WO2021040256A1 - Dispositif électronique et son procédé de recommandation de vêtements - Google Patents

Dispositif électronique et son procédé de recommandation de vêtements Download PDF

Info

Publication number
WO2021040256A1
WO2021040256A1 PCT/KR2020/010208 KR2020010208W WO2021040256A1 WO 2021040256 A1 WO2021040256 A1 WO 2021040256A1 KR 2020010208 W KR2020010208 W KR 2020010208W WO 2021040256 A1 WO2021040256 A1 WO 2021040256A1
Authority
WO
WIPO (PCT)
Prior art keywords
clothing
feed
artificial intelligence
electronic device
combination
Prior art date
Application number
PCT/KR2020/010208
Other languages
English (en)
Korean (ko)
Inventor
이형동
Original Assignee
삼성전자주식회사
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 삼성전자주식회사 filed Critical 삼성전자주식회사
Publication of WO2021040256A1 publication Critical patent/WO2021040256A1/fr

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
    • G06Q50/10—Services
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/08—Learning methods

Definitions

  • the present disclosure relates to an electronic device and a method of recommending clothing thereof, and more particularly, to an electronic device capable of recommending various styles of clothing according to a user, and a method of recommending clothing thereof.
  • artificial intelligence systems are used in various fields. Unlike existing rule-based smart systems, artificial intelligence systems are systems where machines learn, judge, and become smarter. As the artificial intelligence system is used, the recognition rate improves and the user's taste can be understood more accurately, and the existing rule-based smart system is gradually being replaced by a deep learning-based artificial intelligence system.
  • the present disclosure was devised in accordance with the above-described necessity, and the object of the present disclosure is to obtain combinations of clothes of various styles for the input clothing items using an artificial intelligence model learned for each feed provider, and among the obtained clothing combinations. It is to provide an electronic device and a method of recommending clothes thereof, which can recommend different styles of clothing for each user.
  • An electronic device includes a memory including at least one instruction, a processor connected to the memory to control the electronic device, and the processor, by executing the at least one instruction,
  • the input image is input to an artificial intelligence model learned based on a plurality of combination images obtained from a plurality of feed providers and a plurality of clothing items are combined.
  • Acquiring a plurality of combinations of clothes corresponding to each of the plurality of feed providers including clothing items included in an image, and clothing items included in at least one clothing combination among the plurality of clothing combinations Provides information about.
  • the electronic device may output a plurality of clothing combinations for an input clothing item to provide various styles of clothing combinations, and may recommend personalized clothing items based on user selection or usage history information. There will be.
  • FIG. 1 is a diagram illustrating an example of using an electronic device according to an embodiment of the present disclosure
  • FIG. 2 is a diagram for explaining a simple configuration of an electronic device according to an embodiment of the present disclosure
  • FIG. 3 is a block diagram for explaining a specific configuration of the electronic device disclosed in FIG. 2;
  • FIG. 4 is a view for explaining a clothing recommendation system according to an embodiment of the present disclosure.
  • FIG. 5 is a diagram illustrating a method of acquiring input data of an artificial intelligence model according to an embodiment of the present disclosure
  • FIG. 6 is a diagram for explaining various learning methods of an artificial intelligence model
  • FIG. 7 is a diagram for explaining various learning methods of an artificial intelligence model
  • FIG. 8 is a diagram for explaining various learning methods of an artificial intelligence model
  • FIG. 9 is a block diagram illustrating an output of an artificial intelligence model learned according to an embodiment of the present disclosure.
  • FIG. 10 is a diagram illustrating a clothing recommendation user interface (UI) screen of an electronic device according to an embodiment of the present disclosure
  • FIG. 11 is a diagram illustrating a clothing recommendation user interface (UI) screen of an electronic device according to an embodiment of the present disclosure
  • FIG. 12 is a flowchart illustrating a method of recommending clothes by an electronic device according to an embodiment of the present disclosure
  • FIG. 13 is a sequence diagram illustrating an embodiment of recommending clothing according to a user's selection of a feed provider
  • FIG. 14 is a sequence diagram illustrating an embodiment of recommending clothing according to usage history information of a user terminal device
  • 15 is a sequence diagram illustrating an embodiment in which an electronic device downloads a learned artificial intelligence model from an external server
  • 16 is a diagram for explaining a simple configuration of a server according to an embodiment of the present disclosure.
  • 17 is a sequence diagram illustrating an embodiment of a server that recommends clothing according to a user's selection of a feed provider.
  • FIG. 1 is a diagram illustrating an example of using an electronic device according to an embodiment of the present disclosure.
  • the electronic device 100 may include a learned artificial intelligence model that outputs a plurality of clothing combinations.
  • the clothing item may mean a top, a bottom, shoes, a bag, or an accessory.
  • the clothing combination refers to a combination of a plurality of clothing items, and refers to a combination of tops, bottoms, shoes, bags, accessories, and the like.
  • the learned artificial intelligence model may be learned to periodically receive feeds from a plurality of feed providers and output clothing combinations corresponding to each feed provider.
  • the artificial intelligence model may be learned by performing multi-task learning (MTL).
  • Multi-task learning is a method of improving prediction performance by simultaneously learning each task, and the task may mean a set of samples observed in the same environment.
  • multi-task learning when tasks are tasks having high similarity to each other, learning performance may be improved.
  • feeds of brands having similar styles to each other may be grouped into one task to perform learning.
  • Multi-task learning is designed to maximize the performance from a generalization point of view while guaranteeing the performance of the model itself for data collected from various tasks. You can consider both performance. In general, it is necessary to train a model for each task, and sufficient data to train each model is required. In the case of multi-task learning, by sharing a representation between each task, it is possible to exhibit superior performance than each trained model. A detailed learning method of an artificial intelligence model through multi-task learning will be described in more detail below with reference to FIGS. 5 to 8.
  • the feed provider may be a server of each clothing manufacturer, an SNS server of a celebrity, a server of a fashion magazine company, or the like.
  • the feed provided from the feed provider may be a combined image including a plurality of clothing items.
  • the combined image may mean an image including all of a top, a bottom, shoes, a bag, and accessories.
  • the combination image may refer to a photo in which a plurality of clothing items is worn on a person or includes a plurality of clothing items, such as a pictorial photo or styling photo of a clothing manufacturer, and a daily photo of a celebrity.
  • the electronic device 100 may be a clothes manager, as illustrated in FIG. 1, and may be implemented as a smart phone, a refrigerator, an air conditioner, a TV, or the like.
  • the electronic device 100 may receive an input of a clothing item 11 for which a user wants to receive a recommendation for a clothing combination (1). Specifically, the electronic device 100 may acquire an image including the clothing item 11 through the equipped camera 150. In this case, the electronic device 100 may acquire an image including the clothing item 11 through the camera 150 or may receive an image including the clothing item 11 from an external device through a communication unit.
  • the electronic device 100 may receive a selection of a feed provider by the user 10.
  • the electronic device 100 may display a feed provider that the user 10 can select through the provided display 140.
  • the display 150 is shown to be provided only on a part of the front surface of the electronic device 100, but may be applied to the entire front surface and may be provided on a side surface other than the front surface.
  • the camera 150 may be disposed to overlap an area of the display 140.
  • the electronic device 100 may receive a selection of a feed provider by a user's touch.
  • the electronic device 100 may receive a selection of a feed provider based on a user selection input through a button provided in the electronic device 100.
  • the electronic device 100 may receive a selection of a feed provider through a voice input through a microphone provided in the electronic device 100.
  • the electronic device 100 when the electronic device 100 and the terminal device of the user 10 communicate, the electronic device 100 The usage history information may be received, and the user 10 may select a preferred feed provider based on the usage history information. Specifically, the electronic device 100 is based on the information on the feed provider frequently accessed by the user and information on the feed provider to which the user 10 has subscribed in the usage history information of the terminal device of the user 10. You can also choose a feed provider.
  • a plurality of feed providers may be classified into a plurality of groups.
  • the group may be a higher category including a plurality of feed providers providing feeds having the same feature information.
  • the feed provider is a server corresponding to a clothing brand
  • a plurality of brands e.g., A brand, B brand
  • the electronic device 100 may provide a UI screen so that the user 10 selects a group.
  • the electronic device 100 may acquire a clothing combination including the input clothing item 11 by using the learned artificial intelligence model (2). Specifically, the electronic device 100 may obtain information on the clothing item 11 by analyzing an image including the clothing item 11.
  • the electronic device 100 may obtain information on the clothing item 11 by inputting an image including the clothing item 11 into an artificial intelligence model for extracting the clothing item from the combined image.
  • the artificial intelligence model for extracting the clothing item from the combined image may be the learned artificial intelligence model of the present disclosure or a separate artificial intelligence model.
  • the artificial intelligence model for extracting a clothing item from the combined image is to crop the image of the clothing item 11 from the combined image, and the image of the cropped clothing item 100 and the image of the clothing for sale previously received from the server of the clothing manufacturer.
  • the information on the clothing item 11 may be obtained by comparing.
  • the electronic device 100 may obtain a plurality of clothing combinations including the clothing item 11 by inputting the obtained information on the clothing item 11 into the artificial intelligence model learned to output the clothing combination.
  • the electronic device 100 may provide at least one clothing combination from among a plurality of acquired clothing combinations to the user 10.
  • the electronic device 100 may provide information on a clothing item included in at least one clothing combination to the user 10 (3).
  • the electronic device 100 receives and stores information on clothing items sold from clothing manufacturers included in the plurality of feed providers before performing a clothing recommendation operation, and among the information on the stored clothing items, the obtained clothing combination Information on the clothing item included in may be provided to the user 10.
  • the at least one clothing combination including clothing items for which information is provided to the user 10 may be a clothing combination corresponding to a feed provider selected by the user 10.
  • the electronic device 100 is a feed provider included in the selected group.
  • a plurality of clothing combinations corresponding to each of may be provided to the user 10.
  • FIG. 2 is a diagram for explaining a simple configuration of an electronic device according to an embodiment of the present disclosure.
  • the electronic device 100 may include a memory 110 and a processor 120.
  • the memory 110 may store various programs and data necessary for the operation of the electronic device 100. Specifically, at least one command may be stored in the memory 110.
  • the processor 120 may perform an operation of the electronic device 100 by executing a command stored in the memory 110.
  • the memory 110 may be implemented as a nonvolatile memory, a volatile memory, a flash memory, a hard disk drive (HDD), a solid state drive (SSD), or the like.
  • the memory 110 is accessed by the processor 120, and data read/write/edit/delete/update by the processor 120 may be performed.
  • the term memory refers to a memory 110, a ROM (not shown) in the processor 120, a RAM (not shown), or a memory card (not shown) mounted in the electronic device 100 (eg, micro SD Card, memory stick).
  • the memory 110 may store programs and data for configuring various screens to be displayed in the display area of the display.
  • the learned artificial intelligence model for operating the clothing recommendation system may be stored in the memory 110.
  • the learned artificial intelligence model may be learned based on a plurality of combination images in which a plurality of clothing items obtained from a plurality of feed providers are combined.
  • the learned artificial intelligence model may be learned by the electronic device 100, learned by an external server, or downloaded from an external server.
  • the learned artificial intelligence model may include a common parameter irrelevant to a plurality of feed providers and individual parameters corresponding to each of the plurality of feed providers.
  • the common parameter according to the present disclosure may mean overall characteristics of clothing or a specific style according to a multi-task configuration.
  • the overall characteristics of clothing may be characteristics related to category information such as material or T-shirt pants, and the overall characteristics of a specific style may refer to color characteristics unique to casual clothes or characteristics such as vintage.
  • the individual parameter according to the present disclosure may mean a parameter extracted according to the characteristics of each brand because, for example, characteristics of each brand are different even in the same casual category.
  • casual brands have a common parameter called casual, but each brand has its own characteristics, and each brand's unique characteristics (e.g., pattern, trademark, fit, etc.) can be expressed as individual parameters. have.
  • the learned artificial intelligence model may receive a plurality of combined images obtained from a plurality of feed providers as a training image, and may learn a common parameter and a plurality of individual parameters.
  • the learned artificial intelligence model may group a plurality of feed providers into a plurality.
  • the grouped feed providers may share common parameters. That is, common parameters can be shared for each group.
  • the learned artificial intelligence model may learn a common parameter corresponding to a plurality of grouped feed providers by using a plurality of combined images obtained from a plurality of grouped feed providers as a training image.
  • the learned artificial intelligence model may be configured hierarchically.
  • the hierarchical configuration may mean that the artificial intelligence model groups and learns feed providers that provide a combined image having similar feature information among a plurality of feed providers. That is, in the artificial intelligence model, a plurality of feed providers providing a combined image having a common feature may be grouped into a higher category related to the common feature and learned.
  • the common feature refers to the feature of whether features such as categories and styles are common among the combined images.For example, in the case of casual brands, the casual style may be a common feature, and in the case of formal brands, the modern style is It can be a common feature.
  • the upper category is a grouping of common features into a higher category. For example, casual, vintage, and layered may be grouped into one higher category, and modern, minimalism, and dandy may be grouped into another higher category. .
  • the learned artificial intelligence model may obtain a weight by extracting feature information for each of a plurality of combined images obtained from a plurality of feed providers.
  • the feature information may be at least one of a morphological feature, a color combination, and a texture combination of the clothing item included in the combination image.
  • the feature information may be a combination of morphological features, color features, and texture features of tops and bottoms.
  • the morphological characteristics of the clothing item may include the length of the top, the length and shape of the sleeve, the shape of the collar, etc. in the case of a top, and in the case of a bottom, the length of the bottom, the line of the pants, and the rise It may include length and the like.
  • the type, heel height, and shape of the front nose may be included in the morphological characteristics.
  • the texture feature may include a material of a clothing item, a type of cloth, a pattern included in the cloth, and the like.
  • the learned artificial intelligence model compares the weights of the artificial intelligence models acquired for each combination image through the above-described feature information, and group a plurality of feed providers that provide the combined image in which the difference in weight as a result of the comparison is within a preset range. have. That is, the learned artificial intelligence model can group feed providers that provide a similar style of combined image through the combination image analysis.
  • the learned artificial intelligence model may learn a common parameter such that, when a plurality of combined images each obtained from a plurality of grouped feed providers are input, a combination of clothing for each of the grouped feed providers is obtained.
  • the learned common parameter corresponds to a plurality of grouped feed providers, and may be a parameter shared by a plurality of grouped feed providers.
  • the learned artificial intelligence model may use the combined image itself provided from a plurality of feed providers as input data, but in another embodiment, a plurality of clothing items included in the combined image is extracted, and information on the extracted plurality of clothing items Can also be used as input data.
  • the plurality of clothing items included in the combined image may be extracted from the learned artificial intelligence model itself, or may be extracted from a separate artificial intelligence model.
  • the separate artificial intelligence model may be a convolutional neural network (CNN).
  • CNN convolutional neural network
  • information on a plurality of clothing items is extracted from a separate artificial intelligence model may be input into the learned artificial intelligence model, and the learned artificial intelligence model is separate.
  • a combination of information on a plurality of clothing items extracted from the artificial intelligence model of may be used as input data. An operation of extracting information on a plurality of clothing items from the combined image will be described in more detail below with reference to FIG. 5.
  • the processor 120 may be electrically connected to the memory 110 to control overall operation of the electronic device 100. Specifically, the processor 120 may control the electronic device 100 by executing at least one command stored in the memory 110.
  • the processor 120 of the present disclosure inputs the image input to the learned artificial intelligence model to provide a plurality of feed providers corresponding to each of the plurality of feed providers including the clothing item included in the input image.
  • a plurality of combinations of clothes can be obtained.
  • the learned artificial intelligence model may be learned to periodically receive a plurality of feeds from each of the plurality of feed providers and output a clothing combination to each of the feed providers.
  • the processor 120 may output a plurality of combinations of clothes including the input T-shirt, such as an input T-shirt, bottoms, shoes, and accessories, wherein the plurality of clothes combinations are It may correspond to each of a plurality of feed providers. That is, combinations of t-shirts, bottoms, shoes, accessories, etc. input for each of the plurality of feed providers may be output.
  • the processor 120 may extract clothing items included in the input image and input information on the extracted clothing items to the learned artificial intelligence model as input data.
  • the learned artificial intelligence model may output a combination of information on clothing items for each of a plurality of feed providers based on the input information on the clothing item.
  • the processor 120 may provide at least one clothing combination from among a plurality of clothing combinations acquired from the learned artificial intelligence model. Specifically, the processor 120 may provide information on clothing items included in at least one clothing combination. Here, the information on the clothing item may be a manufacturer, article number, size, color, etc. of the clothing item, and the processor 120 may also provide a link to an online purchase site of the corresponding clothing item according to an embodiment.
  • the processor 120 may provide a clothing combination corresponding to at least one feed provider selected from among a plurality of clothing combinations acquired from the learned artificial intelligence model.
  • the selected at least one feed provider may be selected by the user or may be selected based on the usage history information of the user's terminal device. That is, only selected clothing combinations among clothing combinations output from the learned artificial intelligence model may be provided to the user.
  • FIG. 3 is a block diagram illustrating a detailed configuration of the electronic device disclosed in FIG. 2.
  • the electronic device 100 includes a memory 110, a processor 120, a communication unit 130, a display 140, a camera 150, an input unit 160, a speaker 170, and a fashion recommendation application. It may include 180.
  • the communication unit 130 is a component for performing communication with an external electronic device. Meanwhile, the communication connection between the communication unit 130 and an external device may include communicating through a third device (eg, a repeater, a hub, an access point, a server, or a gateway).
  • a third device eg, a repeater, a hub, an access point, a server, or a gateway.
  • the communication unit 130 may communicate with the user terminal device through wired communication or wireless communication to receive usage history information of the user terminal device from the user terminal device.
  • the electronic device 100 is a user terminal through a wired connection, such as a universal serial bus (USB) or a high definition multimedia interface (HDMI), or a wireless connection such as Wi-fi, Bluetooth, and NFC tag with the user terminal device.
  • Device usage history information can be received.
  • the usage history information may include information on a feed provider frequently accessed by a user, a number of accesses, access time, information on a feed provider subscribed to by the user, and the like.
  • the processor 120 may select at least one feed provider from among a plurality of feed providers based on the usage history information of the user terminal device received through the communication unit 130, and at least one feed provider corresponding to the selected at least one feed provider. Clothing combinations can be provided.
  • the communication unit 130 may receive an artificial intelligence model learned from an external server by communicating with an external server.
  • the processor 120 may download a learned artificial intelligence model corresponding to at least one feed provider selected from among a plurality of artificial intelligence models learned in an external server from an external server through the communication unit 130.
  • the electronic device 100 is a tablet PC or a mobile device
  • the fashion recommendation application 180 including a plurality of artificial intelligence models learned from an external server is registered in the application store, the user may )
  • the fashion recommendation application 180 may be downloaded, and a learned artificial intelligence model corresponding to a preferred feed provider may be selected and downloaded through the fashion recommendation application 180.
  • the electronic device 100 since the electronic device 100 does not need to perform a learning operation, it is possible to reduce the data processing burden of the electronic device 100.
  • the display 140 may display various types of information according to the control of the processor 120.
  • the display 140 may display a UI screen for selecting a feed provider.
  • the display 140 may display information on clothing items included in the obtained clothing combination.
  • the display 140 may be implemented as a touch screen together with a touch panel.
  • the camera 150 is disposed on one side of the electronic device 100 and is a component for acquiring an image including a clothing item. In this case, the camera 150 may capture a still image or a video.
  • the input unit 160 may receive a user input for controlling the electronic device 100.
  • the input unit 160 may include a touch panel for receiving a user touch for controlling the electronic device 100, a button for receiving a user manipulation, a microphone for receiving a user's voice, and the like.
  • the example of the input unit 160 shown in FIG. 3 is only an exemplary embodiment, and may be implemented with other input devices (eg, a keyboard, a mouse, a motion input unit, etc.).
  • the speaker 170 is a component that outputs not only various audio data on which various processing tasks such as decoding, amplification, and noise filtering have been performed, but also various notification sounds or voice messages by the processor 120 or a separate audio processing unit.
  • the speaker 170 may output information on the acquired clothing item as a voice message in a natural language format.
  • a configuration for outputting audio may be implemented as a speaker, but this is only an embodiment and may be implemented as an output terminal capable of outputting audio data.
  • the fashion recommendation application 180 may download a plurality of artificial intelligence models learned from an external server. That is, when the electronic device 100 is a tablet PC or a mobile device, when the fashion recommendation application 180 including a plurality of artificial intelligence models learned from an external server is registered in the application store of the electronic device 100, the user The fashion recommendation application 180 may be downloaded from the device 100, and a learned artificial intelligence model corresponding to a preferred feed provider may be selected and downloaded through the fashion recommendation application 180.
  • the electronic device 100 when the electronic device 100 is a clothes manager, a clothes manager (not shown) may further be included, and the clothes manager (not shown) may be used without a separate washing process through washing water. It may be a configuration for removing wrinkles, dust or odors.
  • the clothing management unit (not shown) may include a steam generation unit generating steam, a clothing support unit capable of supporting or fixing clothes inside the electronic device 100, an air circulation unit, and the like.
  • FIG. 4 is a diagram illustrating a clothing recommendation system according to an exemplary embodiment of the present disclosure.
  • the clothing recommendation system 100 may include a plurality of feed providers 200-1 to 200-n and an artificial intelligence model 111.
  • each of the plurality of feed providers 200-1 to 200-n may periodically provide a feed to the artificial intelligence model 111 (1).
  • the feed may be a combination image in which a plurality of clothing items are combined.
  • the artificial intelligence model 111 may learn a clothing combination by using a plurality of combination images obtained from the plurality of feed providers 200-1 to 200-n as a training image (2). Specifically, the artificial intelligence model 111 extracts a plurality of clothing items from the obtained combination image, and uses a combination of information on the extracted clothing item as input data, as a clothing combination for a feed provider who provides the combination image. You can learn. An operation of extracting a plurality of clothing items from the combined image will be described in detail below with reference to FIG. 5.
  • the artificial intelligence model 111 includes a common parameter irrelevant to a feed provider and a plurality of individual parameters corresponding to each of the feed providers, and the combined images obtained from the plurality of feed providers 200-1 to 200-n are used as training data. It may be learned by performing multi-task learning for learning a common parameter and a plurality of individual parameters as input.
  • the artificial intelligence model 111 may be learned in a server or in an electronic device.
  • the user 10 may input a clothing item image into the learned artificial intelligence model 111 (3).
  • the learned artificial intelligence model 111 is stored in the electronic device of the present disclosure, and when the artificial intelligence model 111 is learned in the server, it may be downloaded from the server to the electronic device. That is, the user 10 may input a clothing item image into the artificial intelligence model 111 learned through the electronic device.
  • the learned artificial intelligence model 111 may provide a clothing combination including the input clothing item to the user 10 (4).
  • the learned artificial intelligence model 111 may output a plurality of clothing combinations corresponding to each of the plurality of feed providers 200-1 to 200-n that provided the combination image to the artificial intelligence model 111.
  • only a clothing combination for at least one feed provider selected by the user among the plurality of output clothing combinations may be provided to the user 10.
  • the learned artificial intelligence model 111 may be converted to a plurality of feed providers 200-1 to 200-n) Output clothing combinations including'pants' input for each unit, and only clothing combinations corresponding to at least one feed provider selected by the user 10 among the output n clothing combinations are provided to the user 10. I can.
  • FIG. 5 is a diagram illustrating a method of obtaining input data of an artificial intelligence model according to an embodiment of the present disclosure.
  • the feed provider 1 200-1 may provide a combined image 51 including a plurality of clothing items to the artificial intelligence model 111.
  • the artificial intelligence model 111 may extract a plurality of clothing items 52 to 55 from the provided combination image 51.
  • the artificial intelligence model 111 may extract clothing items such as a top 52, accessories 53, bottoms 54, and shoes 55 included in the combined image 111 through image analysis.
  • FIG. 5 it is illustrated that a plurality of clothing items 52 to 55 are extracted from the combination image 51 input to the artificial intelligence model 111, but in actual implementation, a plurality of clothing items from the combination image 51
  • the operation of extracting (52 to 55) may be performed through a separate artificial intelligence model.
  • the artificial intelligence model 111 acquires information on a plurality of extracted clothing items 52 to 55, and learns a combination for the feed provider 1 200-1 by using the obtained combination of information as input data ( 56) You can.
  • the information on the extracted plurality of clothing items refers to information on clothing items that can be recognized from an image including the clothing item, and may be information such as color, shape, and pattern.
  • information on the plurality of clothing items 52 to 55 may be obtained based on information on sale clothing provided from a plurality of clothing manufacturers. For example, by comparing images of a plurality of clothing items 52 to 55 extracted from images of clothing for sale provided from a clothing manufacturer, information on a plurality of clothing items 52 to 55 among information about clothing for sale is Can be obtained.
  • feed provider 1 (200-1) is shown to provide one combined image 51 to the artificial intelligence model 111, but in actual implementation, feed provider 1 (200-1) Is provided a plurality of combined images to the artificial intelligence model 111 periodically, and a plurality of feed providers other than the feed provider 1 200-1 may also provide the combined images to the artificial intelligence model 111. In this case, a plurality of feed providers other than the feed provider 1 200-1 may operate in the same manner as the feed provider 1 200-1.
  • FIG. 6 to 8 are diagrams for explaining various learning methods of an artificial intelligence model. Specifically, FIG. 6 is a diagram for explaining a method of learning an artificial intelligence model using a hard parameter sharing method of multi-task learning.
  • a plurality of feed providers 200-1 to 200-n may provide a training image to the artificial intelligence model 111.
  • the training image may be a combination image provided from a plurality of feed providers 200-1 to 200-n.
  • the hard parameter sharing method may mean a method in which a hidden layer (hidden layer) is shared between all tasks and an output layer for each task is maintained.
  • the task may mean a process of obtaining individual parameters corresponding to each feed provider.
  • the artificial intelligence model 111 trained in the hard parameter sharing method includes a common parameter 61 irrelevant to a feed provider and a plurality of individual parameters 62-1 corresponding to each of the plurality of feed providers 200-1 to 200-n.
  • the common parameter 61 may be a hidden layer (hidden layer) and an output layer (output layer) for each feed provider of the plurality of individual parameters 62-1 to 62-n.
  • the common parameter 61 is illustrated as being composed of three layers, but the present invention is not limited thereto.
  • a layer consisting of a plurality of individual parameters 62-1 to 62-n is arranged after a layer consisting of the common parameter 61, but a plurality of individual layers between the layers of the common parameter 61 At least one layer composed of the parameters 62-1 to 62-n may be arranged.
  • This hard parameter sharing method has the effect of greatly reducing the risk of overfitting.
  • FIG. 7 is a diagram for explaining a method of learning an artificial intelligence model using a soft parameter sharing method of multi-task learning.
  • a plurality of feed providers 200-1 to 200-n may provide a training image to the artificial intelligence model 111.
  • the training image may be a combination image provided from a plurality of feed providers 200-1 to 200-n.
  • the soft parameter sharing method may refer to a method in which each task has its own parameter, and the distance between its own parameters is normalized so that its own parameters become similar.
  • the artificial intelligence model 111 learned in the soft parameter sharing method may include common parameters 71-1 to 71-n and individual parameters 72-1 to 72-n.
  • the common parameters 71-1 to 71-n in FIG. 7 are their own parameters included in the models for each of the plurality of feed providers 200-1 to 200-n, but they can be viewed as being shared in that they are normalized to each other. Meanwhile, the plurality of individual parameters 72-1 to 72-n may be an output layer for each feed provider.
  • FIG. 8 is a diagram for explaining a method of hierarchically learning an artificial intelligence model.
  • a plurality of feed providers 200-1 to 200 -n may be grouped into a plurality of groups.
  • feed provider 1 (200-1) and feed provider 2 (200-2) are the first group
  • feed provider 3 (200-3) and feed provider 4 (200-4) are the second group
  • feed providers (n-1)(200-(n-1)) and feed provider n may be grouped into an m-th group.
  • the artificial intelligence models 111-1 to 111-m may be trained for each group, and each of the artificial intelligence models 111-1 to 111-m may include separate common parameters and learn them.
  • each group may be composed of a feed provider that analyzes a combined image provided by each feed provider and provides a combined image having similar characteristic information. That is, the group may be a higher category of the included feed provider.
  • the feature information may be at least one of a morphological feature, a color combination, and a texture combination of a clothing item included in the combination image.
  • feed provider 1 200-1 that is'hippie' and feed provider 2 (200-2) that is'hip-hop' are classified into a first group that is'casual'
  • the artificial intelligence model 111-1 corresponding to one group may be trained by receiving a combination image provided by the feed provider 1 200-1 and the feed provider 2 200-2 as a training image.
  • feed provider 3 (200-3) that is a'suit' and feed provider 4 (200-4) that is a'uniform' are classified into a second group, which is a'suit', and an artificial intelligence model 111- corresponding to the second group is classified. 2)
  • the combined image provided by the feed provider 3 (200-3) and the feed provider 4 (200-4) may be input and learned as a training image.
  • the feed provider 1 (200-1) and the feed provider 2 (200-2) respectively use only the artificial intelligence model 111-1 corresponding to the first group. Only the corresponding clothing combination can be obtained. Accordingly, the burden of processing data on the electronic device may be reduced compared to an embodiment in which n combinations of clothes are output.
  • FIG. 8 although it is shown that two feed providers are included in one group, it is not limited thereto, and three or more feed providers may be included in one group, and a different number of feed providers may be included for each group.
  • the artificial intelligence models 111-1 to 111-m are trained in a hard parameter sharing method, but at least some artificial intelligence models may be trained in a soft parameter sharing method. .
  • FIG. 9 is a block diagram illustrating an output of a learned artificial intelligence model according to an embodiment of the present disclosure.
  • the artificial intelligence model 111 may output a plurality of clothing combinations 92-1 to 92-n.
  • the plurality of clothing combinations 92-1 to 92-n may correspond to a plurality of feed providers who have provided the combination image when learning the artificial intelligence model 111.
  • a plurality of clothes combinations 92-1 to 92-n which are a plurality of output data, are obtained for input data, and a plurality of clothes Information on clothing items included in at least one of the combinations 92-1 to 92-n may be provided to the user.
  • the clothing combination provided to the user may correspond to a feed provider selected by the user or may correspond to a user preferred feed provider selected based on the usage history of the user terminal device.
  • FIG. 10 is a diagram illustrating a clothing recommendation user interface (UI) screen of an electronic device according to various embodiments of the present disclosure.
  • the electronic device may display a UI screen 1010 for selecting a feed provider through an equipped display.
  • an object for selecting a plurality of feed providers may be displayed on the UI screen 1010 for selecting a feed provider.
  • the plurality of feed providers displayed on the UI screen 1010 provided training images in the learning stage of the artificial intelligence model.
  • the artificial intelligence model may be learned in the same manner as the artificial intelligence model illustrated in FIGS. 6 and 7.
  • the feed provider may be a server of a clothing manufacturer, an SNS server of a celebrity, a server of a fashion magazine, or the like.
  • the user may select one feed provider on the UI screen 1010 for selecting a feed provider.
  • the user may select a feed provider through a touch screen, a button, a voice input, or the like.
  • the electronic device acquires the history information of the user terminal device, even if the user's selection is not input, the user's preferred feed provider may be selected.
  • the electronic device may display a UI screen 1020 that displays information on a clothing item included in a clothing combination corresponding to the selected feed provider among a plurality of clothing combinations acquired from the learned artificial intelligence model. .
  • the electronic device when'feed provider 2'is selected as shown in FIG. 10, the electronic device includes information on clothing items included in the clothing combination for'feed provider 2'among a plurality of clothing combinations obtained from the artificial intelligence model. Can provide.
  • 'Image, information on'accessories', shoes image, and information on shoes may be included.
  • the information on the clothing item may include at least one of manufacturer information, article number information, color information, size information, price information, and a link to an online purchase site for purchasing the clothing item.
  • the information on the clothing item may include at least one of manufacturer information, article number information, color information, size information, price information, and a link to an online purchase site for purchasing the clothing item.
  • only an image of a clothing item is displayed without information on the clothing item, and when an image of the clothing item is selected, information on the selected clothing item may be provided.
  • FIG. 11 is a diagram illustrating a clothing recommendation user interface (UI) screen of an electronic device according to various embodiments of the present disclosure.
  • the learned artificial intelligence model included in the electronic device of FIG. 11 may be learned in the same manner as the artificial intelligence model illustrated in FIG. 8.
  • the electronic device may display a UI screen 1110 for selecting a style through an equipped display.
  • the style may be a higher concept in which a plurality of feed providers are grouped.
  • a plurality of feed providers providing a combined image having similar feature information through analysis of a combined image obtained from a feed provider may be grouped in a specific style.
  • the style may include suit, casual, sporty, and the like, and may include a clothing manufacturer server and a celebrity SNS server having similar characteristic information for each style.
  • the style is only an example of a criterion in which a plurality of feed providers are grouped, and may be grouped according to various criteria such as color and texture (eg, tweed, herringbone, linen, knit, etc.).
  • an object for selecting a plurality of styles may be displayed on the UI screen 1110 for selecting a style.
  • the plurality of styles displayed on the UI screen 1110 include a plurality of feed providers, respectively, and the plurality of feed providers for each style generate a training image to learn common parameters for each style in the learning stage of the artificial intelligence model. May be provided.
  • the user may select one style on the UI screen 1110 for selecting a style.
  • the user can select a style through a touch screen, a button, or a voice input.
  • the electronic device acquires the history information of the user terminal device, even if the user's selection is not input, the user's preferred style may be identified.
  • the electronic device displays information on a clothing item included in a clothing combination corresponding to each of a plurality of feed providers included in the selected style among a plurality of clothing combinations acquired from the learned artificial intelligence model. ) Can be displayed.
  • the electronic device when'casual' is selected as shown in FIG. 11, the electronic device includes a plurality of clothing providers (A feed provider, B feed provider) included in'casual' among a plurality of clothing combinations obtained from the artificial intelligence model. , C feed provider) It is possible to provide information on clothing items included in the clothing combination for each.
  • a feed provider B feed provider
  • C feed provider It is possible to provide information on clothing items included in the clothing combination for each.
  • information on a clothing item included in a clothing combination for each of a plurality of clothing providers included in the'casual' is displayed on the UI screen 1120 displaying information on the clothing item, and at this time, each of the plurality of clothing providers
  • the information on the clothing item included in the Korean clothing combination may include a'top' image, which is a clothing item input by the user, a matching'bottom' image, a'accessory' image, and a'shoes' image.
  • the artificial intelligence model is hierarchically trained, a plurality of combinations of clothes having similar characteristics can be provided, so that a user's selection range can be broadened.
  • FIG. 12 is a flowchart illustrating a method of recommending clothes by an electronic device according to an embodiment of the present disclosure.
  • the electronic device may obtain a plurality of combined images combined with a plurality of clothing items from a plurality of feed providers (S1210).
  • the plurality of feed providers may be a server of a clothing manufacturer, an SNS server of a celebrity, a server of a fashion magazine, and the like, and may periodically provide a combination image to the electronic device.
  • the combined image is an image showing a combination of a plurality of clothing items, and may be referred to as a feed.
  • the electronic device is a combination of a plurality of clothing corresponding to each of a plurality of feed providers including clothing items included in the input image by inputting the image input to the artificial intelligence model learned based on the plurality of obtained combination images.
  • the learned artificial intelligence model may include a common parameter irrelevant to a feed provider and a plurality of individual parameters corresponding to each of the feed providers.
  • the learned artificial intelligence model may be learned by performing multi-task learning in which a common parameter and a plurality of individual parameters are learned using a combination image obtained from a plurality of feed providers as a training image. have.
  • the learned artificial intelligence model may extract information on a clothing item from an input image, and obtain a plurality of clothing combinations including clothing items by using the extracted clothing item information as input data.
  • the plurality of clothing combinations may be clothing combinations for each of the plurality of feed providers that provide training images in the step of learning the artificial intelligence model.
  • the electronic device may provide information on a clothing item included in at least one clothing combination among a plurality of clothing combinations (S1230). Specifically, an operation command for selecting at least one feed provider from among a plurality of feed providers may be input from the user, or at least one feed provider preferred by the user may be selected based on the usage history information of the user terminal device.
  • the electronic device stores information on clothing items included in a clothing combination corresponding to at least one feed provider selected by a user from among a plurality of clothing combinations acquired from the learned artificial intelligence model, or at least one feed provider selected based on usage history information.
  • the information on the clothing item may include at least one of an image of the clothing item, manufacturer information, article number information, color information, size information, price information, and a link to an online purchase site through which the corresponding clothing item can be purchased.
  • FIG. 13 is a sequence diagram illustrating an embodiment of recommending clothing according to a user's selection of a feed provider.
  • an artificial intelligence model may be trained by an electronic device.
  • a plurality of feed providers 200-1 to 200-n may provide a combined image to the electronic device 100 (S1301 ).
  • the combined image may be an image in which a plurality of clothing items are combined.
  • the plurality of feed providers 200-1 to 200 -n may periodically provide a combined image to the electronic device 100.
  • the electronic device 100 may include a plurality of feed providers 200-1 to 200 -n. If a combination image is provided from ), a clothing item included in each combination image may be extracted (S1302). In this case, the electronic device 100 may extract a plurality of clothing items included in each combination image through analysis of the combination image. For example, the electronic device 100 may extract a plurality of clothing items included in the combined image using the CNN model.
  • the electronic device 100 may learn a combination of a plurality of clothing items based on the extracted information on a plurality of clothing items (S1303). Specifically, the electronic device 100 may learn a combination of clothes corresponding to a feed provider who provided the combined image by using a combination of the extracted information of a plurality of clothing items as input data of the artificial intelligence model.
  • the electronic device 100 may learn the artificial intelligence model by repeatedly performing steps S1302 and S1303.
  • the electronic device 100 may select a feed provider and receive a clothing item image (S1304). Specifically, the electronic device 100 may receive a feed provider selection through a display and an input unit provided in the electronic device 100 and may receive a clothing item image through a camera provided in the electronic device 100. Meanwhile, as another example, the electronic device 100 may obtain usage history information of the user terminal device and select a user preferred feed provider. Alternatively, the electronic device 100 may receive an image including a clothing item through a communication unit.
  • FIG. 13 for convenience of explanation, it is illustrated and described that the selection of the feed provider and the reception of the clothing item image are performed in the same step. Provider selection may be received.
  • the electronic device 100 may acquire a plurality of clothing combinations (S1305). Specifically, when a clothing item image is received, the electronic device 100 extracts information on the clothing item from the image, inputs the information on the extracted clothing item into the learned artificial intelligence model, and receives a plurality of feed providers 200- 1 to 200-n) a plurality of clothing combinations corresponding to each may be obtained.
  • the clothing combination may mean a combination of clothing items that match the clothing items included in the input clothing item image.
  • the electronic device 100 may provide information on a clothing item included in at least one clothing combination corresponding to a feed provider selected from among a plurality of clothing combinations acquired from the learned artificial intelligence model (S1306).
  • the electronic device 100 may include images of a plurality of clothing items included in a clothing combination, manufacturer information, article number information, color information, size information, price information, and links to an online purchase site where the clothing item can be purchased. At least one may be included.
  • the electronic device 100 may present a higher category in which a plurality of feed providers are grouped in step S1304, and if the user selects one of the plurality of upper categories , It is also possible to provide information on clothing items included in a plurality of clothing combinations respectively corresponding to a plurality of feed providers included in the selected category.
  • FIG. 14 is a sequence diagram illustrating an embodiment of recommending clothing according to usage history information of a user terminal device.
  • a plurality of feed providers 200-1 to 200-n provide a combined image to the electronic device 100 (S1401), and the electronic device 100 provides a plurality of feed providers 200-1 to 200-n.
  • a clothing item may be extracted from the combination image obtained from 200-n) (S1402), and a clothing combination may be learned using information on the extracted clothing item (S1403).
  • the information on the extracted clothing item may mean pattern information on the clothing item extracted from the combination image, and may include, for example, color information, shape information, pattern information, and the like. This operation is the same as S1301 to S1303 of FIG. 13 described above, and a duplicate description will be omitted.
  • the electronic device 100 is transferred from the user terminal device 300 to the user terminal device 300.
  • the usage history information of may be transmitted (S1404).
  • the user terminal device 300 may communicate with the electronic device 100 through wireless communication such as Wi-fi, Bluetooth, and NFC tagging, and wired communication such as a USB connection.
  • the usage history information of the user terminal device 300 may include information on a feed provider subscribed by the user, a history of accessing the feed provider, the number of accesses, and access time.
  • the electronic device 100 may receive a clothing item image (S1405).
  • the clothing item image may be input through a camera provided in the electronic device 100 or may be received from an external device through a communication unit.
  • the electronic device 100 may acquire a plurality of clothing combinations (S1406). This operation is the same as step S1305 of FIG. 13 described above, and redundant descriptions will be omitted.
  • the electronic device 100 may provide information on clothing items included in at least one clothing combination corresponding to a feed provider selected based on usage history information among the plurality of clothing combinations. (S1407). In more detail, the electronic device 100 may provide information on clothing items included in at least one clothing combination corresponding to at least one feed provider preferred by the user based on the acquired use history information.
  • 15 is a sequence diagram illustrating an embodiment in which an electronic device downloads a learned artificial intelligence model from an external server.
  • a plurality of feed providers 200-1 to 200-n provide a combined image to the server 400 (S1501), and the server 400 includes a plurality of feed providers 200-1 to 200-
  • a clothing item may be extracted from the combination image obtained from n) (S1502), and a clothing combination may be learned using information on the extracted clothing item (S1503). Since the operation of the server 400 is the same as the operation of the electronic device 100 of S1301 to S1303 of FIG. 13 and S1401 to S1403 of FIG. 14 described above, a duplicate description will be omitted.
  • the electronic device 100 may receive a selection of a feed provider from the user (S1504). Specifically, the selection of a feed provider may be received through a display and an input unit provided in the electronic device 100. As another example, the electronic device 100 may obtain usage history information of the user terminal device and select a user preferred feed provider.
  • the electronic device 100 may request an artificial intelligence model corresponding to the selected feed provider to the server 400 (S1505). Further, the electronic device 100 may receive the learned artificial intelligence model from the server 400 (S1506).
  • the server 400 may upload the learned artificial intelligence model to the application store, and the electronic device 100 may download the learned artificial intelligence model, such as downloading an application.
  • the electronic device 100 is the upper category of the grouped plurality of feed providers. May be selected by a user, and an artificial intelligence model corresponding to the selected category may be downloaded from the server 400.
  • the electronic device 100 may obtain a clothing combination by inputting information on the clothing item to the learned artificial intelligence model (S1508). .
  • the learned artificial intelligence model may be selected according to a feed provider selection after receiving the clothing item image.
  • the electronic device 100 may provide information on a clothing item included in the clothing combination (S1509). Accordingly, there is no need for the electronic device 100 to perform a learning operation, and thus, a data processing burden on the electronic device 100 can be reduced.
  • 16 is a diagram illustrating a configuration of a server according to an embodiment of the present disclosure.
  • the electronic device 100 described above may be implemented as a server 400, and the server 400 according to the present disclosure includes a memory 410, a processor 420, and a recommendation engine 430. ) And a communication unit 440.
  • the memory 410 may store various programs and data required for the operation of the server 400. Specifically, at least one command may be stored in the memory 410.
  • the processor 420 may perform an operation of the server 400 by executing a command stored in the memory 410.
  • the learned artificial intelligence model for operating the clothing recommendation system may be stored in the memory 410. Specifically, the learned artificial intelligence model may be learned based on a plurality of combined images in which a plurality of clothing items obtained from a plurality of feed providers are combined. In this case, the learned artificial intelligence model may be learned by the server 400.
  • the learned artificial intelligence model may receive a plurality of combined images obtained from a plurality of feed providers as a training image, and learn a common parameter and a plurality of individual parameters.
  • the learned artificial intelligence model may group a plurality of feed providers into a plurality.
  • the grouped feed providers may share common parameters. That is, common parameters can be shared for each group.
  • the learned artificial intelligence model may learn a common parameter corresponding to a plurality of grouped feed providers by using a plurality of combined images obtained from a plurality of grouped feed providers as a training image.
  • the learned artificial intelligence model may be configured hierarchically.
  • the hierarchical configuration may mean that the artificial intelligence model groups and learns feed providers that provide a combined image having similar feature information among a plurality of feed providers. That is, in the artificial intelligence model, a plurality of feed providers providing a combined image having a common feature may be grouped into a higher category related to the common feature and learned.
  • the learned artificial intelligence model may obtain a weight by extracting feature information for each of a plurality of combined images obtained from a plurality of feed providers.
  • the feature information may be at least one of a morphological feature, a color combination, and a texture combination of the clothing item included in the combination image.
  • the learned artificial intelligence model may compare weights obtained for each combination image through the above-described feature information, and group a plurality of feed providers providing a combination image in which a difference in weight is within a preset range as a result of the comparison. That is, the learned artificial intelligence model can group feed providers that provide a similar style of combined image through the combination image analysis.
  • the learned artificial intelligence model may learn a common parameter such that, when a plurality of combined images each obtained from a plurality of grouped feed providers are input, a combination of clothing for each of the grouped feed providers is obtained.
  • the learned common parameter corresponds to a plurality of grouped feed providers, and may be a parameter shared by a plurality of grouped feed providers.
  • the learned artificial intelligence model may use the combined image itself provided from a plurality of feed providers as input data, but in another embodiment, a plurality of clothing items included in the combined image is extracted, and information on the extracted plurality of clothing items Can also be used as input data.
  • the processor 420 is electrically connected to the memory 410 to control overall operation of the server 400. Specifically, the processor 420 may control the server 400 by executing at least one command stored in the memory 410.
  • the processor 420 of the present disclosure inputs the image input to the learned artificial intelligence model to provide a plurality of feed providers corresponding to each of the plurality of feed providers including the clothing item included in the input image.
  • a plurality of combinations of clothes can be obtained.
  • the learned artificial intelligence model may be learned to periodically receive a plurality of feeds from each of the plurality of feed providers and output a clothing combination to each of the feed providers.
  • the processor 420 may output a plurality of combinations of clothes including the input T-shirt, such as an input T-shirt, bottoms, shoes, and accessories, wherein the plurality of clothes combinations are It may correspond to each of a plurality of feed providers. That is, combinations of t-shirts, bottoms, shoes, accessories, etc. input for each of the plurality of feed providers may be output.
  • the processor 420 may extract clothing items included in the input image and input information on the extracted clothing items to the learned artificial intelligence model as input data.
  • the learned artificial intelligence model may output a combination of information on clothing items for each of a plurality of feed providers based on the input information on the clothing item.
  • the processor 420 may provide at least one clothing combination from among a plurality of clothing combinations acquired from the learned artificial intelligence model. Specifically, the processor 420 may provide information on clothing items included in at least one clothing combination. Here, the information on the clothing item may be a manufacturer, article number, size, color, etc. of the clothing item, and the processor 420 may also provide a link to an online purchase site of the corresponding clothing item according to an embodiment.
  • the recommendation engine 430 may provide a clothing combination corresponding to a feed provider based on the learned artificial intelligence model. Specifically, when at least one feed provider from among the plurality of feed providers is selected, the recommendation engine 430 may provide a clothing combination corresponding to at least one feed provider selected from among a plurality of clothing combinations obtained from the learned artificial intelligence model. have. In this case, the selected at least one feed provider may be selected by the user or may be selected based on the usage history information of the user's terminal device. That is, only selected clothing combinations among clothing combinations output from the learned artificial intelligence model may be provided to the user.
  • the communication unit 440 is a component that communicates with various types of external devices according to various types of communication methods.
  • the communication unit 440 may be implemented as a Wi-Fi module. That is, the Wi-Fi module of the communication unit 440 may receive connection information (eg, SSID, encryption key information, etc.) received from the user terminal device and perform communication with the user terminal device based on the received connection information. have.
  • connection information eg, SSID, encryption key information, etc.
  • the communication unit 440 may communicate with the user terminal device through wired communication or wireless communication to receive usage history information of the user terminal device from the user terminal device.
  • the communication unit 440 may receive information on a feed provider selected from the user terminal device. In addition, the communication unit 440 may receive a clothing item image from the user terminal device, and transmit information on the clothing item included in the clothing combination generated by the server 400 to the user terminal device.
  • 17 is a sequence diagram illustrating an embodiment of a server that recommends clothing according to a user's selection of a feed provider.
  • a plurality of feed providers 200-1 to 200-n provide a combined image to the server 400 (S1701), and the server is provided from a plurality of feed providers 200-1 to 200-n.
  • a clothing item may be extracted from the obtained combination image (S1702), and a clothing combination may be learned using information on the extracted clothing item (S1703).
  • the user terminal device 300 may receive a selection of a feed provider from the user (S1704). Specifically, a feed provider selection may be received through a display and an input unit provided in the user terminal device 300. As another example, the user terminal device 300 may obtain the user's usage history information and select a user preferred feed provider.
  • the user terminal device 300 may transmit the selected feed provider information to the server 400 (S1705).
  • the user terminal device 300 may receive the clothing image (S1706) and transmit the received clothing image to the server 400 (S1707).
  • the server 400 may obtain a plurality of clothing combinations (S1708). Specifically, when a clothing item image is received, the server 400 extracts information on the clothing item from the received image, inputs the information on the extracted clothing item into the learned artificial intelligence model, and provides a plurality of feed providers 200. -1 ⁇ 200-n) it is possible to obtain a plurality of clothing combinations corresponding to each.
  • the server 400 may obtain information on a clothing item included in at least one clothing combination corresponding to a feed provider selected from among a plurality of clothing combinations acquired from the learned artificial intelligence model (S1709). Then, the acquired clothing item information may be transmitted to the user terminal device 300 (S1710).
  • embodiments described above may be implemented in a recording medium that can be read by a computer or a similar device using software, hardware, or a combination thereof.
  • embodiments described in the present disclosure include Application Specific Integrated Circuits (ASICs), digital signal processors (DSPs), digital signal processing devices (DSPDs), programmable logic devices (PLDs), field programmable gate arrays (FPGAs). ), processors, controllers, micro-controllers, microprocessors, and electric units for performing other functions.
  • ASICs Application Specific Integrated Circuits
  • DSPs digital signal processors
  • DSPDs digital signal processing devices
  • PLDs programmable logic devices
  • FPGAs field programmable gate arrays
  • processors controllers, micro-controllers, microprocessors, and electric units for performing other functions.
  • the embodiments described herein may be implemented by the processor itself.
  • a non-transitory readable medium may be mounted and used in various devices.
  • the non-transitory readable medium refers to a medium that stores data semi-permanently and can be read by a device, rather than a medium that stores data for a short moment, such as a register, cache, and memory.
  • programs for performing the above-described various methods may be provided by being stored in a non-transitory readable medium such as a CD, DVD, hard disk, Blu-ray disk, USB, memory card, ROM, or the like.
  • a method according to various embodiments disclosed in this document may be provided by being included in a computer program product.
  • Computer program products can be traded between sellers and buyers as commodities.
  • the computer program product may be distributed online in the form of a device-readable storage medium (eg, compact disc read only memory (CD-ROM)) or through an application store (eg, Play StoreTM).
  • an application store eg, Play StoreTM
  • at least some of the computer program products may be temporarily stored or temporarily generated in a storage medium such as a server of a manufacturer, a server of an application store, or a memory of a relay server.

Landscapes

  • Engineering & Computer Science (AREA)
  • Business, Economics & Management (AREA)
  • Theoretical Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • General Health & Medical Sciences (AREA)
  • Health & Medical Sciences (AREA)
  • Tourism & Hospitality (AREA)
  • General Physics & Mathematics (AREA)
  • Computational Linguistics (AREA)
  • Biomedical Technology (AREA)
  • Data Mining & Analysis (AREA)
  • Molecular Biology (AREA)
  • Computing Systems (AREA)
  • General Engineering & Computer Science (AREA)
  • Biophysics (AREA)
  • Mathematical Physics (AREA)
  • Software Systems (AREA)
  • Evolutionary Computation (AREA)
  • Artificial Intelligence (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Economics (AREA)
  • Human Resources & Organizations (AREA)
  • Marketing (AREA)
  • Primary Health Care (AREA)
  • Strategic Management (AREA)
  • General Business, Economics & Management (AREA)
  • Image Analysis (AREA)
  • Information Retrieval, Db Structures And Fs Structures Therefor (AREA)

Abstract

La présente invention concerne un dispositif électronique. Le dispositif électronique selon la présente invention comprend : une mémoire comprenant au moins une instruction et un processeur connecté à la mémoire pour commander le dispositif électronique. Lorsqu'une image comprenant des articles vestimentaires est entrée par exécution d'au moins une instruction, le processeur obtient une pluralité de combinaisons de vêtements correspondant à chaque fournisseur parmi une pluralité de fournisseurs de flux comprenant les articles vestimentaires inclus dans l'image entrée par l'intermédiaire d'une entrée de l'image entrée dans un modèle d'intelligence artificielle instruit sur la base d'une pluralité d'images combinées, dans lesquelles une pluralité d'articles vestimentaires obtenus à partir d'une pluralité de fournisseurs de flux sont combinés et fournit des informations sur des articles vestimentaires inclus dans au moins une des combinaisons de vêtements.
PCT/KR2020/010208 2019-08-27 2020-08-03 Dispositif électronique et son procédé de recommandation de vêtements WO2021040256A1 (fr)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
KR1020190105410A KR20210025399A (ko) 2019-08-27 2019-08-27 전자 장치 및 이의 의류 추천 방법
KR10-2019-0105410 2019-08-27

Publications (1)

Publication Number Publication Date
WO2021040256A1 true WO2021040256A1 (fr) 2021-03-04

Family

ID=74684517

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/KR2020/010208 WO2021040256A1 (fr) 2019-08-27 2020-08-03 Dispositif électronique et son procédé de recommandation de vêtements

Country Status (2)

Country Link
KR (1) KR20210025399A (fr)
WO (1) WO2021040256A1 (fr)

Families Citing this family (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
KR102318952B1 (ko) * 2021-03-31 2021-10-27 임정현 인공지능을 활용한 의류 추천 및 구매 방법, 장치 및 시스템

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
KR20090058389A (ko) * 2007-12-04 2009-06-09 동명대학교산학협력단 지능형 패션 코디네이터 시스템 및 운영 방법
KR20100004062A (ko) * 2008-07-02 2010-01-12 팔로 알토 리서치 센터 인코포레이티드 패션 관련 정보에 기초하여 소셜 네트워킹을 조성하는 방법
KR20120114806A (ko) * 2011-04-08 2012-10-17 주식회사 케이티 가상 옷장 제공 시스템 및 방법
KR101725960B1 (ko) * 2015-06-30 2017-04-11 주식회사 바디엘 의상 코디 시스템 및 방법
KR20190093813A (ko) * 2018-01-19 2019-08-12 네이버 주식회사 인공지능 기반 상품 추천 방법 및 그 시스템

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
KR20090058389A (ko) * 2007-12-04 2009-06-09 동명대학교산학협력단 지능형 패션 코디네이터 시스템 및 운영 방법
KR20100004062A (ko) * 2008-07-02 2010-01-12 팔로 알토 리서치 센터 인코포레이티드 패션 관련 정보에 기초하여 소셜 네트워킹을 조성하는 방법
KR20120114806A (ko) * 2011-04-08 2012-10-17 주식회사 케이티 가상 옷장 제공 시스템 및 방법
KR101725960B1 (ko) * 2015-06-30 2017-04-11 주식회사 바디엘 의상 코디 시스템 및 방법
KR20190093813A (ko) * 2018-01-19 2019-08-12 네이버 주식회사 인공지능 기반 상품 추천 방법 및 그 시스템

Also Published As

Publication number Publication date
KR20210025399A (ko) 2021-03-09

Similar Documents

Publication Publication Date Title
WO2018117704A1 (fr) Appareil électronique et son procédé de fonctionnement
WO2020153796A1 (fr) Dispositif électronique et procédé de fonctionnement associé
WO2020251238A1 (fr) Procédé d'obtention d'informations utilisateur d'intérêt sur la base de données d'image d'entrée et procédé de personnalisation de conception d'objet
WO2014092451A1 (fr) Dispositif et procédé de recherche d'informations et support d'enregistrement lisible par ordinateur associé
WO2019022472A1 (fr) Dispositif électronique et son procédé de commande
WO2019059505A1 (fr) Procédé et appareil de reconnaissance d'objet
WO2019139270A1 (fr) Dispositif d'affichage et procédé de fourniture de contenu associé
WO2019143227A1 (fr) Dispositif électronique produisant une image en rapport avec un texte et son procédé de fonctionnement
WO2020262746A1 (fr) Appareil à base d'intelligence artificielle pour recommander un parcours de linge, et son procédé de commande
WO2020032567A1 (fr) Dispositif électronique destiné à fournir des informations sur un élément sur la base d'une catégorie d'élément
WO2020222623A9 (fr) Système et procédé pour construire automatiquement un contenu destiné à des ventes stratégiques
EP3539056A1 (fr) Appareil électronique et son procédé de fonctionnement
WO2017030255A1 (fr) Appareil d'affichage grand format et son procédé de commande
WO2019054791A1 (fr) Procédé et appareil d'exécution de contenu
WO2021150033A1 (fr) Dispositif électronique et procédé de commande de dispositif électronique
WO2017131348A1 (fr) Appareil électronique et son procédé de commande
WO2016153200A2 (fr) Procédé d'essayage virtuel et programme informatique
WO2018182068A1 (fr) Procédé et appareil de fourniture d'informations de recommandation pour un article
WO2020171567A1 (fr) Procédé permettant de reconnaître un objet et dispositif électronique le prenant en charge
WO2017069422A1 (fr) Dispositif de terminal utilisateur et procédé destinés à la fourniture de service web associé
WO2021132922A1 (fr) Dispositif informatique et procédé de fonctionnement associé
WO2019054792A1 (fr) Procédé et terminal de fourniture de contenu
WO2021040256A1 (fr) Dispositif électronique et son procédé de recommandation de vêtements
WO2020138957A1 (fr) Procédé de recommandation dynamique de catalogue et dispositif électronique associé
WO2018043859A1 (fr) Dispositif d'analyse automatique de préférence d'article de location au moyen d'une image d'utilisateur et procédé correspondant

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 20856360

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 20856360

Country of ref document: EP

Kind code of ref document: A1