KR20240028655A - Apparatus and method for recommending apparel styling - Google Patents

Abstract

This relates to a style recommendation device and method. The style recommendation device and method generate user attributes based on answers written by a user based on a survey, and classify a plurality of stored styles into clothing types appropriate for the situation. You can filter using a dictionary. In addition, the style recommendation device and method calculates the similarity of the filtered results by comparing whether they are similar to user attributes, filters them using this, delivers the results to SNS acquaintances, receives rankings and opinions, and generates and scores evaluation weights. It may include a device and a method for providing a recommended style to a user.

Description

Style recommendation device and method {APPARATUS AND METHOD FOR RECOMMENDING APPAREL STYLING}

This is a device and method that recommends a style suitable for the user based on the user's input information.

As online shopping has become a common consumption pattern and meta-shopping mall services that collect and compare products from multiple shopping malls are increasing, consumers are showing a tendency to obtain various clothing information online and purchase products that suit them personally.

As the number of users of diverse social network services (SNS) expands, the role and influence of individuals participating in the network are increasing. Additionally, unlike other industries, the fashion industry is easily influenced by social awareness, so the influence of influencers with a large number of followers is clearly increasing.

Conventional fashion style recommendation has the limitation of recommending clothing with similar shapes and colors based on information entered by the user, or only recommending products in the same category as the style that the user usually wears. Clothing styling services that utilize users' personal networks are lacking.

Republic of Korea Patent Publication No. 2020-0111591 (“Clothing size recommendation method considering body measurements and clothing data”, Na Woo-sik, 2020.09.29)

The purpose is to provide a style recommendation device and method that provides recommended styles to users based on user attributes, clothing wearing situations, and SNS feedback.

One embodiment of the style recommendation method includes generating user attributes based on answers to an inputted user questionnaire, and generating a candidate style list by filtering a previously stored style database based on a situational clothing dictionary in which appropriate clothing types for each situation are specified. calculating the similarity between the style attributes of the candidate style list and the user attributes, generating a user style list based on the calculated similarity, and delivering styles included in the user style list to the user's SNS acquaintances. Receiving rankings and opinions of styles, scoring the received rankings and opinions to generate evaluation weights for styles included in the style list, reflecting the evaluation weights in the user style list to provide the information to the user. It may include determining a recommended style to be proposed, extracting an image of the recommended style, cropping or resizing the image to fit the user's body type among the user attributes, and providing the image to the user.

In addition, according to one embodiment, the step of generating the candidate style list includes comparing the appropriate clothing type of the situation clothing dictionary with the clothing type of the style included in the style database, and matching the clothing type in the style database is preset. A step of filtering out styles that are below the standard may be further characterized.

In addition, according to one embodiment, the step of generating the candidate style list includes determining a clothing type that matches the user's situation by searching the situational clothing dictionary, and comparing the clothing type constituting the style with the determined clothing type. Calculating the situational suitability of each style and generating the candidate style list based on the situational suitability may be further characterized.

One embodiment of the style recommendation device includes a communication unit that receives user answers or transmits and receives data generated in the process to the server, a storage unit that includes a style database and a situational clothing dictionary, and a user analysis unit that generates the user answers as user attributes. A filtering unit for generating a candidate style list by filtering the style database based on the situational clothing dictionary, calculating a similarity between the candidate style list and the user attributes, and generating a user style list based on the calculated similarity. Recommendation unit, delivers the styles included in the user's style list to the user's SNS acquaintances, receives rankings and opinions of the styles, scores the received rankings and opinions, and determines the evaluation weight of the styles included in the style list and a feedback recommendation unit that reflects the evaluation weight in the user style list to determine a recommended style to suggest to the user, extracts an image of the recommended style, and crops or resizes the image to fit the user's body type among the user attributes. It may be modified to include an extraction unit provided to the user.

In addition, according to one embodiment, the filtering unit compares the appropriate clothing type of the situational clothing dictionary with the clothing type of the style included in the style database, and filters styles in the style database in which the matching degree of clothing type is less than a preset standard. It may be characterized as a decision to further include.

In addition, according to one embodiment, the filtering unit searches the situational clothing dictionary to determine a clothing type that matches the user's situation, and compares the clothing type constituting the style with the determined clothing type to calculate the situational suitability of each style. And, the candidate style list may be generated based on the situation suitability.

The style recommendation device and method according to an embodiment of the present invention can generate user attributes based on a user survey, select styles matching the user attributes, and provide the styles to the user.

In addition, the style recommendation device and method according to an embodiment of the present invention can provide the user with a style optimized for the situation by providing styles using a situational clothing dictionary in which appropriate clothing types are stored according to the clothing wearing situation. .

In addition, the style recommendation device and method according to an embodiment of the present invention can provide the user with a style that reflects a plurality of evaluation opinions by determining a style to match the user based on the feedback results of SNS acquaintances.

Additionally, the style recommendation device and method according to an embodiment of the present invention can provide the optimal style to the user by recommending a style to the user by comprehensively reflecting user attributes, clothing wearing situation, and SNS feedback.

1 is a style recommendation system according to an embodiment.
Figure 2 is a detailed block diagram of a style recommendation device according to an embodiment.
Figure 3 is an example diagram of a user survey according to an embodiment.
Figure 4 is an example of a style database according to one embodiment.
Figure 5 is an example of a user body type decision tree according to one embodiment.
Figure 6 is an example of generating a candidate style list using a situational clothing dictionary according to an embodiment.
Figure 7 is an example of creating a user style list using the similarity between user attributes and style attributes according to an embodiment.
Figure 8 is an example of determining a recommended style using opinions and feedback received from SNS acquaintances according to an embodiment.
Figure 9 is an example of providing a recommended style to a user according to an embodiment.
Figure 10 is a flowchart for explaining a style recommendation method according to an embodiment.

The advantages and features of the present invention and methods for achieving them will become clear with reference to the embodiments described below along with the accompanying drawings. However, the present invention is not limited to the embodiments disclosed below and may be implemented in various different forms. The present embodiments are merely provided to ensure that the disclosure of the present invention is complete and to be understood by those skilled in the art. It is provided to fully inform those who have the scope of the invention, and the present invention is only defined by the scope of the claims.

The terms used in this specification will be briefly explained, and the present invention will be described in detail.

The terms used in the present invention are general terms that are currently widely used as much as possible while considering the function in the present invention, but this may vary depending on the intention or precedent of a person working in the art, the emergence of new technology, etc. In addition, in certain cases, there are terms arbitrarily selected by the applicant, and in this case, the meaning will be described in detail in the description of the relevant invention. Therefore, the terms used in the present invention should be defined based on the meaning of the term and the overall content of the present invention, rather than simply the name of the term.

When it is said that a part "includes" a certain element throughout the specification, this means that, unless specifically stated to the contrary, it does not exclude other elements but may further include other elements. Additionally, terms such as “unit,” “module,” and “unit” used in the specification refer to a unit that processes at least one function or operation, and include software, hardware components such as FPGA or ASIC, or software and hardware. It can be implemented by combining . However, terms such as “part”, “module”, and “unit” are not limited to software or hardware. “Parts”, “modules”, “units”, etc. may be configured to reside in an addressable storage medium and may be configured to reproduce one or more processors 300. Accordingly, as an example, terms such as "part", "module", and "unit" refer to components such as software components, object-oriented software components, class components, and task components, and processes. Contains fields, functions, properties, procedures, subroutines, segments of program code, drivers, firmware, microcode, circuits, data, databases, data structures, tables, arrays, and variables. .

Below, with reference to the attached drawings, embodiments of the present invention will be described in detail so that those skilled in the art can easily implement the present invention. In order to clearly explain the present invention in the drawings, parts unrelated to the description are omitted.

Terms containing ordinal numbers, such as “first,” “second,” etc., may be used to describe various components, but the components are not limited by the terms. Terms are used only to distinguish one component from another. For example, a first component may be named a second component, and similarly, the second component may also be named a first component without departing from the scope of the present invention. The term “and/or” includes a combination of a plurality of related items or any one item among a plurality of related items.

Hereinafter, an embodiment of the style recommendation device 1 will be described with reference to the attached drawings.

Figure 1 shows the overall system of the style recommendation system 1.

The style recommendation system 1 may include a user terminal 20. Here, the user terminal 20 can receive answers to user questionnaires, user SNS information, etc., and transmit them to the style recommendation device 10. Additionally, the user terminal 20 may receive a user style list, recommended styles, etc. for SNS acquaintance voting from the style recommendation device 10 and output them.

The style recommendation system 1 may include a style recommendation device 10.

The style recommendation device 10 may receive user information input from the user terminal 20 through a network, and provide a style most suitable to the user's taste and situation among pre-stored styles based on the user information. Hereinafter, the style recommendation device 10 will be described in detail with reference to the drawings.

Figure 2 shows a block diagram of the style recommendation device 10.

Referring to FIG. 2 , the style recommendation device 10 may include a communication unit 100, a storage unit 200, and a processor 300.

As shown in FIG. 3, the communication unit 100 receives input from the user terminal 20 to generate user attributes of the style recommendation device 10, such as user gender, body size, weight, price range for clothing size, and body complex. , information on budget costs and necessary situations can be received, and the received information can be transmitted to the storage unit 200 and the processor 300.

The communication unit 100 may receive SNS information used by the user, SNS follower and SNS acquaintance information, etc. to generate evaluation weights.

The communication unit 100 receives at least one command for starting the operation of the style recommendation device and method (e.g., a command for running a styling program (which may be referred to as an app, application, or software, etc.)), This may be transmitted to the storage unit 200 and the processor 300.

Depending on the embodiment, the communication unit 100 may receive the above-described information by transmitting the above-described information through a wired or wireless communication network. If necessary, the communication unit 100 can also transfer the received data to the storage unit 200, and in this case, the storage unit 200 can temporarily or non-temporarily store these data.

The communication unit 100 may transmit the recommended style image to the user terminal 20.

The communication unit 100 may provide the above-described information to the user in a visual or auditory form, or may transmit the above-described information to another device through a communication module or a wired or wireless communication network.

The storage unit 200 temporarily or It can be stored non-temporarily.

The storage unit 200 can store an operating system, applications, etc. necessary for the operation of the style recommendation device 10, and can store various data generated during the operation of the style recommendation device 10. For example, the storage unit 200 may store a candidate style list, a user style list, and a recommended style.

The storage unit 200 may include a style database 210.

Style database 210 includes style attributes. At this time, the style stored in the style database 210 consists of a plurality of clothing, and the style attribute consists of information necessary for style matching.

As shown in FIG. 4, style attributes may include information such as style image, suitable gender, preferred age, preferred body type, body complement, and style price range that constitute the style. The style database 210 is not limited to the attributes shown in FIG. 4 and may be an example.

Style attributes may be created for a style recommendation device or may be created by analyzing information about a style modeled by an expert.

Style attributes can be stored in the form of a style attribute vector as shown in Table 1. To this end, the processor 300 extracts key words by performing morphological analysis and stopword removal, which are natural language processing processes, on the style attributes, and stores the style attributes in the form of a style attribute vector as shown in Table 1 depending on the presence or absence of the word. .

In other words, the style attribute vector displays the style attribute by matching 1 if the style attribute is applicable, and 0 if it is not applicable.

By managing style attributes in the form of style attribute vectors in this way, processing efficiency, such as filtering and user matching, of styles stored in the style database 210 can be improved.

Style ID Preferred age
20's Preferred age
30s Preferred body type: hourglass shape Style_Basic Style_Chic … AttributeN-1 AttributeN One One 0 One One 0 … 0 One 2 0 One 0 0 One … One One 3 One 0 One 0 0 … 0 0 … … … … … … … … … i One 0 One 0 One One 0

The storage unit 200 may include at least one of a main memory and an auxiliary memory. For example, the main memory may be implemented using a semiconductor storage medium such as ROM and/or RAM, and the auxiliary memory may be a flash memory device (Solid State Drive (SSD) ), etc.), SD (Secure Digital) cards, hard disk drives (HDD), compact disks, DVDs, or laser disks, etc. can be implemented based on devices that can store data permanently or semi-permanently. .

The processor 300 may include a user analysis unit 310, a filtering unit 320, a style recommendation unit 330, a feedback recommendation unit 340, and an extraction unit 350.

The processor 300 may execute a program stored in the storage unit 200 to execute at least one operation. The processor 300 may include, for example, a central processing unit (CPU), a graphics processing unit (GPU), a microcontroller unit (MCU), and an application processor (AP). ), an electronic control unit (ECU), a microprocessor (Micom), and/or at least one electronic device capable of performing various calculations and control processing. These processing or control devices may be implemented, for example, by using one or more semiconductor chips, circuits, or related components alone or in combination.

Hereinafter, the user analysis unit 310 will be described with reference to FIG. 5.

The user analysis unit 310 may perform morpheme analysis and stopword removal, which are natural language processing processes, on the input answer to extract key words and generate user attributes.

The user analysis unit 310 may store user attributes in the form of a user attribute vector expressed as a vector as shown in Table 2, depending on the presence or absence of words included in the user attributes.

In other words, the user attribute vector displays the user attribute by matching 1 if the user attribute is applicable, and 0 if it is not applicable.

In this way, by managing user attributes in the form of user attribute vectors, processing efficiency, such as calculating similarity between user attribute vectors and stored style attribute vectors, and determining body shape, can be improved.

User ID 20's inverted triangle triangle chic lovely … AttributeN-1 AttributeN One One 0 One One 0 … 0 One 2 0 One 0 0 One … One One 3 One 0 One 0 0 … 0 0 … … … … … … … … … I One 0 One 0 One One 0

The user analysis unit 310 may include user attribute vectors for a plurality of users in the form of a vector as shown in Table 2. The user analysis unit 310 may determine the user's body type based on the body size input by the user. there is. The user analysis unit 310 may analyze the user's body shape based on the size correlation between chest circumference, hip circumference, and waist circumference input by the user.

In one embodiment, the user analysis unit 310 uses a body shape decision tree to determine the user's body type, as shown in FIG. 5. If "|chest circumference-hip circumference|>5", if chest circumference>hip circumference, the user is classified into an inverted triangle body shape. In the case of chest circumference ≤ hip circumference, the user can be determined to have a triangular body shape.

The body shape decision tree of the user analysis unit 310 can determine the user as having an hourglass body shape if "|chest circumference-hip circumference|≤5", waist circumference<hip circumference, waist circumference/hip circumference<0.7, and waist circumference When circumference/butt circumference ≥0.7, the user can be determined to have a straight body type.

The body shape decision tree of the user analysis unit 310 may determine that the user has a round body type if "|chest circumference - hip circumference|≤5" and waist circumference ≥ hip circumference.

When determining the user's body type, the user analysis unit 310 is not limited to the formula or standard for determining the user's body type of the user decision tree using "chest circumference, hip circumference, and waist circumference" such as an embodiment of the body type decision tree. In addition, it may include other measurements, formulas, and the user's body type determined based on the user's answers.

Hereinafter, the filtering unit 320 will be described with reference to FIG. 6.

The filtering unit 320 may filter the style database 210 to generate a candidate style list.

The candidate style list includes styles corresponding to the user's situation, and can be created by filtering the styles included in the style database 210 using a situational clothing dictionary.

The filtering unit 320 may generate a candidate style list including styles matching the user's situation based on the situational clothing dictionary.

The situational clothing dictionary is data in which experts classify clothing types appropriate for each situation. As shown in Figure 6, the situational clothing dictionary includes situations such as going to work, dating, outings, interviews, etc. and information on clothing types appropriate for each situation. You can.

The filtering unit 320 may filter the style database 210 using situational suitability.

The filtering unit 320 may search the situational clothing dictionary to determine a clothing type that matches the user's situation, and calculate the situational suitability of each style by comparing the determined clothing type with the clothing type constituting the style. For example, situational suitability can be calculated according to Equation 1.

Situational suitability can be calculated by comparing the clothing types stored in the situational clothing dictionary with the types of clothing that make up the style.

Situational suitability is a standard that allows you to know how many types of clothing are included in a style appropriate for a situation, based on the total number of clothing styles for a certain situation in the situational clothing dictionary.

Referring to Figure 6, for example, if the user's situation is dating, the clothing types suitable for the dating situation in the situation clothing dictionary determine the clothing types suitable for the dating situation: shirts, knits, skirts, sneakers, dresses, and the clothing that makes up each style. Situational suitability is calculated by comparing the type and suitable clothing type.

In Style 1, which consists of three items: a shirt, a skirt, and shoes, the shirt and skirt are clothing types that are appropriate for a dating situation, and the shoes are a clothing type that is not appropriate for a dating situation. Therefore, the total number of clothes that make up the style of Style 1 is 3, and the number of situation-appropriate style clothes suitable for the situation is 2. Therefore, the situational suitability of Style 1 is * Calculated as 100, the situational suitability can be calculated as a value of 67%.

The filtering unit 320 may further filter the candidate style list filtered for the same situation using the situation suitability based on the situation suitability of the score preset by the user.

The filtering unit 320 may compare the appropriate clothing type in the situational clothing dictionary with the clothing type of the style included in the style database, and may include a style filtering step in which the match between the clothing types in the style database is less than a preset standard.

For example, in the candidate style list containing 5 styles filtered by dating situation, if the user wants a style with high situational suitability even for the same dating situation, filter by the situational suitability criterion of 67% or more preset by the user and select that standard. A candidate style list containing three styles that satisfy may further be included.

Hereinafter, the style recommendation unit 330 will be described with reference to FIG. 7 .

The style recommendation unit 330 may calculate the similarity by comparing the style attributes of the styles in the candidate style list filtered by situation with the user's user attributes.

As an example of similarity calculation, the style recommendation unit 330 may calculate a similarity value through the inter-vector cosine similarity formula of Equation 2 between the stored style attribute vector and the user attribute vector.

(A: user attribute vector of user 1, B: style attribute vector of each style)

For example, the similarity is a value calculated through Equation 1 between the user attribute vector of user 1 and the style attribute vectors of styles 1 to 3 in the candidate style list. The similarity has a value in the range of 0 to 1 and is close to 1. This may mean that it is a style attribute vector similar to the user attribute vector of user 1.

The style recommendation unit 330 may generate a user style list including a plurality of styles by filtering based on a similarity value preset by the user.

Hereinafter, the feedback recommendation unit 340 will be described with reference to FIG. 8.

The feedback recommendation unit 340 may deliver the user style list to the terminal of the user registered as the user's SNS acquaintance. To this end, the user terminal 20 may transmit the SNS address or access link of the user's SNS acquaintance to the feedback recommendation unit 340.

The user's SNS acquaintances can input the style ranking and opinion of the user style list through each user terminal 20.

The feedback recommendation unit 340 may generate evaluation weights by analyzing style rankings and opinions received from the user's SNS acquaintances.

The feedback recommendation unit 340 may receive the style ranking of the user style list, score the styles in order of ranking, and calculate the ranking score. For example, as shown in Figure 8, if SNS acquaintance 1 determines style 1 as first priority, the ranking score can be calculated as +2 points, and if style 3 is decided as second priority, +1 point can be calculated.

The feedback recommendation unit 340 extracts key words by performing morphological analysis and stopword removal, which are natural language processing processes, on the received opinions, and according to the presence or absence of the word, as shown in Table 3, with a value of 1 if it exists and 0 if it does not exist. An opinion vector can be created.

The opinion vector is the result of natural language processing of the received opinion text to calculate the positive score of the feedback recommendation unit 340 and stored as a value of 0 or 1, and is the opinion received through a sentiment classification process using the opinion vector. can be scored.

The positive score can be defined as the sum of the opinion vectors through sentiment classification, scoring +1 if positive and -1 if negative.

feedback number skirt point necklace accessory not really bad overall cute It's not right Feedback 1 One One One One One 0 0 0 0 Feedback 2 0 0 0 0 0 One One One 0 Feedback 3 0 0 0 0 0 0 0 0 One Feedback 4 0 0 0 0 0 0 0 0 0 … … … … … … … … … … … 2 3 3 One One 2 One One One

The feedback recommendation unit 340 may include positive or negative sentiment analysis of the opinion vector in Table 4 using the existing defined sentiment classification in Table 4.

Sentiment classification related words Positive It suits, it fits, it’s cute, it’s suitable, it has personality, it’s cool, it’s pretty… denial It doesn't suit me, it doesn't suit me, it's too much, it doesn't fit, it's plain, I wish I could change it...

Sentiment classification can be composed of a database of words representing positive and negative, and can be created by classifying evaluated opinions about style into positive and negative, and storing related words corresponding to each positive and negative. Based on the emotional classification in Table 4, the feedback recommendation unit 340 can calculate a positive score of +1 if the opinion vector is classified as positive, such as "suitable, correct, cute, appropriate, unique, cool, pretty, etc." , "Not very good, doesn't suit, is excessive, doesn't fit, is plain, etc." If the content is classified as negative, it can be calculated as a positive score of -1 and added together to calculate a positive score in the evaluation weight.

For example, the feedback recommendation unit 340 may calculate positive score values such as +7.5 for style 1 and +5 for style 3 in FIG. 8 and include them in the evaluation weight.

The evaluation weight of the feedback recommendation unit 340 includes a style score that is the sum of the ranking score and the positive score, and the feedback recommendation unit 340 may determine the highest style in the user style list as the recommended style based on the style score.

The extraction unit 350 will be described with reference to FIG. 9 .

The extraction unit 350 extracts an image of a recommended style and crops or resizes the image to fit the user's body shape determined through the body shape decision tree of the user analysis unit 310 to create a image to fit the user's body type as shown in FIG. 9. there is.

For example, if the user has an inverted triangle body type, the extractor 350 may generate a recommended style image by enlarging the upper part of the image to fit the inverted triangle body type compared to the straight body type, as shown in FIG. 9. .

The style recommendation device 10 may transmit and output the generated recommended style image to the user terminal 20 through a network.

Figure 10 is a flowchart for explaining a style recommendation method according to an embodiment.

According to the style recommendation method according to an embodiment of the present invention, the style recommendation device 10 generates user attributes based on the user's questionnaire (S100).

According to the style recommendation method according to an embodiment of the present invention, the style recommendation device 10 generates a candidate style list by filtering the style database 210 based on the situational clothing dictionary (S200).

According to the style recommendation method according to an embodiment of the present invention, the style recommendation device 10 calculates the similarity between user attributes and style attributes and reflects this to generate a user style list (S300).

According to the style recommendation method in one embodiment of the present invention, the style recommendation device 10 generates an evaluation weight by scoring the rankings and feedback received from SNS acquaintances, and determines a recommended style by reflecting the scores (S400).

According to the style recommendation method according to an embodiment of the present invention, the style recommendation device 10 outputs a recommended style tailored to the user's body type (S500).

Those skilled in the art related to the embodiments of the present invention will understand that the above-described material may be implemented in a modified form without departing from its essential characteristics. Therefore, the disclosed methods should be considered from an illustrative rather than a restrictive perspective. The scope of the present invention is indicated in the claims, not the detailed description of the invention, and all differences within the equivalent scope should be construed as being included in the scope of the present invention.

1: Style recommendation system 320: Filtering unit
10: Style recommendation device 330: Style recommendation unit
20: User terminal 340: Feedback recommendation unit
100: Communication Department 350: Extraction Department
200: storage unit
210:Database
300: Processor
310: User analysis department

Claims (6)

Creating user properties based on answers to the input user questionnaire;
generating a candidate style list by filtering a previously stored style database based on a situational clothing dictionary in which appropriate clothing types according to the situation are specified;
calculating a similarity between style attributes of the candidate style list and the user attribute, and generating a user style list based on the calculated similarity;
delivering styles included in the user style list to the user's SNS acquaintances and receiving rankings and opinions of the styles;
generating evaluation weights for styles included in the style list by scoring the received rankings and opinions;
determining a recommended style to suggest to the user by reflecting the evaluation weight in the user style list; and
Extracting an image of the recommended style, cropping or resizing the image to fit the user's body type among the user attributes, and providing the image to the user:
A style recommendation method that includes . According to paragraph 1,
The step of generating the candidate style list is,
A style recommendation method further comprising comparing the appropriate clothing type in the situational clothing dictionary with the clothing type of the style included in the style database, and filtering out styles in the style database whose matching degree of clothing type is less than a preset standard. According to paragraph 1,
The step of generating the candidate style list is,
The situational clothing dictionary is searched to determine a clothing type that matches the user's situation, the clothing types constituting the style are compared with the determined clothing type to calculate the situational suitability of each style, and the candidate is based on the situational suitability. A style recommendation method further comprising: generating a style list. A communication unit that receives user responses or transmits and receives data generated in the process to and from the server;
a repository containing a style database and a situational clothing dictionary;
a user analysis unit that generates the user answers as user attributes;
a filtering unit generating a candidate style list by filtering the style database based on the situational clothing dictionary;
a style recommendation unit that calculates a similarity between the candidate style list and the user attributes and generates a user style list based on the calculated similarity;
Deliver the styles included in the user style list to the user's SNS acquaintances, receive rankings and opinions of the styles, score the received rankings and opinions to determine the evaluation weight of the styles included in the style list, and a feedback recommendation unit that determines a recommended style to suggest to the user by reflecting the evaluation weight in a user style list;
A style recommendation device comprising: an extraction unit that extracts the image of the recommended style, crops or resizes the image to fit the user's body type among the user attributes, and provides the image to the user. According to paragraph 4,
The filtering unit,
Comparing the appropriate clothing type in the situational clothing dictionary with the clothing type of the style included in the style database, and filtering out styles in the style database whose matching degree of clothing type is less than a preset standard. According to paragraph 4,
The filtering unit,
The situational clothing dictionary is searched to determine a clothing type that matches the user's situation, the clothing types constituting the style are compared with the determined clothing type to calculate the situational suitability of each style, and the candidate is based on the situational suitability. A style recommendation device that creates a style list.