Method for clustering fashion item

Abstract

To provide a method in coordinating fashion for replacing a fashion item that is out of stock with a fashion item that is in stock or with an item that has a higher profit rate, while keeping the style of the coordination.SOLUTION: One aspect of the present embodiment is a method for supporting fashion selection. In the method, a vector is generated for each classification of a tag that is learned in advance based on image data of a received commodity, and a clustering model is generated based on the generated vector. The method also includes steps of extracting commodities based on a difference between a vector created from the commodity image and a vector of commodities included in the clustering model and transmitting the extracted commodity data to a user terminal of a user.SELECTED DRAWING: Figure 8

Description

[Technical field]

[0001] The present invention relates to a method for clustering fashion items. [Background technology]

[0002] Traditionally, in the world of clothing retail, inventory management has mainly been carried out by classifying product inventory and products actually on display in stores by units such as size and color, and various technologies have been proposed for selling appropriate clothing to customers.

[0003] For example, Patent Document 1 discloses a technology that aims to realize a technology for proposing appropriate products according to customer attributes. [Prior art documents] [Patent documents]

[0004] [Patent Document 1] Patent Publication No. 2022-064035 Summary of the Invention [Problem to be solved by the invention]

[0005] However, even though the technology disclosed in Patent Document 1 can suggest appropriate products according to the attributes of a customer, it is limited to processing assuming only that the suggested products are in stock. On the other hand, in actual clothing sales, when a suggested product is out of stock, there is a growing need to replace it with a product that is in stock, and furthermore, from the business side, there is an increasing need to suggest products with higher profit margins.

[0006] Therefore, an object of the present invention is to replace out-of-stock fashion items with in-stock fashion items or with items with higher profit margins when putting together an outfit while maintaining the style of the outfit. [Means for solving the problem]

[0007] In one aspect of the present invention, a method is provided for generating a vector for each pre-trained tag classification based on image data of a product. Effect of the Invention

[0008] According to the present invention, when putting together an outfit, it is possible to replace an out-of-stock fashion item with an in-stock fashion item, or to replace it with an item with a higher profit margin, while maintaining the style of the outfit. [Brief description of the drawings]

[0009] [Figure 1] 1 is a block diagram showing a system for supporting fashion selection according to a first embodiment of the present invention. [Diagram 2] 2 is a functional block diagram showing the server terminal 100 of FIG. 1. [Diagram 3] 2 is a functional block diagram showing a user terminal 200 of FIG. 1. [Figure 4] FIG. 2 is a diagram showing an example of user data stored in the server 100. [Diagram 5] FIG. 2 is a diagram showing an example of learning data stored in the server 100. [Figure 6] 3 is a flowchart showing an example of a method for supporting fashion selection according to the first embodiment of the present invention. [Figure 7] 10 is a flowchart showing another example of the method for supporting fashion selection according to the first embodiment of the present invention. [Figure 8] 11 is a flowchart showing yet another example of the method for supporting fashion selection according to the first embodiment of the present invention. [Figure 9] 1 is an example of image processing of a fashion item according to the first embodiment of the present invention. [Figure 10] 1A to 1C are diagrams illustrating a process of generating a vector from an image of a fashion item according to a first embodiment of the present invention. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0010] Hereinafter, an embodiment of the present invention will be described with reference to the drawings. Note that the embodiment described below does not unduly limit the contents of the present invention described in the claims. Furthermore, not all of the components shown in the embodiment are necessarily essential components of the present invention.

[0011] <Configuration> 1 is a block diagram showing a system for supporting fashion selection according to a first embodiment of the present invention. The system 1 includes a server terminal 100 that receives image information on a specific product or service from a user and extracts products based on the difference with a vector included in a clustering model that has already been trained, and user terminals 200A and 200B that are managed by the user. For convenience of explanation, each terminal is described as being a single terminal or a specific number of terminals, but the number of each terminal is not limited.

[0012] The server terminal 100 and the user terminals 200A and 200B are connected to each other via a network NW1. The network NW is configured by the Internet, an intranet, a wireless LAN (Local Area Network), a WAN (Wide Area Network), or the like.

[0013] The server terminal 100 may be, for example, a general-purpose computer such as a workstation or a personal computer, or may be configured as a cloud computing system, a cluster or multicomputer consisting of multiple computers, a virtual machine virtually constructed by software, or a quantum computer.

[0014] The user terminals 200A and 200B are, for example, information processing devices such as personal computers and tablet terminals, but may also be configured as smartphones, mobile phones, PDAs, or the like.

[0015] In this embodiment, the system 1 is described as having a configuration including a server terminal 100 and user terminals 200A, 200B, and users of each terminal use their respective terminals to operate the server terminal 100, but the server terminal 100 may be configured as a stand-alone system, and the server terminal itself may have a function for each user to directly operate it. Hereinafter, the user terminals 200A, 200B will be collectively referred to as user terminal 200, as necessary.

[0016] Fig. 2 is a functional block configuration diagram of the server terminal 100 of Fig. 1. The server terminal 100 includes a communication unit 110, a storage unit 120, and a control unit .

[0017] The communication unit 110 is a communication interface for communicating with the user terminal 200 and the company terminal 300 via the network NW1, and communication is performed according to a communication protocol such as TCP / IP (Transmission Control Protocol / Internet Protocol).

[0018] The storage unit 120 stores programs for executing various control processes and various functions in the control unit 130, input data, etc., and is composed of a RAM (Random Access Memory), a ROM (Read Only Memory), etc. The storage unit 120 also has a user data storage unit 121 that stores various data related to users, and a learning data storage unit 122 that stores a clustering model generated by machine learning of product data, etc. A database (not shown) that stores various data may be constructed outside the storage unit 120 or the server terminal 100.

[0019] The control unit 130 controls the overall operation of the server terminal 100 by executing a program stored in the storage unit 120, and is composed of a CPU (Central Processing Unit), a GPU (Graphics Processing Unit), etc. Functions of the control unit 130 include an information receiving unit 131 that receives information from each user terminal, and an information processing unit 132 that processes the information. The information receiving unit 131 and the information processing unit 132 are started by a program stored in the storage unit 120 and executed by the server terminal 100, which is a computer (electronic calculator).

[0020] The information receiving unit 131 receives information from the user terminal 200 via the communication unit 110. For example, information on a specific product is received from the user terminal 200 via a user interface displayed on the user terminal 200.

[0021] The information processing unit 132 analyzes the product information received from the user terminal 200, and performs predetermined information processing, such as proposing products based on the object model.

[0022] The control unit 130 may also have a screen generation unit (not shown), which generates screen information to be displayed via the user interface of the user terminal 200 upon request. For example, using image and text data (not shown) stored in the storage unit 120 as material, the control unit 130 generates a user interface by arranging various images and text in predetermined areas of the user interface based on predetermined layout rules. Processing related to the image generation unit may also be executed by a GPU (Graphics Processing Unit).

[0023] Fig. 3 is a functional block diagram showing the user terminal 200 of Fig. 1. The user terminal 200 includes a communication unit 210, a display operation unit 220, a storage unit 230, and a control unit 240.

[0024] The communication unit 210 is a communication interface for communicating with the server terminal 100 via the network NW, and communication is performed according to a communication protocol such as TCP / IP.

[0025] The display operation unit 220 is a user interface used for inputting instructions by the provider, analyzing the input data according to the input data such as text, audio, and images from the control unit 240, and displaying the text, audio, and images as output, and is composed of a display, keyboard, and mouse when the user terminal 200 is configured as a personal computer, and is composed of a touch panel, etc. when the user terminal 200 is configured as a smartphone or tablet terminal. This display operation unit 220 is started up by a control program stored in the storage unit 230 and executed by the user terminal 200, which is a computer (electronic calculator).

[0026] The storage unit 230 stores programs for executing various control processes and functions in the control unit 240, input data, etc., and is composed of a RAM, a ROM, etc. The storage unit 230 also temporarily stores the contents of communication with the server terminal 100.

[0027] The control unit 240 controls the overall operation of the user terminal 200 by executing the programs stored in the storage unit 230, and is composed of a CPU, a GPU, and the like.

[0028] FIG. 4 is a diagram showing an example of user data stored in the server 100. As shown in FIG.

[0029] The user data 1000 shown in Fig. 4 stores various data related to users. For convenience of explanation, Fig. 4 shows an example of one user (for example, a user identified by a user ID "10001"), but information on multiple users can be stored. Examples of various data related to a user may include, but are not limited to, basic information about the user (e.g., login ID, password, user name, name, address, age, contact information, occupation), coordination data (clothing preferences, favorite brands, frequency of use of different clothing styles (pants, skirts, dresses), favorite colors, requests for stylist suggestions (style similar to your usual self and preferences, suggestions of items different from your usual style), items / designs / silhouettes / colors / patterns / materials to avoid, etc.), budget, physical information (height, weight, shoe size), body type (torso length, shoulder width, arm width, arm length, bust size, stomach size, hip size, thigh width, leg length), bone structure (thick upper body, high bust top, center of gravity seems somewhat high), etc.), stylist knowledge data (past coordination suggestions by stylists, etc.), and data on products purchased by the user.

[0030] FIG. 5 is a diagram showing an example of the learning data stored in the server 100. As shown in FIG.

[0031] The learning data 2000 stores various data related to fashion items. The various data includes, for example, product data related to products registered as products sold by the service provider who operates the server terminal 100, vectorized data of information included in the images of the product data, group data of similar vectors, and a set of exchangeable product data (RKU: Resemblance Keeping Unit) for coordinating outfits.

[0032] <Processing flow> FIG. 6 is a flowchart showing an example of a method for supporting fashion selection according to the first embodiment of the present invention.

[0033] First, in the process of step S101, the control unit 130 of the server terminal 100 performs a learning process in the information processing unit 132 to generate a clustering model based on the product data stored in the learning data 122 of the storage unit 120. This process will be described in detail with reference to FIG.

[0034] Next, in the process of step S102, the information receiving unit 131 receives the product image of the fashion item received from the user terminal 200. The information processing unit 132 of the server terminal 100 performs an inference process on products similar to the received product image based on the clustering model generated in step 101. This will also be described in detail with reference to FIG. 8.

[0035] FIG. 7 is a flowchart showing another example of the method for supporting fashion selection according to the first embodiment of the present invention.

[0036] First, in the process of step S201, the information processing unit 132 refers to the product data in the learning data storage unit 122 of the memory unit 120. In particular, in this example, image data of the product is referred to.

[0037] Next, in step S202, the information processing unit 132 performs image recognition processing on the reference image data. Specifically, as shown in Fig. 9, semantic segmentation is performed on the image downloaded from the product URL, and only pixels of a specified category, such as tops, bottoms, and outerwear, are extracted.

[0038] Next, in step S203, the information processing unit 132 performs vectorization on the extracted pixels based on a predetermined tag classification. Specifically, as shown in Fig. 10, for each tag classification (e.g., a function for creating a compressed representation of color, pattern, category, sleeve length, and material) previously learned by metric learning, the output of semantic segmentation that brings similar attributes closer together is mapped, and a vector for each tag classification is obtained. Furthermore, each vector is combined to obtain one vector.

[0039] Next, in the process of step S204, the information processing unit 132 of the control unit 130 maps the product information to a vector space in which the distances between exchangeable products are close to each other based on the vector acquired in step S203, and generates a clustering model. Here, the clustering model is preferably mapped to configure the above-mentioned exchangeable product data (RKU), and can also be associated with structured data that manages the RKU.

[0040] In the process of step S205, the information processing unit 132 stores the clustering model created in step S204 in the learning data storage unit 122 of the memory unit 120. As described above, by inputting the product data managed by the business operator and executing this learning process, a clustering model that allows interchangeable product data to be mutually referenced is generated.

[0041] FIG. 8 is a flowchart showing yet another example of the method for supporting fashion selection according to the first embodiment of the present invention.

[0042] In step S301, the information receiving unit 131 receives, from the user terminal 200, a product image of a fashion item such as the user's casual clothes.

[0043] In step S302, the information processing unit 132 performs image recognition processing based on the received product image. Specifically, as shown in Fig. 9, semantic segmentation is performed on the image downloaded from the product URL, and only pixels of a specified category, such as tops, bottoms, and outerwear, are extracted.

[0044] In step S303, the information processing unit 132 performs vectorization based on a predetermined tag classification for the pixels of the category extracted in step S302. Specifically, as shown in Fig. 10, the output of semantic segmentation that brings similar attributes closer together for each tag classification (e.g., a function that creates a compressed representation of color, pattern, category, sleeve length, and material) previously learned by metric learning is mapped, and a vector for each tag classification is obtained. Furthermore, each vector is combined to obtain one vector.

[0045] In the process of step S304, the information processing unit 132 refers to the clustering model stored in the data storage unit 122 of the recording unit 120, and extracts products with similar values ​​based on the difference between the vector acquired in step S303 and the vector included in the clustering model. The information processing unit 132 can transmit information on products with high similarity values ​​(higher scores) among the extracted products to the user terminal. The user can purchase the product by selecting a details button, a purchase button, or the like for the product displayed on the user terminal 200. As described above, in response to input of product data to be coordinated (in this example, an image of the user's private clothes), exchangeable items can be extracted and proposed to the user based on the generated clustering model.

[0046] In addition, even if the service provider operating the server terminal 100 does not have in stock the fashion items coordinated with the user, the service provider can input data (images, etc.) relating to the fashion items and use the means detailed in Figure 8 to extract similar products while maintaining the overall style of the coordination and suggest them to the user.

[0047] Although the embodiments of the present invention have been described above, they can be embodied in various other forms and can be implemented with various omissions, substitutions, and modifications. These embodiments and modifications, as well as omissions, substitutions, and modifications, are included in the technical scope of the claims and their equivalents. [Explanation of symbols]

[0048] 1 System 100 Server terminal, 110 Communication unit, 120 Storage unit, 130 Control unit, 200 User terminal, 300 Corporate user terminal, NW1 Network

Claims

1. A method for supporting fashion selection, comprising generating a vector for each pre-learned tag classification based on image data of a product.

2. The method of claim 1 , further comprising generating a clustering model based on the generated vectors.

3. The method according to claim 1 , wherein a product is extracted based on a difference between a vector created from a product image transmitted from a user terminal and a vector of the product included in the clustering model.

4. The method of claim 1 , further comprising transmitting the extracted product data together with the product to a user terminal of the user.

5. The method according to claim 1 , wherein the control unit transmits a product purchase page to the user terminal in response to a request from the user terminal to view the product data.

Images