and feel.

At this time, the style database management unit 160 defines a style label in advance, generates a neural network model that learns characteristics of an image corresponding to the style label, classifies objects in the style image, and extracts a style label for the object. can do. Then, the style database manager 160 may assign a corresponding style label to an image that matches a specific pattern with a random probability through a neural network model that has learned a pattern of an image corresponding to each style label.

According to another embodiment of the present invention, the style database management unit 160 forms an initial neural network model by learning characteristics of an image corresponding to each style label, and applies a large number of style image objects to it to further refine the neural network model. It can also be expanded.

Meanwhile, according to another embodiment of the present invention, the style database manager 160 may apply style images to a neural network model formed in a hierarchical structure formed of a plurality of layers without additional learning about a style label.

Furthermore, a weight is given to the feature information of the style image according to the request of the layer, and product images are clustered using the processed feature information, and the celebrity look, magazine look, summer look, feminine look, and sexy look are added to the clustered style image group. Style labels that are interpreted post-mortem as look, office look, drama look, and Chanel look can be given.

Meanwhile, the style database management unit 160 may pre-classify items having a very high appearance rate, such as white shirts, jeans, and black skirts.

For example, jeans are a basic item in fashion, so their appearance in style images is very high. Therefore, no matter what item the user inquires about, the probability of matching jeans as a coordination item will be much higher than that of other items.

Accordingly, the style database management unit 160 according to an embodiment of the present invention pre-classifies items with a very high appearance rate in the style image as buzz items, and changes the version of the style book to include a buzz item and a buzz item. Can be created by

According to another embodiment of the present invention, the buzz item may be classified by reflecting time information. For example, when considering the fashion cycle of fashion items, items that briefly become fashionable and disappear for a month or two, fashionable items that return every season, and items that continue to be fashionable for a certain period of time can be considered.

Therefore, by reflecting time information in the classification of buzz items, if a specific fashion item has a very high appearance rate during a certain period, the item can be classified as a buzz item together with information about the period. Classifying the buzz items as described above has an effect of being able to recommend the recommended item in consideration of whether the item to be recommended is in fashion or irrelevant to the fashion in a subsequent item recommendation step.

In addition, the style database management unit 160 may generate category information of a corresponding product based on the product image content. In this case, the style database management unit 160 may use some of the label information as category information.

For example, when the label for an arbitrary product image is extracted as a woman, top, blouse, linen, stripe, long sleeve, blue, and office look, the style database management unit 160 provides the service server for women, tops, and blouses. Labels can be used as information on the category of the product, and labels for linen, stripes, long sleeves, blue, and office look can be used as label information describing characteristics of products other than the category.

In the product database 170, product information for each product (eg, product origin, size, sales location, wearing shot, etc.) is correspondingly stored.

The product database management unit 180 receives product information for each product from the online shopping mall server 200 and stores product information for each product in the product database 170. At this time, the product information includes detailed product information such as the origin, size, sales place, and wearing shots of products sold in the online market.

To this end, the product database management unit 180 may include a crawler, a parser, and an indexer to collect web documents of an online store and access text information such as product images, product names, and prices included in the web documents.

For example, a crawler may collect a list of web addresses of online stores, check websites, and extract related product information by tracking links. At this time, the parser analyzes the web document collected during the crawling process to extract product information such as product images, product prices, and product names included in the page, and the indexer may index the corresponding location and meaning.

At this time, the product database management unit 180 may collect and index product information from a website of any online store, but may receive product information in a preset format from an affiliate market.

To this end, the product database management unit 180 structures text information such as a product name, product category, etc. in addition to the image of the product and stores it in the product database management unit 180.

In an embodiment, the product database management unit 180 may extract feature information of a product image and index feature information of the images for efficiency of search.

More specifically, the product database management unit 180 may detect characteristic areas of product images (Interest Point Detection). The feature region refers to a main region for extracting a feature descriptor, that is, a feature descriptor of an image for determining whether or not the images are identical or similar.

According to an embodiment of the present invention, such a characteristic area is a contour line included in an image, a corner such as a corner among the contours, a blob that is distinguished from the surrounding area, an area that is invariant or co-variable according to the transformation of the image, or more than the ambient brightness. They can be poles with dark or bright features, and can target patches (fragments) of the image or the entire image.

Furthermore, the product database management unit 180 may extract a feature descriptor from the feature region (Descriptor Extraction). A feature descriptor is a representation of the features of an image as vector values.

According to an embodiment of the present invention, such a feature descriptor may be calculated by using the location of the feature region for the corresponding image, or brightness, color, sharpness, gradient, scale, or pattern information of the feature region. For example, the feature descriptor may be calculated by converting a brightness value, a brightness change value, or a distribution value of the feature region into a vector.

Meanwhile, according to an embodiment of the present invention, a feature descriptor for an image is not only a local descriptor based on a feature region as described above, but also a global descriptor, a frequency descriptor, a binary descriptor, or It can be expressed as a neural network descriptor.

More specifically, the feature descriptor is a global descriptor that converts and extracts the brightness, color, sharpness, gradient, scale, pattern information, etc. of each of the entire image or the image divided by a random criterion, or Global descriptor) can be included. For example, the feature descriptor is a frequency descriptor that converts and extracts the number of times that specific descriptors classified in advance are included in an image, the number of global features such as a conventionally defined color table, etc., into vector values, and each descriptor In the layer of a binary descriptor or neural network that extracts bit by bit whether the values of elements are included or whether the size of each element value constituting the descriptor is larger or smaller than a specific value, converts it to an integer type and uses it. It may include a neural network descriptor for extracting image information used for learning or classification.

Furthermore, according to an embodiment of the present invention, a feature information vector extracted from a product image can be converted to a lower dimension. For example, feature information extracted through an artificial neural network corresponds to 40,000-dimensional high-dimensional vector information, and it is appropriate to convert it into a low-dimensional vector in an appropriate range in consideration of the resources required for search.

The feature information vector may be converted using various dimensional reduction algorithms such as PCA and ZCA, and the feature information converted into a lower dimensional vector may be indexed into a corresponding product image.

Unlike the above embodiment, the product database management unit 180 may extract a style label for the meaning of the image by applying a machine learning technique based on the product image. The style label may be expressed as an abstracted value, but may be expressed in a text form by interpreting the abstracted value.

In one embodiment, the product database management unit 180 defines a style label in advance, generates a neural network model that learns the characteristics of an image corresponding to the style label, classifies objects in the product image, and classifies the object in the style. Labels can be extracted. In this case, the product database management unit 180 may assign a corresponding style label to an image that matches a specific pattern with a random probability through a neural network model that has learned a pattern of an image corresponding to each style label.

In another embodiment, the product database management unit 180 forms an initial neural network model by learning the characteristics of an image corresponding to each style label, and applies a large number of product image objects to it to more elaborately extend the neural network model. May be. Further, the product database management unit 180 may create a new group including the product if the product is not included in any group.

In the above-described embodiment, the product database management unit 180 utilizes as meta information about products such as women's bottoms, skirts, dresses, short sleeves, long sleeves, patterns, materials, colors, and abstract feelings (innocent, chic, vintage, etc.) A possible style label can be defined in advance, a neural network model learning the characteristics of an image corresponding to the style label can be created, and the neural network model can be applied to the product image of an advertiser to extract a style label for an advertisement target product image. have.

In another embodiment, the product database management unit 180 may apply product images to a neural network model formed in a hierarchical structure formed of a plurality of layers without additional learning about a style label. Furthermore, a weight may be assigned to the feature information of the product image according to a request of a corresponding layer, and product images may be clustered using the processed feature information.

In the above embodiment, additional analysis may be required in order to check whether corresponding images are clustered according to which attribute of the feature value, that is, to connect the clustering result of the images with a concept that can be recognized by a real human being.

For example, the product database management unit 180 classifies products into three groups through image processing, and A for the features of the first group, B for the features of the second group, and C for the features of the third group. In the case of extracting the style label, A, B, and C need to be interpreted post-hoc that, for example, each means a woman's top, blouse, and checkered pattern.

Thereafter, the product database management unit 180 post-results to the clustered style image group with women's bottoms, skirts, dresses, short sleeves, long sleeves, patterns of patterns, materials, colors, abstract feelings (innocent, chic, vintage, etc.). A style label that can be interpreted may be assigned, and style labels assigned to a style image group to which an individual product image belongs may be extracted as a style label of a corresponding product image.

Meanwhile, the product database management unit 180 according to an embodiment of the present invention may express the style label extracted from the product image as text, and the style label in the form of text may be used as tag information of the product.

In the past, tag information of a product was subjectively directly assigned by a seller, so it was inaccurate and less reliable. Product tags subjectively given by sellers act as noise, thereby lowering the efficiency of search.

However, as in the embodiment of the present invention, when style label information is extracted based on a product image, and the extracted style label information is converted into text and used as tag information of a corresponding product, the tag information of the product is used as the image of the product. Based on the mathematical extraction without human intervention, the reliability of the tag information is increased and the accuracy of the search is improved.

In addition, the product database management unit 180 may generate category information of a corresponding product based on the product image content. In this case, style label information and category information may be respectively generated, but style label information may be used as category information, and category information may be used as style label information.

For example, when a style label for an arbitrary product image is extracted as a woman, top, blouse, linen, stripe, long sleeve, blue, office look, the product database management unit 180 is a style label for women, tops, and blouses. Is used as the category information of the corresponding product, and style labels for linen, stripes, long sleeves, blue, and office look can be used as style label information describing the characteristics of products other than the category. Alternatively, the product database management unit 180 may index a corresponding product without distinguishing between the style label and the category information.

Meanwhile, the product database management unit 180 filters product images and stores only product images in the product database 170 excluding portions that may act as noise for image search.

More specifically, the product database management unit 180 may determine whether to filter the product image by comparing the style label extracted from the product image with the category and/or tag information directly assigned by the seller.

According to an embodiment of the present invention, if there are multiple images for a specific product, and the style label extracted from one of the images and the category assigned by the seller for the product are different, the image or a specific object in the image is a product database It can be filtered at (170).

For example, there are first to third product images for product A, and in the first product image (women's top, jacket), in the second product image (women's top, jacket) and (sunglasses, round), 3 In the product image, a case in which the style label (sunglasses, round) is extracted can be considered. At this time, if the sales category of product A is “sunglasses”, the product database management unit 180 may configure the product database 170 with only the second and third product images excluding the first product image.

This filtering is to reduce the noise of image search. In the above example, if product A is actually for sunglasses, and if the database is configured including all the first to third product description images, even if the query image is a jacket, it is determined that it is similar to the first product image, so that product A for sunglasses is It may be determined as an advertisement item. Therefore, product images that may reduce the accuracy of the search are filtered and the product database 170 is built.

Referring to FIG. 2, the fashion item advertisement server 100 extracts a holding fashion item image from a user database in which style information for each user is stored (step S210).

The fashion item advertisement server 100 extracts a recommended style image matching the holding fashion item image from the style database in which the style image is stored (step S220).

The fashion item advertisement server 100 extracts an image of a product matching the recommended style image from a product database in which product information for each product sold in an online shopping mall is stored, and determines it as a recommended fashion item image (step S230).

The fashion item advertisement server 100 generates a customized advertisement image by mapping the recommended fashion item image and the holding fashion item image to the user image (step S240).

The fashion item advertisement server 100 provides a user-customized advertisement image to the user terminal (step S250).

Referring to FIG. 3, the fashion item advertisement server 100 extracts a holding fashion item image from a style database in which style information for each user is stored (step S310).

The fashion item advertisement server 100 extracts style label and category information by analyzing the held fashion item image (step S320).

In an embodiment, the fashion item extracting unit 110 may extract a style label from the held fashion item image and cluster the holding fashion item image based on the style label. It is appropriate that the style label is extracted as about the look and feel and trend of the fashion item's appearance and feel. Meanwhile, the fashion item extracting unit 110 may use style label information as category information.

The fashion item advertisement server 100 analyzes the possessed fashion item image and calculates a feature value (step S330).

The fashion item advertisement server 100 selects at least one style image group having a characteristic value of the fashion item image and a distance value within a certain range (step S340), and selects a style image belonging to the style image group as a recommended style image. It may be determined (step S350).

The fashion item advertisement server 100 extracts a matching fashion item of a category different from the holding fashion item from among fashion items included in the recommended style image (step S360), and The image may be extracted and determined as a recommended fashion item image (step S370).

As described above, although the present invention has been described by the limited embodiments and drawings, the present invention is not limited to the above embodiments, which is various modifications and variations from these descriptions to those of ordinary skill in the field to which the present invention belongs. Transformation is possible. Accordingly, the idea of the present invention should be grasped only by the scope of the claims set forth below, and all equivalent or equivalent modifications thereof will be said to belong to the scope of the idea of the present invention.

Explanation of sign

100: fashion item advertisement server

110: fashion item extraction unit

120: Fashion item advertisement provider

130: user database

140: user database management unit

150: style database

160: style database management unit

170: product database

180: product database management unit

200: online shopping mall server

300: user terminal

Claims

A user database storing style information for each user;

A style database in which style images are stored;

A product database storing product information for each product sold in an online shopping mall;

A recommended fashion item image by extracting a fashion item image from the user database, extracting a recommended style image matching the fashion item image from the style database, and extracting an image of a product matching the recommended style image from the product database A fashion item extracting unit determined to be; And

And a fashion item advertisement providing unit for generating a customized advertisement image by mapping the recommended fashion item image and the holding fashion item image to a user image, and providing the user customized advertisement image to a user terminal.

Fashion item advertisement server.
The method of claim 1,

Analyzing the holding fashion item image to extract style label and category information, selecting at least one style image group having a characteristic value of the holding fashion item image and a distance value within a certain range, and a style image belonging to the style image group Characterized in that determining the recommended style image

Fashion item advertisement server.
The method of claim 2,

The fashion item extraction unit

Among the fashion items included in the recommended style image, a matching fashion item of a category different from the holding fashion item is extracted, and an image of a product corresponding to the matching fashion item from among the products stored in the product database is extracted as a recommended fashion item image Characterized by determining

Fashion item advertisement server.
The method of claim 3,

Among the matching fashion items, an item corresponding to a preferred color, preferred price point, and brand extracted from the user database is determined as a recommended fashion item image.

Fashion item advertisement server.
In the user-customized fashion item advertisement method executed in the fashion item advertisement server,

Extracting an image of a fashion item held by a style database in which style information for each user is stored;

Extracting a recommended style image matching the holding fashion item image from a style database in which a style image is stored;

Extracting an image of a product matching the recommended style image from a product database in which product information for each product sold in an online shopping mall is stored, and determining a recommended fashion item image;

Generating a customized advertisement image by mapping the recommended fashion item image and the holding fashion item image to a user image; And

And providing the user-customized advertisement image to a user terminal.

How to advertise custom fashion items.
The method of claim 5,

Extracting the recommended style image matching the holding fashion item image

Analyzing the holding fashion item image to extract style label and category information;

Analyzing the holding fashion item image to calculate a feature value;

Selecting at least one style image group having a characteristic value of the fashion item image and a distance value within a predetermined range; And

And extracting a style image belonging to the style image group as a recommended style image.

How to advertise custom fashion items.
The method of claim 6,

The step of extracting an image of a product matching the recommended style image and determining it as a recommended fashion item image

Extracting a matching fashion item of a category different from the holding fashion item from among fashion items included in the recommended style image; And

And extracting an image of a product corresponding to the matching fashion item from among the products stored in the product database and determining a recommended fashion item image.

How to advertise custom fashion items.
The method of claim 6,

The step of extracting an image of a product matching the recommended style image and determining it as a recommended fashion item image

And determining an item corresponding to a preferred color, preferred price point, and brand extracted from the user database among the matching fashion items as a recommended fashion item image.

How to advertise custom fashion items.