KR20210016593A - Method and system for recommending product based on artificial intelligence - Google Patents

Abstract

The present discloses a method and a system for recommending product based on artificial intelligence. In regard to the method and system for recommending product, it is possible to recommend a product matching a given product based on the degree of matching between products through a DNN learning model using a matching product set and a non-matching product set, and recommend the product of a coordination set similar to an actual coordination set for the given product based on a generation model using GAN. The present invention provides the method capable of providing a product coordinator function based on artificial intelligence.

Description

Artificial intelligence-based product recommendation method and its system {METHOD AND SYSTEM FOR RECOMMENDING PRODUCT BASED ON ARTIFICIAL INTELLIGENCE}

The description below relates to a technology that automatically recommends suitable products for a given product.

The fashion coordinator system is a system for recommending a product or service as a system that helps a user select one clothing that they want to wear, and then selects another appropriate clothing.

The system for recommending a product or service is largely divided into a collaborative filtering method using similarity between users and a content-based filtering method using characteristic values of products and users.

For example, in Korean Patent Registration No. 10-0882716 (registration date February 02, 2009), product information on products is automatically displayed using the product code, product name, or model name of the product recognized by the information recommendation agent installed on the user terminal. A recommended technique is disclosed.

It provides a method and system capable of providing product coordinator functions based on artificial intelligence.

Provides a method and system for recommending products that match a given product based on the degree of matching between products through a deep neural network (DNN) learning model using matching product sets and non-matching product sets.

Provides a method and system for recommending a product of a coordination set similar to an actual coordination set for a given product based on a generative model.

A computer-implemented product recommendation method, the method comprising: collecting a correct answer image set consisting of matching product images; And building a product recommendation model based on a generative model by learning the set of correct answer images, wherein the product recommendation model receives a given product image as an input and includes the input product image and the input product image through the generation model. In the step of generating information on the characteristics of other suitable products, determining a product similar to the generated characteristic as a recommended product, and building, in order to advance the generation model, a determination model for determining the authenticity of the generated result is combined together. A coordination set generated by a machine whether the discrimination model is an actual coordination set existing in the DB using the given product image and the recommended product image resulting from the generation of the generated model using the DB in which the actual coordination sets are accumulated It provides a product recommendation method, characterized in that it determines whether the recognition authenticity.

According to an aspect, in the product recommendation model, a latent space following a specific distribution is given as irregular noise to the generated model, and information indicating a product recommendation range for the given product image is specified and specified in the latent space. The recommended product may be determined within a product recommendation range.

According to another aspect, in the collecting step, individual product images may be recognized from images including a plurality of products, and a suitable product image set may be formed.

According to another aspect, the building may include learning the correct answer image set by using an encoder-decoder model as the generation model.

According to another aspect, the constructing may include learning the correct answer image set using a variational auto-encoder (VAE), a model having a conditional variational distribution as the generation model. have.

There is provided a computer program stored in a computer-readable recording medium to execute the product recommendation method on a computer.

There is provided a computer-readable recording medium in which a program for executing the above product recommendation method on a computer is recorded.

A computer system, comprising: at least one processor embodied to execute a command readable by a computer, the at least one processor comprising: an image collection unit for collecting a correct answer image set consisting of matching product images; And a learning processing unit for building a product recommendation model based on a generative model by learning the correct answer image set, wherein the product recommendation model receives a given product image as an input and the input product image through the generation model. It generates information on the characteristics of other products that match with and determines a product similar to the generated characteristics as a recommended product, and the learning processing unit includes a discrimination model for determining the authenticity of the generated result in order to advance the generation model. A coordination set generated by a machine whether the discrimination model is an actual coordination set existing in the DB using the given product image and the recommended product image resulting from the generation of the generated model using the DB in which the actual coordination sets are accumulated It provides a computer system, characterized in that it determines whether the recognition authenticity.

According to embodiments of the present invention, a product matching a given product may be recommended based on a degree of matching between products through a DNN learning model using a matching product set and a non-matching product set.

According to embodiments of the present invention, in recommending a coordination using a generation model, a product of a coordination set similar to an actual coordination set may be recommended for a given product through the advancement of the generation model using GAN.

1 is a diagram showing an example of a network environment according to an embodiment of the present invention.
2 is a block diagram illustrating an internal configuration of an electronic device and a server according to an embodiment of the present invention.
3 is a diagram illustrating an example of components that may be included in a processor of a server according to an embodiment of the present invention.
4 is a diagram illustrating a product recommendation method that can be performed by a processor of a server according to the first embodiment of the present invention.
5 is an exemplary diagram for explaining a process of collecting an image set for a correct answer.
6 is an exemplary diagram for describing a process of collecting an incorrect answer image set.
7 is an exemplary diagram showing a DNN-based product recommendation model according to the first embodiment of the present invention.
8 is a diagram showing a product recommendation method that can be performed by a processor of a server according to a second embodiment of the present invention.
9 is an exemplary diagram showing a product recommendation model based on a generation model according to a second embodiment of the present invention.
10 is an exemplary diagram for explaining a process of upgrading a generation model according to a second embodiment of the present invention through a GAN learning technique.
11 is an exemplary diagram for explaining a process of upgrading a generated model according to a second embodiment of the present invention through layer elaboration.

Hereinafter, embodiments of the present invention will be described in detail with reference to the accompanying drawings.

Embodiments of the present invention relate to a technology for automatically recommending suitable products for a given product.

The product recommendation method according to embodiments of the present invention may be performed through an electronic device or a computer device such as a server, which will be described later. In this case, a computer program according to an embodiment of the present invention may be installed and driven in the computer device, and the computer device may perform a product recommendation method according to an embodiment of the present invention under control of the driven computer program. . The above-described computer program may be combined with a computer device and stored in a computer-readable recording medium to execute a product recommendation method on a computer. For example, the server serves as a server that provides product information, and in particular, as an artificial intelligence-based product coordinator, it is possible to automatically recommend a product that matches the product for a given product. According to an embodiment, it may also be considered that the electronic device performs a product coordinator function.

Hereinafter, a specific embodiment of the fashion coordination recommendation model will be described, but the embodiments of the present invention are not limited thereto, and the embodiments of the present invention can be applied to various coordination recommendation models such as space coordination, home coordination, food coordination, and color coordination in addition to fashion coordination.

1 is a diagram showing an example of a network environment according to an embodiment of the present invention. The network environment of FIG. 1 shows an example including a plurality of electronic devices 110, 120, 130, and 140, a plurality of servers 150 and 160, and a network 170. 1 is an example for explaining the present invention, and the number of electronic devices or servers is not limited as in FIG. 1.

The plurality of electronic devices 110, 120, 130, and 140 may be a fixed terminal implemented as a computer device or a mobile terminal. Examples of the plurality of electronic devices 110, 120, 130, 140 include smart phones, mobile phones, navigation systems, computers, notebook computers, digital broadcasting terminals, personal digital assistants (PDAs), portable multimedia players (PMPs). ), tablet PC, game console, wearable device, internet of things (IoT) device, virtual reality (VR) device, and augmented reality (AR) device. As an example, in FIG. 1, the shape of a smartphone is shown as an example of the electronic device 1 110, but in the embodiments of the present invention, the electronic device 1 110 substantially connects the network 170 using a wireless or wired communication method. Through this, it may mean one of various physical computer devices capable of communicating with other electronic devices 120, 130, and 140 and/or the servers 150 and 160.

The communication method is not limited, and not only a communication method using a communication network (for example, a mobile communication network, wired Internet, wireless Internet, broadcasting network, satellite network, etc.) that the network 170 may include, but also short-range wireless communication between devices is included I can. For example, the network 170 includes a personal area network (PAN), a local area network (LAN), a campus area network (CAN), a metropolitan area network (MAN), a wide area network (WAN), and a broadband network (BBN). , Internet, and the like. In addition, the network 170 may include any one or more of a network topology including a bus network, a star network, a ring network, a mesh network, a star-bus network, a tree or a hierarchical network, etc. Not limited.

Each of the servers 150 and 160 is a computer device or a plurality of computers that communicates with a plurality of electronic devices 110, 120, 130, and 140 through a network 170 to provide commands, codes, files, contents, services, etc. It can be implemented with devices. For example, the server 150 may be a system that provides a first service to a plurality of electronic devices 110, 120, 130, 140 accessed through the network 170, and the server 160 is also a network ( It may be a system that provides a second service to a plurality of electronic devices 110, 120, 130, and 140 accessed through 170). As a more specific example, the server 150 is a service (for example, a product recommendation service, etc.) through an application as a computer program installed and driven in a plurality of electronic devices 110, 120, 130, 140 May be provided to a plurality of electronic devices 110, 120, 130, and 140 as a first service. As another example, the server 160 may provide a service for distributing the above-described application installation and running file to the plurality of electronic devices 110, 120, 130, and 140 as the second service.

2 is a block diagram illustrating an internal configuration of an electronic device and a server according to an embodiment of the present invention. In FIG. 2, internal configurations of the electronic device 1 110 and the server 150 are described as examples of the electronic device. In addition, other electronic devices 120, 130, 140 or the server 160 may also have the same or similar internal configuration as the electronic device 1 110 or the server 150 described above.

The electronic device 1 110 and the server 150 may include memories 211 and 221, processors 212 and 222, communication modules 213 and 223, and input/output interfaces 214 and 224. The memories 211 and 221 are non-transitory computer-readable recording media, such as random access memory (RAM), read only memory (ROM), disk drive, solid state drive (SSD), flash memory, and the like. It may include a permanent mass storage device. Here, a non-destructive mass storage device such as a ROM, SSD, flash memory, or disk drive may be included in the electronic device 1 110 or the server 150 as a separate permanent storage device that is separate from the memories 211 and 221. In addition, the memories 211 and 221 contain an operating system and at least one program code (for example, a browser installed and driven in the electronic device 1 110 or an application installed in the electronic device 1 110 to provide a specific service). Code) can be saved. These software components may be loaded from a computer-readable recording medium separate from the memories 211 and 221. Such a separate computer-readable recording medium may include a computer-readable recording medium such as a floppy drive, disk, tape, DVD/CD-ROM drive, and memory card. In another embodiment, software components may be loaded into the memories 211 and 221 through communication modules 213 and 223 other than a computer-readable recording medium. For example, at least one program is a computer program installed by files provided by a file distribution system (for example, the server 160 described above) that distributes the installation files of developers or applications through the network 170 It may be loaded into the memories 211 and 221 based on (for example, the above-described application).

The processors 212 and 222 may be configured to process instructions of a computer program by performing basic arithmetic, logic, and input/output operations. The instructions may be provided to the processors 212 and 222 by the memories 211 and 221 or the communication modules 213 and 223. For example, the processors 212 and 222 may be configured to execute a command received according to a program code stored in a recording device such as the memories 211 and 221.

The communication modules 213 and 223 may provide a function for the electronic device 1 110 and the server 150 to communicate with each other through the network 170, and the electronic device 1 110 and/or the server 150 A function for communicating with another electronic device (for example, electronic device 2 120) or another server (for example, server 160) may be provided. As an example, a request generated by the processor 212 of the electronic device 1 110 according to a program code stored in a recording device such as a memory 211 is transmitted through the network 170 under the control of the communication module 213. 150). Conversely, control signals, commands, contents, files, etc. provided under the control of the processor 222 of the server 150 are transmitted via the communication module 223 and the network 170 to the communication module of the electronic device 1 110 ( 213) through the electronic device 1 (110). For example, control signals, commands, contents, files, etc. of the server 150 received through the communication module 213 may be transmitted to the processor 212 or the memory 211, and the contents or files are electronic devices 1 110 may be stored in a storage medium (permanent storage device described above) that may further include.

The input/output interface 214 may be a means for an interface with the input/output device 215. For example, the input device may include a device such as a keyboard, a mouse, a microphone, and a camera, and the output device may include a device such as a display, a speaker, and a haptic feedback device. As another example, the input/output interface 214 may be a means for interfacing with a device in which input and output functions are integrated into one, such as a touch screen. The input/output device 215 may be configured with the electronic device 1110 and one device. In addition, the input/output interface 224 of the server 150 may be a means for an interface with a device (not shown) for input or output that is connected to the server 150 or that the server 150 may include. As a more specific example, the processor 212 of the electronic device 1 110 uses data provided by the server 150 or the electronic device 2 120 to process a command of a computer program loaded in the memory 211 The service screen or content to be displayed may be displayed on the display through the input/output interface 214.

In addition, in other embodiments, the electronic device 1 110 and the server 150 may include more components than those of FIG. 2. However, there is no need to clearly show most of the prior art components. For example, the electronic device 1 110 may be implemented to include at least some of the input/output devices 215 described above, or other components such as a transceiver, a global positioning system (GPS) module, a camera, various sensors, and a database. It may also contain more elements. As a more specific example, when the electronic device 1 110 is a smartphone, an acceleration sensor or a gyro sensor, a camera module, various physical buttons, buttons using a touch panel, input/output ports, and vibration Various components such as a vibrator may be implemented to be further included in the electronic device 1 (110).

Hereinafter, a specific embodiment of an artificial intelligence-based product recommendation method and system will be described.

3 is a block diagram illustrating an example of components that may be included in a processor of a server according to an embodiment of the present invention.

A product recommendation system may be configured in the server 150 according to the present embodiment. The product recommendation system can automatically find suitable products for a given product through image learning using DNN and GAN.

In order to perform the product recommendation method to be described below, the processor 222 of the server 150 is a component, as shown in FIG. 3, the image collection unit 310, the learning processing unit 320, and the product recommendation. It may include a unit 330. Depending on the embodiment, components of the processor 222 may be selectively included or excluded from the processor 222. Further, according to an embodiment, components of the processor 222 may be separated or merged to express the function of the processor 222.

The components of the processor 222 and the processor 222 may control the server 150 to perform steps included in the product recommendation method to be described below. For example, the processor 222 and the components of the processor 222 may be implemented to execute an instruction according to the code of the operating system included in the memory 221 and the code of at least one program.

Here, the components of the processor 222 may be expressions of different functions of the processor 222 performed by the processor 222 according to an instruction provided by the program code stored in the server 150. . For example, the image collection unit 310 may be used as a functional representation of the processor 222 that controls the server 150 according to the above-described command so that the server 150 collects images usable as training data.

Embodiment 1

4 is a flowchart illustrating a product recommendation method that can be performed by a processor of a server according to the first embodiment of the present invention.

In step S410, the processor 222 may read a required command from the memory 221 loaded with a command related to control of the server 150. In this case, the read command may include a command for controlling the processor 222 to execute steps S420 to S440 to be described later.

In step S420, the image collection unit 310 collects a product image set (hereinafter, referred to as'correct answer image set') as correct answer learning data, and a product image set that does not match as incorrect answer learning data (hereinafter,'incorrect answer image set). ') can be collected. First, the image collection unit 310 may collect a large amount of images in which a plurality of products are recognized in one image using an object detection technique. For example, it is possible to collect images of products coordinated by MD (merchandisers) or sellers such as fashion pictorials, magazines, shopping malls, etc., and images of coordinated products with a coordination evaluation score higher than a certain level by general users. It can be used as an image for correct answer coordination to create correct answer learning data. The image collection unit 310 may create a plurality of product images by imagely copying or cutting individual products from the corresponding image for each coordinated product image collected previously, and use them as a correct answer image set. In order for a machine to learn the harmony between products, it needs images of'comparable products' and'non-matching products' just like humans. If such work is not done logically, it can be judged that the machine does not match'comparable products'. Accordingly, the image collection unit 310 may collect an incorrect answer image set together with a correct answer image set as data for learning matching between products. As an example, the image collection unit 310 may collect an incorrect image set through negative sampling excluding a correct answer image set and a similar product image using correct answer image sets.

5 is an exemplary diagram for explaining a process of collecting an image set for a correct answer.

5, the image collection unit 310 acquires a product image 500 in which a plurality of products are coordinated. At this time, two objects recognized in the product image 500, that is, a top and a bottom, are'matched' with each other. It can be defined as'product'. In other words, the image collection unit 310 recognizes the top image 501 and the bottom image 502 from the product image 500 including a plurality of products, for example, a top and a bottom, and configures the correct answer image set. One correct answer image set consisting of a plurality of product images can be created by using one coordinating product image. Through this method, the image collection unit 310 may accumulate correct answer image sets for use as learning data by generating a matching product image set for each coordinating product image.

6 is an exemplary diagram for describing a process of collecting an incorrect answer image set. For example, as the correct answer image sets are accumulated, we have the image sets of the bottom B that match the top A. In the following description, matching products are indicated by'>' and similar products are indicated by'~'.

6 shows a process of investigating bottoms that do not fit the top A0.

First, in step S61, the image collection unit 310 searches for products A1 and A2 similar to the top A0 (A0 to A1, A2). Similar products can be found based on the similarity between product images of correct answer image sets.For example, through image analysis, features related to color, texture, and other attributes are extracted and based on this The degree of similarity between images can be determined. This is exemplary, and the method of calculating the image similarity may use at least one of various known algorithms.

Next, in step S62, the image collection unit 310 checks the bottom B0 in the correct answer image set including the top A0 (A0>B0), and the similar product is included in each of the products A1 and A2 similar to the top A0. In the set of correct answer images, check the bottoms B1 and B2 (A1>B1, A2>B2).

Subsequently, in step S63, the image collection unit 310 searches for similar products for each of the bottoms B0, B1, and B2 ([B0~B0'], [B1~B1'], [B2~B2', B2") ]).

Finally, the image collection unit 310 defines that the bottoms matching the top A0 as well as the bottoms matching the top A0, the bottoms matching other tops similar to the top A0, and other bottoms similar to these bottoms all match the top A0 (A0 >B0,B0',B1,B1',B2,B2',B2"), other products are defined as not matching the top A0.

In this way, the image collection unit 310 may sample at least some of the product images other than the correct answer image set and the similar product image to create an image set that does not match the image A0 (negative sampling).

In FIG. 4 again, in step S430, the learning processing unit 320 may build a product recommendation model including a matching learning result between images by performing machine learning by using the correct answer image set and the incorrect answer image set together. At this time, machine learning uses DNN, the input of the model is two product images for determining the degree of matching, and the output outputs the degree of matching for the two product images. For example, referring to FIG. 7, the learning processing unit 320 receives an upper image 701 and a lower image 702 as inputs, and vectorizes the features extracted from each image 701 and 702, and The DNN model 700 can be trained to map the image vectors of and below to a vector in a multidimensional vector space, but in the case of a correct answer image set, the distance between the two image vectors becomes close, and in the case of an incorrect answer image set, the distance between the two image vectors is increased. .

Again, in step S440 of FIG. 4, the product recommendation unit 330 may automatically recommend a product that matches the given product through a DNN-based product recommendation model. When a new product image is input, the product recommendation unit 330 passes the product recommendation model and calculates all the matching degrees between the corresponding product image and the candidate product images, and recommends the product with the highest matching degree or a plurality in the order of matching degree. Can recommend dog products. In other words, the product recommendation unit 330 vectorizes the characteristics extracted from a given product image, maps it to a vector in a multidimensional vector space, and calculates a degree of matching based on the distance with another image vector in the vector space. According to this, recommended products can be provided.

Accordingly, the product recommendation method according to the present invention recognizes a plurality of products from one image and learns them as a matching product set on the assumption that a coordinating product consists of matching products. In particular, by performing machine learning on the matching product image set and the non-matching product image set using the DNN method, more accurate and effective product recommendation is possible through a model that learns not only matching products but also non-matching products.

Embodiment 2

8 is a flowchart illustrating a product recommendation method that can be performed by a processor of a server according to a second embodiment of the present invention.

In step S810, the processor 222 may read a required command from the memory 221 loaded with a command related to control of the server 150. In this case, the read command may include a command for controlling the processor 222 to execute steps S820 to S840 to be described later.

In step S820, the image collection unit 310 may collect a matching product image set as data for learning the matching between products. The image collection unit 310 may mass-collect images in which a plurality of products are recognized in one image using an object recognition technology. For example, it is possible to collect images of products coordinated by MDs or sellers such as fashion pictorials, magazines, shopping malls, etc., and images of coordinated products with a coordination evaluation score higher than a certain level by general users. It can be used as learning data for learning. The image collection unit 310 may accumulate image sets for use as learning data by generating a matching product image set for each coordinating product image. The detailed collection process of learning data is as described with reference to FIG. 5.

In step S830, the learning processing unit 320 may build a product recommendation model based on the generation model by learning the product image set collected as the training data. In this case, the product recommendation model may receive a product image as an input, generate information on a characteristic of another product matching the input product, and then determine a product similar to the generated characteristic as a recommended product. The generation model vectorizes the features extracted from a given product image and makes it an input vector. The image feature vector may be expressed in the form of an embedding vector through a pre-trained image learning model, and may be used as a feature of a product recommendation model. In addition, the generation model derives an output vector expressing the characteristics of a suitable product and outputs it. The learning direction of the product recommendation model is to learn a relationship between products that match each other by extracting feature vectors for each product from a matching product image set.

For example, as shown in FIG. 9, an encoder-decoder model used in language processing such as machine translation may be used as the generation model 900. For example, the learning processing unit 320 receives a vector corresponding to the upper image 901 and generates a vector of the lower image 902 that is suitable through the generation model 900 so that these vectors are the closest distance in the vector space. Learn to have.

In FIG. 8 again, in step S840, the product recommendation unit 330 may automatically recommend a product suitable for a given product through a product recommendation model based on a generation model. In the product recommendation model based on the generation model, an output vector expressing the characteristics of the product suitable for the product is derived from the characteristic vector of a given product image, and the distance between the derived output vector and the characteristic vector extracted from the candidate product images is , For example, a product image having the closest distance between vectors may be provided as a recommended product.

Further, the learning processing unit 320 may learn together with a discriminator that determines the authenticity of the generated result by using the GAN in order to advance the generated model 900, which is also called adversarial training. do. GAN is mainly used in image processing fields such as image synthesis or text-to-image synthesis, and artificial intelligence dialog that improves conversation performance by determining whether it is machine language or human language. And, in the present invention, it is used to advance the generation model for product recommendation. The discrimination model can determine whether a given image set is generated as a result of the generation model or corresponds to an actual coordination set in the DB using the DB in which the coordination sets are accumulated. Referring to FIG. 10, the learning processing unit 320 generates a product image 901 and a generation model 900 given by using a matching product image set, that is, a DB 1010 in which real data are accumulated. A discrimination model 1000 that determines whether the recommended product image 902 is an actual coordination set or a machine-generated coordination set (generated data) is trained.

By performing the advancement of the generative model using such a GAN, artificial mass production of negative image sets is unnecessary in order to learn the characteristics to resemble similar to the trend of the coordinating set that real humans think are suitable. Even with image sets, high product recommendation performance can be expected in a semi-supervised method.

In addition, it is also possible to enhance the structure of the generation model 900 itself in addition to confrontational learning for learning the discrimination model 1000. As an example, there is a method of elaborating a layer of the generated model 900, and signals such as color, pattern, and style of a product can be compressedly utilized through a convolution layer and a max pooling layer. As another example, a Highway layer may be added to the generation model 900 for improving learning performance or speed, or a variational auto-encoder (VAE), a model having a conditional variational distribution, may be used. Referring to FIG. 11, a latent space (Z) 1100 that follows a specific distribution as random noise is assigned to a generated model 900 composed of an encoder-decoder model, thereby enabling precise estimation. In other words, by specifying information indicating the product recommendation range in the potential space 1100 of the generation model 900, products within the range may be recommended. For example, when a top is given, it is possible to recommend sophisticated products within the specified category by specifying information on at least one category to be recommended from among product classification categories such as bottoms, bags, shoes, and accessories. In addition, through mixed modeling using deep learning and probabilistic machine learning, it is possible to easily process learning of a probability distribution space, which is difficult with only a simple DNN.

As described above, according to embodiments of the present invention, a product matching a given product may be recommended based on a degree of matching between products through a DNN learning model using a matching product set and a non-matching product set. In addition, according to embodiments of the present invention, in recommending a coordination using a generated model, a product having a coordination set similar to an actual coordination set may be recommended for a given product through the advancement of the generated model using GAN.

The apparatus described above may be implemented as a hardware component, a software component, and/or a combination of a hardware component and a software component. For example, the devices and components described in the embodiments are a processor, a controller, an arithmetic logic unit (ALU), a digital signal processor, a microcomputer, a field programmable gate array (FPGA), a programmable gate array (PLU). It may be implemented using one or more general purpose computers or special purpose computers, such as a logic unit), a microprocessor, or any other device capable of executing and responding to instructions. The processing device may execute an operating system (OS) and one or more software applications executed on the operating system. In addition, the processing device may access, store, manipulate, process, and generate data in response to the execution of software. For the convenience of understanding, although it is sometimes described that one processing device is used, one of ordinary skill in the art, the processing device is a plurality of processing elements and/or a plurality of types of processing elements. It can be seen that it may include. For example, the processing device may include a plurality of processors or one processor and one controller. In addition, other processing configurations are possible, such as a parallel processor.

The software may include a computer program, code, instructions, or a combination of one or more of these, configuring the processing unit to behave as desired or processed independently or collectively. You can command the device. Software and/or data may be embodyed in any type of machine, component, physical device, computer storage medium or device to be interpreted by the processing device or to provide instructions or data to the processing device. have. The software may be distributed over networked computer systems and stored or executed in a distributed manner. Software and data may be stored on one or more computer-readable recording media.

The method according to the embodiment may be implemented in the form of program instructions that can be executed through various computer means and recorded in a computer-readable medium. In this case, the medium may be one that continuously stores a program executable by a computer, or temporarily stores a program for execution or download. In addition, the medium may be a variety of recording means or storage means in a form in which a single or several pieces of hardware are combined, but is not limited to a medium directly connected to a computer system, but may be distributed on a network. Examples of media include magnetic media such as hard disks, floppy disks, and magnetic tapes, optical recording media such as CD-ROMs and DVDs, magnetic-optical media such as floptical disks, and And a ROM, RAM, flash memory, and the like, and may be configured to store program instructions. In addition, examples of other media include an app store that distributes applications, a site that supplies or distributes various software, and a recording medium or a storage medium managed by a server.

As described above, although the embodiments have been described by the limited embodiments and the drawings, various modifications and variations are possible from the above description to those of ordinary skill in the art. For example, the described techniques are performed in a different order from the described method, and/or components such as a system, structure, device, circuit, etc. described are combined or combined in a form different from the described method, or other components Alternatively, even if substituted or substituted by an equivalent, an appropriate result can be achieved.

Therefore, other implementations, other embodiments, and claims and equivalents fall within the scope of the claims to be described later.

Claims (8)

In the computer-implemented product recommendation method,
Collecting a correct answer image set consisting of matching product images; And
Building a product recommendation model based on a generative model by learning the correct answer image set
Including,
The product recommendation model receives a given product image as an input, generates information on a characteristic of another product matching the input product image through the generation model, and determines a product similar to the generated characteristic as a recommended product,
The step of constructing,
In order to advance the generated model, a discrimination model for determining the authenticity of the generated result is learned together, and the discrimination model uses the DB in which the actual coordination sets are accumulated, and the given product image and the recommended product that is the result of the generation To determine whether the image is an actual coordination set existing on the DB or a coordination set generated by a machine.
Product recommendation method, characterized in that. The method of claim 1,
In the product recommendation model, a latent space following a specific distribution is given as irregular noise to the generated model, and information indicating a product recommendation range for the given product image is specified in the potential space, and within the specified product recommendation range. Determining the recommended product
Product recommendation method, characterized in that. The method of claim 1,
The collecting step,
Recognizing individual product images from images containing multiple products and composing them into a set of matching product images
Product recommendation method, characterized in that. The method of claim 1,
The step of constructing,
Learning the correct answer image set using an encoder-decoder model as the generation model
Product recommendation method comprising a. The method of claim 1,
The step of constructing,
Learning the correct answer image set using a VAE (variational auto-encoder) which is a model having a conditional variational distribution as the generation model
Product recommendation method comprising a. A computer program stored in a computer-readable recording medium for executing the product recommendation method of any one of claims 1 to 5 on a computer. A computer-readable recording medium in which a program for causing a computer to execute the product recommendation method of any one of claims 1 to 5 is recorded. In a computer system,
At least one processor implemented to execute computer-readable instructions
Including,
The at least one processor,
An image collection unit that collects a correct answer image set consisting of matching product images; And
Learning processing unit for building a product recommendation model based on a generative model by learning the correct answer image set
Including,
The product recommendation model receives a given product image as an input, generates information on a characteristic of another product matching the input product image through the generation model, and determines a product similar to the generated characteristic as a recommended product,
The learning processing unit,
In order to advance the generated model, a discrimination model for determining the authenticity of the generated result is learned together, and the discrimination model uses the DB in which the actual coordination sets are accumulated, and the given product image and the recommended product that is the result of the generation of the generated model To determine whether the image is an actual coordination set existing on the DB or a coordination set generated by a machine or not
Computer system, characterized in that.

Images