KR102388944B1 - System and method for recommending interior remodeling style using non-contact type - Google Patents

Abstract

The present invention relates to a non-face-to-face interior remodeling style recommendation system and method. The interior remodeling style recommendation system of the present invention receives an interior image of a current living space and interior taste information about an interior style preferred by a customer from the customer through a communication network. The interior remodeling style recommendation system of the present invention extracts interior configurations for each living space through image analysis and object recognition from interior images, and analyzes and diagnoses the interior status of the customer. The interior remodeling style recommendation system of the present invention selects and recommends an interior remodeling style suitable for the customer's living space and taste by using an expert recommendation model of an expert group. According to the present invention, customers can consult on the interior anytime, anywhere in a non-face-to-face manner. By recommending a customized interior remodeling style suitable for a customer's interior condition, it is possible to enhance customer satisfaction with the interior remodeling style.

Description

Non-face-to-face interior remodeling style recommendation system and method {SYSTEM AND METHOD FOR RECOMMENDING INTERIOR REMODELING STYLE USING NON-CONTACT TYPE}

The present invention relates to a non-face-to-face interior remodeling style recommendation system and method, and more specifically, by receiving an interior image of a customer's living space and interior taste information of an interior style preferred by the customer, image analysis and object recognition A non-face-to-face interior remodeling style recommendation system and method that extracts interior composition using technology and recommends a customized interior remodeling style that responds to customers' interior images and interior taste information using a clustering artificial intelligence algorithm will be.

The interior design of a building is to design an aesthetic and rational interior space, and it is an important architectural factor that affects the living, physical health and emotional satisfaction of residents. It is a total design task that creates a harmonious and functional new space in consideration of the correlation between architecture, the city, and the environment.

For example, the interior plays a very important function in changing the space inside a building, and the elements constituting the interior, such as a room, living room, furniture, wallpaper, floor, kitchen equipment, lighting equipment, entrance hall, bathroom, windows, A veranda, etc., can be changed in an appropriate and harmonious manner according to the needs of consumers or consumers. Therefore, it is desirable to construct the interior in consideration of the functional characteristics and design concept of the interior space, as well as the cost and time as well as the structure and movement. In addition, as opportunities and media to experience various interior design examples are abundant, the importance of interior design and consumer interest are increasing.

However, consumers' interior tastes differ from individual to individual, and the interior space needs to be designed differently according to the consumer's taste or lifestyle. Therefore, if consumers can participate from the interior design process, they can directly select the interior design they prefer, increasing the value of interior design and customer satisfaction.

However, most of the housing market, especially the apartment housing market, is traditionally a supplier-oriented market, so it is difficult to reflect the opinions of the end user, the resident. By guessing or predicting a preferred interior style and selecting an interior design, it is difficult to secure user satisfaction with the interior design.

In general, consumers start by meeting with an interior contractor for interior construction. To this end, consumers have to have several face-to-face consultations through multiple interior contractors, make an appointment with interior contractors, and repeatedly request information or requests for the interior style that consumers want from different interior contractors several times. There are inconveniences such as having to inform you of matters. In addition, it is often the case that the quality of interior consultation varies greatly depending on the experience and experience of consulting experts at interior construction companies. And since most interior construction companies are formed in the center of cities, for example, consumers in rural areas or small towns have a limited choice of interior construction companies.

Recently, the remodeling market for existing houses is increasing instead of the decrease in new construction works, and the size of the indoor interior market is growing rapidly. However, as described above, since the interior market is a market centered on companies rather than consumers, interior remodeling also has disadvantages such as construction defects and after-sales service, and disputes about this are continuously occurring. In addition, since construction quality varies greatly depending on the experiences and work methods of individual interior contractors, consumer complaints are always latent.

Korean Patent Publication No. 10-2067645 (published on January 17, 2020) Korean Patent Publication No. 10-2276091 (published on July 12, 2021) Korean Patent Publication No. 10-1869356 (published on June 20, 2018) Korean Patent Publication No. 10-2021-0083573 (published on July 07, 2021)

An object of the present invention is to analyze and diagnose an interior image of a customer's living space by using a non-face-to-face object recognition technology, and to recommend a customized interior remodeling style optimized for a customer using a clustering artificial intelligence algorithm. It is to provide an interior remodeling style recommendation system and method of the method.

Another object of the present invention is to receive an image of the customer's interior, recognize the interior configuration, analyze the interior style according to the customer's preference, and match it with interior remodeling style information recommended by a group of experts to customize the interior remodeling style to the customer To provide a non-face-to-face interior remodeling style recommendation system and method for recommending

Another object of the present invention is to cluster the customer's interior style by analyzing and learning the customer's interior information, and predicting and recommending the customer's preferred interior remodeling style by comparing and analyzing the customer's feedback information and the interior remodeling style of the expert group. It is to provide a non-face-to-face interior remodeling style recommendation system and method therefor.

In order to achieve the above objects, the non-face-to-face interior remodeling style recommendation system of the present invention receives interior images for each living space and interior taste information for the interior styles preferred by the customer from the customer, and performs object recognition and analysis from the interior image. It diagnoses the interior condition of the customer through the analysis and uses the expert recommendation model for various interior styles recommended in advance by a group of experts to recommend an interior remodeling style suitable for the customer's living space and taste. The non-face-to-face interior remodeling style recommendation system of the present invention allows customers to consult on interior anytime and anywhere in a non-face-to-face manner, and recommends a customized interior remodeling style suitable for the customer's living space and taste. satisfaction can be increased.

The non-face-to-face interior remodeling style recommendation system of the present invention according to this feature includes; an interior recommendation server connected to each of a plurality of customer terminals through a communication network; When the interior image of the living space is transmitted, interior taste analysis information for analyzing the interior style preferred by the customer is transmitted to the customer terminal, and interior preference information selected from the interior preference analysis information is transmitted from the customer terminal to the interior style Interior request processing unit receiving a request for a recommendation; The interior image transmitted to the interior request processing unit is recognized and analyzed to extract the interior configuration for a plurality of living spaces, extract the inference information that infers the meaning of the extracted interior configuration, and the extracted interior configuration and inference information an interior image analysis unit for generating interior preference information from the interior request processing unit as interior configuration information of a customer; The expert recommendation model for a plurality of interior styles generated after being recommended by a group of experts in advance is grouped and classified, and the interior configuration information of the customer generated from the interior image analysis unit responds to the expert recommendation model to provide an interior according to the customer's taste an interior style classification unit that groups and classifies styles; And from the interior style classification unit, the interior style according to the customer's taste is compared with the expert recommendation model, and a customized interior remodeling style is selected and recommended to the customer, and the recommended interior remodeling style, company information on the interior construction company, and construction and an interior style recommendation unit that generates recommendation report information including estimate information including period, construction cost, and construction materials and transmits it to the customer terminal.

In this feature, the interior request processing unit; When the recommendation report information generated from the interior style recommendation unit is transmitted to the customer terminal, and feedback information including at least satisfaction with the recommended interior remodeling style of the recommendation report information is transmitted from the customer terminal, the feedback information is reflected The report information is transmitted to the interior style recommendation unit to generate the report information.

In this feature, the interior style recommendation unit extracts and recommends the interior remodeling style having the highest probability and the interior configuration information of the customer according to the set probability distribution of the expert recommendation model.

In this feature, the interior recommendation server includes at least an entrance, a living room, a kitchen, a room, a bathroom, and a window / veranda in the customer's living space, and provides a plurality of interior styles for each of the living spaces. Modern, natural, Nordic , classic, modern classic, and casual at least, and recommends an interior remodeling style by classifying and comparing the expert recommendation model and the customer's interior configuration information according to the color combination of each interior style and the structure of the living space.

According to another feature of the present invention, there is provided a processing method of a non-face-to-face interior remodeling style recommendation system.

The processing method of the non-face-to-face interior remodeling style recommendation system according to this feature includes: receiving, by an interior recommendation server, an interior image for a customer's living space from a customer terminal through a communication network, and receiving a request for a recommendation of an interior remodeling style; The interior recommendation server analyzes the interior image to recognize the interior configuration, extracts the recognized interior configuration for each living space, and transmits interior taste analysis information for the customer's preferred interior style for each living space to the customer terminal step; receiving, by the interior recommendation server, selected interior preference information from among interior preference analysis information from the customer terminal; generating, by the interior recommendation server, the customer's interior configuration information through the customer's interior configuration and interior taste information, and grouping and classifying the interior configuration information to correspond to expert recommendation models recommended by a group of experts; and the interior recommendation server compares and determines the expert recommendation model and the customer's interior configuration information, generates recommendation report information including the interior remodeling style of the expert recommendation model matching the customer's interior configuration information, and transmits it to the customer terminal including the steps of

In this feature, in generating the recommendation report information, the interior recommendation server recommends the interior remodeling style most similar to the interior configuration information for each customer's living space according to the set probability distribution of the expert recommendation model, respectively, and the recommended interior Recommendation report information is generated by including at least a remodeling style, company information for an interior construction company, a construction period, a construction cost, and a construction material.

In this feature, the processing method further includes the step of the interior recommendation server receiving feedback information on the recommendation report information from the customer terminal, generating the recommendation report information reflecting the feedback information, and transmitting it to the customer terminal do.

As described above, the non-face-to-face interior remodeling style recommendation system of the present invention enables a consumer to consult on an interior in a non-face-to-face manner anytime, anywhere by requesting and receiving an interior style recommendation request through a communication network.

In addition, the non-face-to-face interior remodeling style recommendation system of the present invention can increase consumer satisfaction with interior remodeling styles by recommending customized interior remodeling styles suitable for consumers' living spaces and tastes based on artificial intelligence and machine learning. there is.

In addition, the non-face-to-face interior remodeling style recommendation system of the present invention provides images and taste information on the interior situation of the consumer before the interior remodeling construction, so that the consumer can participate from the process of selecting the interior remodeling style, so that the consumer himself You can choose your preferred interior remodeling style, which can improve your satisfaction with interior design.

In addition, the non-face-to-face interior remodeling style recommendation system of the present invention selects and recommends an interior remodeling style that is most similar to the consumer's living space, interior composition, interior style, and interior taste through an expert recommendation model recommended by a group of experts, You can receive recommendations for customized interior remodeling styles suitable for consumers.

1 is a block diagram showing a network configuration of a non-face-to-face interior remodeling style recommendation system according to the present invention;
Figure 2 is a block diagram showing the configuration of the interior recommendation server shown in Figure 1, and
3 is a flowchart illustrating a processing procedure of a non-face-to-face interior remodeling style recommendation system according to the present invention.

Embodiments of the present invention may be modified in various forms, and the scope of the present invention should not be construed as being limited by the embodiments described below. This example is provided to more completely explain the present invention to those of ordinary skill in the art. Therefore, the shapes of the components in the drawings are exaggerated in order to emphasize a clearer description.

The non-face-to-face interior remodeling style recommendation system according to the present invention receives the interior image of the current living space and interior taste information on the interior style preferred by the customer from the customer through the communication network, and through object recognition and analysis from the interior image It diagnoses the interior condition of consumers and selects and recommends interior remodeling styles suitable for the customer's living space and taste through the expert recommendation model for various interior styles recommended in advance by a group of experts.

Therefore, the interior remodeling style recommendation system of the present invention enables customers to consult on interior anytime and anywhere in a non-face-to-face manner, analyzes the customer's living space and taste, and recommends a customized interior remodeling style preferred by the customer. satisfaction can be increased.

Hereinafter, an embodiment of the present invention will be described in detail with reference to the accompanying drawings 1 to 3 .

1 is a block diagram illustrating a network configuration of a non-face-to-face interior remodeling style recommendation system according to the present invention.

Referring to FIG. 1 , the interior remodeling style recommendation system 2 of the present invention receives a request for an interior style recommendation for remodeling an indoor interior in a non-face-to-face manner from a customer, To recommend a remodeling style, based on artificial intelligence and machine learning, it analyzes and diagnoses the customer's living space and interior taste, and recommends a customized interior remodeling style to the customer.

To this end, the interior remodeling style recommendation system 2 of the present invention provides an interior image of the customer's living space and an interior style that the customer prefers by using the customer terminal 200 as the interior recommendation server 100 through the communication network 4 . When interior preference information is selected and transmitted, the interior recommendation server 100 extracts the interior configuration for each living space of the current customer by recognizing the image analysis and objects according to the interior configuration of the customer's living space from the interior image, and the extracted By analyzing and learning the interior composition and customer's interior taste information, it selects and recommends an interior remodeling style suitable for the customer from among the expert recommendation models generated by the expert group. In this case, the interior recommendation server 100 may receive feedback on satisfaction, requirements, etc. for the interior remodeling style recommended to the customer, and update and recommend the customized interior remodeling style to the customer.

Therefore, the interior remodeling style recommendation system (2) of the present invention allows customers (ie, consumers) to consult about indoor interiors anytime and anywhere in a non-face-to-face manner, and provides the customer's living space and the interior style preferred by the customer to a group of experts. You can select a customized interior remodeling style through expert recommendation models.

To this end, the interior remodeling style recommendation system 2 of the present invention includes a communication network 4 , a plurality of customer terminals 200 , and an interior recommendation server 100 . In addition, the interior remodeling style recommendation system 2 of the present invention may further include a plurality of interior company terminals 300 .

Specifically, the communication network 4 includes, for example, a wired/wireless communication network, a mobile communication network, and the like, and is provided as a single communication network of each of these or a complex communication network in which each of them is combined. The communication network 4 is connected to each of the plurality of customer terminals 200, the interior recommendation server 100, and each of the plurality of interior company terminals 300 and the interior recommendation server 100 to perform data communication with each other.

The customer terminal 200 is a communicationable electronic device provided by a customer (consumer) who wants to receive an interior remodeling style recommendation in a non-face-to-face manner, for example, a smart phone, a tablet computer, a personal computer, etc., and is provided with a mobile application and By accessing the interior recommendation server 100 through a web service, it is requested to recommend an interior remodeling style suitable for the customer's living space and preferred by the customer. In this embodiment, the customer terminal 200 is provided with, for example, a smart phone, and interworks with the interior recommendation server 100 using a mobile application. In this case, the mobile application may be downloaded and installed, for example, from the interior recommendation server 100 or from a mobile application service market such as an Android play store or an Apple app store.

That is, the customer terminal 200 accesses the interior recommendation server 100 using a mobile application, and captures interior images of the customer's current living spaces through the camera 210 and transmits it to the interior recommendation server 100 . Alternatively, an interior image for each living space stored in the storage device 220 is selected and transmitted to the interior recommendation server 100, thereby requesting a recommendation for an interior remodeling style. At this time, the customer terminal 200 uses the camera 210 to shoot interior images for each of a plurality of living spaces, for example, a hallway, a living room, a kitchen, a room, a bathroom, and a space such as a window and a veranda. Alternatively, a plurality of interior images for each living space stored in the storage device 220 are selected and transmitted. Here, the interior image is provided as an image or video including the overall structure, color, interior style, etc. of each of the customer's current living spaces.

When the customer terminal 200 transmits an interior image to the interior recommendation server 100 and requests a recommendation of an interior remodeling style, the customer's preferred interior remodeling style for each living space, that is, interior taste information, is provided from the interior recommendation server 100. Receives interior taste analysis information for analysis. The customer terminal 200 selects the interior taste information preferred by the customer from among the transmitted interior taste analysis information, and transmits it to the interior recommendation server 100 . Here, the interior taste analysis information includes representative images representing various interior styles for each of the plurality of living spaces in order to determine the interior remodeling style preferred by the customer. Accordingly, the interior preference information selects a representative image preferred by the customer from among representative images for each living space for each living space and transmits the selected representative image to the interior recommendation server 100 .

In an embodiment of the present invention, the living space includes at least six spaces, for example, a hallway, a living room, a kitchen, a room, a bathroom, and a window and a balcony, and the interior taste analysis information includes six interior styles, for example , modern, natural, nordic, classic, modern classic and casual. In addition, the interior taste analysis information includes representative images showing various colors and structures according to each of the six interior styles for each of the six spaces. Therefore, the customer terminal 200 selects representative images such as interior style, color and structure for each living space according to the customer's taste from the interior taste analysis information transmitted from the interior recommendation server 100, and through this, the customer's own interior It is transmitted to the interior recommendation server 100 to be used as taste information.

In addition, the customer terminal 200 receives the recommendation report information to be described later from the interior recommendation server 100, checks the recommended interior remodeling style, and generates feedback information including satisfaction and requirements for the interior recommendation server ( 100) is sent. Accordingly, the customer terminal 200 may check the updated interior remodeling style by receiving the recommendation report information in which the feedback information is reflected.

The customer terminal 200 is given identification information (eg, ID) through the process of member registration and customer authentication of the interior recommendation server 100 . The identification information may include, for example, a customer name, contact information, e-mail address, password, address, and the like. Accordingly, the interior recommendation server 100 stores and manages the interior image and interior taste information by matching the identification information of the customer.

The interior company terminal 300 is an electronic device capable of communication provided by an interior construction company, for example, a smart phone, a personal computer, or the like. Of course, the interior company terminal 300 may be provided as a server of the interior construction company. The interior company terminal 300 is given identification information (eg, ID) through the process of member registration and customer authentication of the interior recommendation server 100, and through this, company information about the interior construction company, construction history information and the like are registered in the interior recommendation server 100 .

The interior company terminal 300 receives real-time or periodic requests for quotations for various interior remodeling styles from the interior recommendation server 100 through the communication network 4, and the construction period for the interior remodeling style, construction costs, materials, etc. Generates estimate information including the and transmits to the interior recommendation server (100).

And the interior recommendation server 100 is provided as at least one server to provide a service for recommending an interior remodeling style in a non-face-to-face manner, and a plurality of customer terminals 200 and a plurality of interior companies through the communication network 4 Each of the terminals 300 are connected. The interior recommendation server 100 is a web server equipped with a hardware and software platform, an image analysis algorithm, an object recognition algorithm, an artificial intelligence algorithm, a machine learning algorithm, etc. for recommending an interior remodeling style to each of the plurality of customer terminals 200 . , mobile application servers, and the like.

Therefore, when an interior style recommendation for interior remodeling is requested from the customer terminal 200, the interior recommendation server 100 analyzes and diagnoses the customer's living space and interior taste, and responds to the interior remodeling style generated in advance. Using the expert recommendation model of a group of experts, it selects and recommends customized interior remodeling styles to customers in a non-face-to-face manner.

To this end, the interior recommendation server 100 first builds an expert recommendation model for various interior styles recommended for each living space by a group of experts in advance, learns using a machine learning algorithm, and through this, determines the interior styles in similarity. According to the classification, it is grouped into a plurality of groups. That is, the interior recommendation server 100 builds an expert recommendation model for a plurality of interior styles by generating various interior styles according to living spaces, interior configurations, etc. from a group of experts in advance.

In an embodiment of the present invention, the expert recommendation model is generated by receiving a plurality of interior style models recommended by the expert group according to the interior style, color, space shape, and basic information of the customer for each living space. At this time, the living space includes an entrance, a kitchen, a living room, a bathroom, a room, and a window and a veranda, and the interior style includes a modern, a natural, a Nordic, a classic, a modern classic, and a casual. The color is composed by selectively combining the color of the point color, ceiling, wall, and floor for each interior style of the living space. And basic information includes the customer's age group, number of household members, housing type (apartment, villa, house, officetel, etc.), house size, number of rooms, budget (budget amount per pyeong), region, family composition (married, single, whether there are children, etc.), office space desired, and whether or not pets are present. In the expert recommendation model, a proposal probability is set for each interior style model.

The space type is, for example, a library type for the structure of a living room or room, a home cafe type, a general type, etc., a folding door for windows and verandas, an extension, a general window, a hanging type, a general type, a 3-dong middle door, a sliding middle door, a hinged door Lights, ceiling fans for living space, system air conditioners, stand air conditioners, wall-mounted air conditioners, etc., straight, L-shaped, U-shaped, island-shaped, etc. for living room structure, full upper cabinet for kitchen, shelf, mixed, double door refrigerator, refrigerator 2 Large, kimchi refrigerator, induction, etc., dry bathroom for bathroom, wet bath, bathtub type, shower partition type, shower booth type, etc., dressing room for room, drawing room, study, children's room, general room, etc. Environmental conditions such as general door, sliding door, pocket door, wainscoting, painting, wallpaper, tile, plain wood pattern, herringbone wood pattern, indirect lighting, recessed lighting, general lighting, point lighting, rail lighting, etc. are included.

Accordingly, the customer's interior configuration is extracted and analyzed according to the living space, interior style, color, and spatial form in response to the expert recommendation model, and is stored in the customer's interior configuration information.

The interior recommendation server 100 sets the number of groups to be classified using, for example, a K-Means Clustering Algorithm, clusters and classifies various interior styles, and repeats learning to recommend experts create a model

The interior recommendation server 100 stores and manages the interior image by matching it with the identification information of the customer when an image or video interior image of the customer's living space is transmitted from the customer terminal 200 . The interior recommendation server 100 analyzes the living space according to the interior configuration for each living space by using the image analysis algorithm for the interior image, and uses the data learned about the interior configurations in advance using the machine learning algorithm to configure the interior It recognizes objects according to the The interior recommendation server 100 learns in advance about the interior configuration for each living space, and thereafter it accompanies the task of increasing the accuracy through repeated learning.

The interior recommendation server 100 infers the meaning of the interior configuration by learning about the extracted interior configurations using a machine learning algorithm. The interior recommendation server 100 recognizes the size, color, number, brand, etc. of the object together with the extracted interior configuration, for example, a sofa, a refrigerator, a shoe cabinet, and the like, and through this, a customer according to the customer's current lifestyle, etc. Generate reasoning information to infer the interior state of That is, the interior recommendation server 100 infers the meaning of things by recognizing interior accessories present in each of the entrance hall, living room, kitchen, room, bathroom and window/veranda using an object recognition algorithm, and through this, the current customer's interior status can be determined. For example, in the case of a bed in the interior configuration, the customer's family composition can be checked through the type of bed such as single or double, and the need for a study can be determined by analyzing the presence or absence of a bookshelf, the size of the entrance hall and the kitchen Based on the number and size of refrigerators, the size of sofa, etc., it is possible to infer the current interior state of the customer.

The interior recommendation server 100 transmits interior preference analysis information for analyzing the interior preference information of the customer for each living space to the customer terminal 200 . The interior recommendation server 100 receives interior preference information selected by the customer among interior preference analysis information from the customer terminal 200 .

The interior recommendation server 100 selects and categorizes interior preference information and interior configuration information of customers generated and stored through interior configuration and inference information extracted from an interior image according to specific tastes. In this case, the interior recommendation server 100 groups and classifies the interior styles of the selected customers using, for example, a genetic algorithm and a K-Nearest Neighbor Algorithm. Such clustering of customer's interior configuration information can not only utilize genetic algorithms, K-nearest neighbor algorithms, etc., but also can be variously changed and applied using algorithms that process the same function.

The interior recommendation server 100 analyzes and diagnoses interior configuration information for the customer's own living space and interior taste information transmitted from the customer terminal 200, and repeats learning to respond to the customer's interior configuration information to the expert recommendation model to cluster them into a plurality of groups.

The interior recommendation server 100 recommends an interior remodeling style suitable for the customer by comparing and discriminating the interior style of the customer whose interior configuration information is selected with the expert recommendation model of the expert group. To this end, the interior recommendation server 100 generates recommendation report information based on the expert recommendation model of the expert group and the interior style of the selected customer through a recommendation algorithm. At this time, the recommended report information includes company information on the most suitable interior remodeling style for each of the recommended living spaces or an interior contractor with experience in the interior style, construction period, and construction cost for each interior remodeling style. and estimate information such as construction materials.

In an embodiment of the present invention, the expert recommendation model sets a probability distribution that suggests which interior style is empirically suitable for each interior style for four environmental conditions. The clustering of customers' interior styles recognizes and analyzes interior images, extracts interior composition information for each living space, and learns them based on machine learning and clusters them into a plurality of similar groups.

Here, the four environmental conditions include interior style, color combination, spatial form, and basic customer information, interior style includes modern, natural, northern Europe, modern classic and casual, and color combination includes multiple The colors are combined, and the space type includes options or configuration types added for each space, and basic information includes the number of household members, house size, number of rooms, area, housing type, companion animal, etc.

Therefore, as for the customer's interior style, the probabilistic model recommended by experts is set for each interior style of the expert recommendation model of the expert group, and the clustering process is equally processed to select the interior remodeling style with the highest probability of being proposed according to a plurality of clustering groups. It selects and recommends an interior style suitable for the customer's interior configuration information.

Also, when the feedback information on the recommendation report information is transmitted from the customer terminal 200 , the interior recommendation server 100 updates the recommendation report information by analyzing it, and transmits it to the customer terminal 200 .

As described above, the interior recommendation server 100 of the present invention receives the customer's interior information, that is, interior image, interior taste information, inference information, and feedback information from the customer terminal 200, based on artificial intelligence and machine learning. Analyzes and diagnoses the customer's interior configuration information, categorizes the interior configuration information, compares and identifies it with the expert recommendation model of the expert group, and recommends a customized interior remodeling style to the customer.

Next, the configuration and function of the interior recommendation server according to an embodiment of the present invention will be described in detail. That is, FIG. 2 is a block diagram illustrating the configuration of the interior recommendation server shown in FIG. 1 .

Referring to FIG. 2 , the interior recommendation server 100 according to an embodiment of the present invention provides data preprocessing technology, data clustering algorithm, genetic algorithm, interior style clustering algorithm, etc. to recommend interior remodeling style to customers in a non-face-to-face manner. based on artificial intelligence and machine learning.

To this end, the interior recommendation server 100 includes a server unit 110 and a database unit 150 . The server unit 110 is connected to each of a plurality of customer terminals 200 and a plurality of interior company terminals 300 through a communication network 4, and a web server and an app server, artificial intelligence and machine that process data communication to be possible. It is provided as a web server and a mobile application server for receiving a request for recommending an interior remodeling style from the customer terminal 200 based on the learning and recommending a customized interior remodeling style to the customer. The database unit 150 stores and manages various information according to the overall processing process of the interior recommendation server 100 processed by the server unit 110 . In this embodiment, the database unit 150 is included in the interior recommendation server 100, but may be provided as an independent database server.

In this embodiment, the server unit 110 includes a control unit 112 , a communication unit 114 , a client management unit 116 , an interior request processing unit 118 , an interior image analysis unit 120 , an interior style classification unit 122 , and an interior and a style recommendation unit 124 . Here, the server unit 110 is provided in the form of an integrated server, but a plurality of them may be provided in the form of an independent server. For example, each of the interior request processing unit 118 , the interior image analysis unit 120 , the interior style classification unit 122 , and the interior style recommendation unit 124 may be provided as an independent server.

Specifically, the control unit 112 controls the customer terminal 200 and the interior company terminal 300 to interwork with each of the operations to process the overall operation of the interior recommendation server 100 . That is, the control unit 110 includes a communication unit 114 , a client management unit 116 , an interior request processing unit 118 , an interior image analysis unit 120 , an interior style classification unit 122 , an interior style recommendation unit 124 , and a database. Each function of the units 150 is controlled to be organically processed with each other.

The communication unit 114 is provided with, for example, a wired/wireless communication module, a mobile communication module, etc., and is connected to each of the customer terminal 200 and the interior company terminal 300 through the communication network 4 so that mutual data communication is performed. handle Of course, the communication unit 114 may include, for example, a network device such as a gateway, a router, a switch, a hub, and a repeater for data communication.

The client management unit 116 includes customer identification information for member registration of the customer terminal 200, and residence type information such as apartment, villa, officetel, etc. Registers and manages company information including customer information to be used, identification information for an interior construction company of the interior company terminal 300, a construction company name, address, contact information, construction history, and the like.

When the interior image for each living space is transmitted from the customer terminal 200, the interior request processing unit 118 receives the customer's interior recommendation request, and stores the interior image in the database unit 150 in response to the customer's identification information, manage At this time, when interior recommendation requests are received from a plurality of different customer terminals 200 , the interior request processing unit 118 registers them in a scheduling queue in the order of reception and sequentially processes them.

The interior request processing unit 118 transmits interior preference analysis information for analyzing the interior style preference of the customer by analyzing the interior configuration and inference information extracted from the interior image analysis unit 120 to the customer terminal 200 . At this time, the interior request processing unit 118 transmits the interior taste analysis information to the customer terminal 200 as various representative images representing interior styles for each living space according to interior style, color, configuration and shape of space, and the like. Accordingly, the customer terminal 200 selects interior preference information for each living space preferred by the customer from among the transmitted interior preference analysis information, and transmits the selected interior preference information to the interior recommendation server 100 . The interior request processing unit 118 receives the interior taste information selected from the customer terminal 200 and stores and manages the information in the database unit 150 in response to the identification information of the customer.

When the interior image and interior taste information are transmitted from the customer terminal 200 and the interior recommendation request is received, the interior request processing unit 118 transmits the occurrence of an event according to the reception of the interior recommendation request to the interior image analysis unit 120 .

In addition, the interior request processing unit 118 transmits the recommendation report information including the interior remodeling style recommended by the interior style recommendation unit 124 to the customer terminal 200 , and feedback on the recommendation report information from the customer terminal 200 . The information is received and stored in the database unit 150 and transmitted to the interior style recommendation unit 124 .

The interior image analysis unit 120 receives the event occurrence from the interior request reception unit 118, analyzes images and recognizes objects from each of the interior images and interior taste information for each living space, and extracts the interior configuration provided for each living space of the customer. . The interior image analysis unit 120 infers the meaning of the extracted interior configuration to generate inference information on the interior state of the customer. At this time, the interior image analysis unit 120 extracts and infers the interior configuration through the learning data for various interior configurations registered in advance, and applies the extracted interior configuration and inference information and interior taste information to the identification information of the customer. It is matched and stored in the interior configuration information of the database unit 150 .

The interior style classification unit 122 clusters and classifies the interior configuration information of the customer's interior image, interior taste information, and inference information using an expert recommendation model recommended by a group of experts in advance. To this end, the interior style classification unit 122 processes the data pre-processing process for clustering the customer's interior configuration information through the expert recommendation model in advance to structure the data, distribute the data through the data partitioning process, and use this as the learning data. save as The interior style classification unit 122 groups and classifies the interior configuration information of the customer by matching the expert recommendation model through the learning data.

The interior style recommendation unit 124 compares and analyzes the interior configuration information of the customer clustered by the interior style classification unit 124 with the expert recommendation model, and according to the probability distribution of the expert recommendation model, the interior style closest to the customer's preferred interior style An interior style having a probability of preference is extracted. The interior style recommendation unit 124 recommends the extracted interior style as an interior remodeling style. A plurality of extracted interior styles are provided for each living space according to a plurality of interior styles, colors, spatial structures or shapes, and among them, an interior style having a high probability distribution is selected and recommended as an interior remodeling style. At this time, the interior style recommendation unit 124 is most suitable with the recommended interior remodeling style, or company information on an interior contractor who has experience in the same interior remodeling style construction, construction period, construction cost, construction materials, etc. By including the included estimate information, the recommendation report information is generated, stored, and transmitted to the customer terminal 200 . Here, the recommendation report information further includes a plurality of images for confirming the interior remodeling style for each of the living spaces.

Accordingly, the customer terminal 200 checks the recommendation report information, generates feedback information including customer satisfaction and requirements, and transmits it to the interior recommendation server 100 , and the interior style recommendation unit 124 provides feedback information The recommendation report information is updated by reflecting the , and it is retransmitted to the customer terminal 200 .

And the database unit 150 includes the client information 152 including the customer information of the customer terminal 200 and the company information of the interior company terminal 300, and the interior state for each living space transmitted from the customer terminal 200 The interior image 154 that represents and includes the interior configuration, the interior configuration information 156 that includes the customer's preferred interior style through the interior configuration, inference information, and interior taste information extracted from the interior image, and the customer by living space Interior taste analysis information 158 including various representative images for analyzing the interior taste of 160), the expert recommendation model 162 that includes various interior styles recommended by the expert group in advance and is grouped and classified according to similar characteristics, and the satisfaction and requirements for the interior remodeling style recommended by the customer terminal 200 It stores and manages at least the feedback information 164 including feedback information 164 and the recommended interior remodeling style, company information on interior construction companies, construction period, construction cost, and recommended report information 166 including construction materials.

And the processing method of the non-face-to-face interior remodeling style recommendation system according to the present invention will be described in detail.

3 is a flowchart illustrating a processing procedure of a non-face-to-face interior remodeling style recommendation system according to the present invention. The procedure of this embodiment is processed by the configuration of the remodeling style recommendation system shown in FIGS. 1 and 2 .

Referring to FIG. 3 , the interior remodeling style recommendation system 2 of the present invention transmits an interior image for each living space from one customer terminal 200 to the interior recommendation server 100 in step S400 to request an interior recommendation. At this time, the customer terminal 200 uses the camera 210 to capture and transmit each interior image for each living space, or select and transmit an interior image for each living space stored in the storage device 220 .

In step S410, when the interior image is transmitted from the customer terminal 200, the interior recommendation server 100 stores it, and gives an image ID corresponding to the customer's identification information so that the customer and the interior image are matched. In step S420, the interior recommendation server 100 analyzes the interior image, recognizes each interior configuration for each living space, and extracts the interior configuration for each living space. At this time, the interior recommendation server 100 further generates inference information by inferring the meaning of the extracted interior configuration.

In step S430 , the interior recommendation server 100 transmits interior taste analysis information to the customer terminal 200 . In this case, the interior taste analysis information is transmitted in the form of a representative image representing a plurality of interior styles, colors, and spatial structures for each living space. In step S440, the customer terminal 200 selects a representative image of the interior style preferred by the customer from among the interior taste analysis information transmitted from the interior recommendation server 100 for each living space, and through this, the customer's interior preference information is provided to the interior recommendation server Send to (100).

In step S450, the interior recommendation server 100 analyzes the interior configuration, inference information, and interior preference information to generate the customer's interior configuration information, and the generated interior configuration information of the customer is similar to the expert recommendation model recommended by the expert group. It is grouped and classified by interior style according to the In step S460, the interior recommendation server 100 compares and determines the expert recommendation model of the expert group and the interior configuration information of the customer. In step S470, the interior recommendation server 100 recommends an interior remodeling style most similar to the customer's interior configuration information according to the probability distribution of the expert recommendation model. At this time, the interior recommendation server 100 generates recommendation report information including the recommended interior remodeling style and company information on the construction company, construction period, construction cost, and construction materials, and transmits it to the customer terminal 200 .

Next, in step S480, the customer terminal 200 checks the recommended recommendation report information, and selectively generates feedback information including satisfaction and requirements for the recommended interior remodeling style as needed, and the interior recommendation server 100 send to Accordingly, the interior recommendation server 100 reflects the feedback information, updates the recommendation report information, and retransmits it to the customer terminal 200 .

In the above, the configuration and operation of the non-face-to-face interior remodeling style recommendation system and the method according to the present invention have been illustrated according to the detailed description and drawings, but this is merely described by way of example, and does not depart from the technical spirit of the present invention. Various changes and modifications are possible within the scope that does not exist.

2: Interior Remodeling Style Recommendation System
4: communication network
100: interior recommendation server
200: user terminal
210: camera
220: storage device
300: interior company terminal

Claims (7)

In the interior remodeling style recommendation system:
Including a; interior recommendation server connected to each of a plurality of customer terminals through a communication network;
The interior recommendation server,
When the interior image of the customer's living space is transmitted from the customer terminal, interior taste analysis information for analyzing the interior style preferred by the customer is transmitted to the customer terminal, and interior taste selected from the interior taste analysis information from the customer terminal an interior request processing unit receiving the information and receiving a request for interior style recommendation;
The interior image transmitted to the interior request processing unit is recognized and analyzed to extract an interior configuration for a plurality of living spaces, generate inference information inferring the meaning of the extracted interior configuration, and extract the interior configuration and inference information and an interior image analysis unit receiving interior taste information from the interior request processing unit and generating the interior configuration information of the customer;
The expert recommendation model for a plurality of interior styles generated after being recommended by a group of experts in advance is grouped and classified into a plurality of groups according to the degree of similarity, and the interior configuration information of the customer generated by the interior image analysis unit corresponds to the expert recommendation model. an interior style classification unit for clustering and classifying interior styles according to customer's taste; and
By comparing the interior style according to the customer's taste classified by the interior style classification unit with the expert recommendation model, a customized interior remodeling style is selected and recommended to the customer, and the recommended interior remodeling style and company information on the interior construction company; An interior style recommendation unit that generates recommendation report information including estimate information including construction period, construction cost, and construction materials and transmits it to the customer terminal;
The interior request processing unit;
When the customer terminal transmits the recommendation report information generated from the interior style recommendation unit to the customer terminal, and feedback information including at least the customer's satisfaction and requirements for the recommended interior remodeling style of the recommendation report information is transmitted from the customer terminal, feedback and transmit the information to the interior style recommendation unit to generate the recommended report information, and
The interior style recommendation unit;
An interior remodeling style recommendation system, characterized in that, according to the probability distribution set for each interior style of the expert recommendation model, each of the interior remodeling styles having the probability of preference most similar to the interior configuration information for each living space of the customer is extracted and recommended.
delete delete The method according to claim 1,
The interior recommendation server,
The customer's living space includes at least the entrance, living room, kitchen, room, bathroom, and windows and veranda, and includes at least Modern, Natural, Nordic, Classic, Modern Classic and Casual for each of the living spaces in a plurality of interior styles. And, an interior remodeling style recommendation system, characterized in that it recommends an interior remodeling style by classifying and comparing the expert recommendation model and the customer's interior configuration information according to the color combination of each interior style and the structure of the living space.
A method of processing an interior remodeling style recommendation system, comprising:
receiving, by an interior recommendation server, an interior image for a customer's living space from a customer terminal through a communication network, and requesting a recommendation for an interior remodeling style;
The interior recommendation server analyzes the interior image to recognize the interior configuration, extracts the recognized interior configuration for each living space, and transmits interior taste analysis information for the customer's preferred interior style for each living space to the customer terminal step;
receiving, by the interior recommendation server, interior preference information selected from among interior preference analysis information from the customer terminal;
generating, by the interior recommendation server, interior configuration information of the customer through the interior configuration and interior taste information of the customer, and classifying the interior configuration information by clustering and classifying the interior configuration information to correspond to expert recommendation models recommended by a group of experts; and
The interior recommendation server compares and determines the expert recommendation model and the customer's interior configuration information, generates recommendation report information including the interior remodeling style of the expert recommendation model matching the customer's interior configuration information, and transmits it to the customer terminal step; including;
Generating the recommendation report information includes:
The interior recommendation server extracts and recommends each of the interior remodeling styles having the probability of preference most similar to the interior configuration information for each living space of the customer according to the probability distribution set for each interior style of the expert recommendation model, and the recommended interior remodeling style; generating recommendation report information by including estimate information including at least company information, construction period, construction cost, and construction materials for the interior construction company; And
The processing method is
The interior recommendation server further receives feedback information including at least customer satisfaction and requirements for the recommended interior remodeling style of the recommendation report information from the customer terminal, generates recommendation report information reflecting the feedback information, and sends it to the customer terminal Processing method of the interior remodeling style recommendation system, characterized in that it further comprises the step of transmitting. delete delete

Images