KR20230011115A - System for providing image based personal perfume recommendation service - Google Patents

Abstract

The present invention provides a system for providing an image-based personal perfume recommendation service, capable of providing a platform which can help reasonable decision-making and purchase of a customer. The system for providing an image-based personal perfume recommendation service includes a user terminal and a recommendation service-providing server. The user terminal inputs user information including age and sex and selects an image associating a fragrance which a user wants from at least one image. Also, the user terminal receives a recommendation about a perfume based on the user information and the image. The recommendation service-providing server includes: a database unit mapping at least one fragrance and the at least one image and storing the image mapped with the fragrance; a perfume storing unit mapping at least one fragrance and the at least one perfume and storing the perfume mapped with the fragrance; and a recommendation unit extracting the perfume and the flagrance stored by being mapped in advance with the at least one image selected by the user terminal and the user information which is input by the user terminal and recommending the extracted flagrance and the perfume for the user terminal.

Description

Image-based personal perfume recommendation service system {SYSTEM FOR PROVIDING IMAGE BASED PERSONAL PERFUME RECOMMENDATION SERVICE}

The present invention relates to an image-based personal perfume recommendation service providing system, and provides a platform capable of finding and recommending a desired scent simply by selecting an image corresponding to the desired scent based on cross-association between the image and the scent.

Modern consumers not only use perfume as a means to pursue individuality and create an image, but also distinguish and use perfume according to the time, place, situation, mood, and characteristics of the event, just like clothes. Just as perfumers get inspiration from spices and specific situations or atmospheres in the creative process of creating fragrances, perfume consumers conversely associate the colors of raw materials or unique situations and atmospheres through the fragrance of perfumes. In addition, as colors affect not only human emotions but also smell, hearing, touch, and balance, certain scents evoke specific colors. Fragrance goes beyond simply pleasing the mood and expands its application area to the purpose of treatment through sedation and stimulation, so the image felt through fragrance is also important in the consumption of perfume.

At this time, after collecting situational information, a method of recommending a perfume or custom-made perfume by extracting a target scent related to a propensity has been researched and developed. Announced on September 30) and Korean Registered Patent No. 10-1913729 (Announced on November 1, 2018), refer to situational information collected from users' social media to determine a target fragrance index associated with a propensity among fragrance indexes , A configuration for extracting an object related to a target scent index from the user's social media, deriving scent-related preference information by referring to attributes, and recommending a scent corresponding to the scent-related preference information by referring to a database defining associations. and a configuration for injecting fragrance of perfume raw materials into the space of the chamber, detecting a customer's bio-signal in response to the injected fragrance, and controlling the injection and detection to derive the optimal fragrance for the customer.

However, in the former case, although context information is extracted from the user's social media, not all of the user's context information is uploaded to social media, and there is a possibility that the scent corresponding to the context information is not the user's favorite scent. cannot be recommended Even in the latter case, if the fragrance is continuously sprayed, the degree of stimulation must increase according to Weber's law to recognize the difference in fragrance, and the sense of smell is paralyzed before the optimal fragrance is derived, so that no fragrance can be smelled. Therefore, it is required to research and develop a platform that can recommend scents reflecting individual memories and inspirations by allowing users to select scents that come to mind through associations based on images.

An embodiment of the present invention uses the human characteristic of identifying the scent of an object through the visual information recognized by the eyes based on the cross-modal association formed by the olfactory and visual stimuli, and the perceived visual information. evokes an image of a specific object to guess the scent based on the experience of the associated image, builds a database by mapping and storing the fragrance by image, and selects the desired image from the user. By extracting the corresponding scent and then recommending a suitable perfume, it eliminates the problem of inaccurate word-based recommendation as the general public and experts use different terms, and a professional perfumer interprets the sensory analysis data from the image. It is possible to provide a method for providing an image-based personal perfume recommendation service that enables automatic recommendation by constructing a data set to be input by big data and artificial intelligence through the dataset operation for extracting fragrance. However, the technical problem to be achieved by the present embodiment is not limited to the technical problem as described above, and other technical problems may exist.

As a technical means for achieving the above-described technical problem, an embodiment of the present invention inputs user information including age and gender, selects an image reminiscent of a scent desired by the user from among at least one image, and A user terminal for recommending perfume based on information and images, a database unit for mapping and storing at least one fragrance and at least one image, a perfume storage unit for mapping and storing at least one fragrance and at least one perfume, and a user terminal and a recommendation service providing server including a recommendation unit for extracting user information input in and at least one image selected in the user terminal and pre-mapped and stored scents and perfumes, and recommending them to the user terminal.

According to any one of the above-mentioned problem solving means of the present invention, based on the cross-modality association relationship formed by the olfactory and visual stimuli, using the human characteristic of identifying the scent of an object through visual information recognized by the eyes And, the perceived visual information evokes the image of a specific object, so that the scent is guessed based on the experience of the associated image, a database is built by mapping and storing the fragrance by the image, and the user's desired When an image is selected, it extracts the scent corresponding to the image and then recommends a suitable perfume, eliminating the problem of inaccurate word-based recommendation as the general public and experts use different terms, and sensory analysis by a professional perfumer By analyzing data and extracting fragrance from images, it is possible to build big data and even a data set to be input by artificial intelligence, enabling automatic recommendation. The number of perfume brands and products, including Niche perfume, Considering the growing trend of the online market for products and products, it is possible to provide a platform that can help consumers make rational decisions and purchases.

1 is a diagram for explaining a system for providing an image-based personal perfume recommendation service according to an embodiment of the present invention.
FIG. 2 is a block diagram illustrating a recommendation service providing server included in the system of FIG. 1 .
3 and 4 are diagrams for explaining an embodiment in which an image-based personal perfume recommendation service according to an embodiment of the present invention is implemented.
5 is an operational flowchart illustrating a method of providing an image-based personal perfume recommendation service according to an embodiment of the present invention.

Hereinafter, embodiments of the present invention will be described in detail so that those skilled in the art can easily practice the present invention with reference to the accompanying drawings. However, the present invention may be embodied in many different forms and is not limited to the embodiments described herein. And in order to clearly explain the present invention in the drawings, parts irrelevant to the description are omitted, and similar reference numerals are attached to similar parts throughout the specification.

Throughout the specification, when a part is said to be "connected" to another part, this includes not only the case where it is "directly connected" but also the case where it is "electrically connected" with another element interposed therebetween. . In addition, when a part "includes" a certain component, this means that it may further include other components, not excluding other components, unless otherwise stated, and one or more other characteristics. However, it should be understood that it does not preclude the possibility of existence or addition of numbers, steps, operations, components, parts, or combinations thereof.

As used throughout the specification, the terms "about", "substantially", etc., are used at or approximating that value when manufacturing and material tolerances inherent in the stated meaning are given, and do not convey an understanding of the present invention. Accurate or absolute figures are used to help prevent exploitation by unscrupulous infringers of the disclosed disclosure. The term "step of (doing)" or "step of" as used throughout the specification of the present invention does not mean "step for".

In this specification, a "unit" includes a unit realized by hardware, a unit realized by software, and a unit realized using both. Further, one unit may be realized using two or more hardware, and two or more units may be realized by one hardware. On the other hand, '~ unit' is not limited to software or hardware, and '~ unit' may be configured to be in an addressable storage medium or configured to reproduce one or more processors. Thus, as an example, '~unit' refers to components such as software components, object-oriented software components, class components, and task components, processes, functions, properties, and procedures. , subroutines, segments of program code, drivers, firmware, microcode, circuitry, data, databases, data structures, tables, arrays and variables. Functions provided within components and '~units' may be combined into smaller numbers of components and '~units' or further separated into additional components and '~units'. In addition, components and '~units' may be implemented to play one or more CPUs in a device or a secure multimedia card.

In this specification, some of the operations or functions described as being performed by a terminal, device, or device may be performed instead by a server connected to the terminal, device, or device. Likewise, some of the operations or functions described as being performed by the server may also be performed by a terminal, apparatus, or device connected to the server.

In this specification, some of the operations or functions described as mapping or matching with the terminal mean mapping or matching the terminal's unique number or personal identification information, which is the terminal's identifying data. can be interpreted as

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

1 is a diagram for explaining a system for providing an image-based personal perfume recommendation service according to an embodiment of the present invention. Referring to FIG. 1 , the image-based personal perfume recommendation service providing system 1 includes at least one user terminal 100, a recommendation service providing server 300, at least one perfumer terminal 400, and at least one corporate server. (500). However, since the image-based personal perfume recommendation service providing system 1 of FIG. 1 is only an embodiment of the present invention, the present invention is not limitedly interpreted through FIG. 1 .

At this time, each component of FIG. 1 is generally connected through a network (Network, 200). For example, as shown in FIG. 1 , at least one user terminal 100 may be connected to a recommendation service providing server 300 through a network 200 . In addition, the recommendation service providing server 300 may be connected to at least one user terminal 100, at least one perfumer terminal 400, and at least one corporate server 500 through the network 200. In addition, at least one perfumer terminal 400 may be connected to the recommendation service providing server 300 through the network 200 . In addition, at least one enterprise server 500 may be connected to at least one user terminal 100, recommendation service providing server 300, and at least one perfumer terminal 400 through the network 200.

Here, the network refers to a connection structure capable of exchanging information between nodes such as a plurality of terminals and servers, and examples of such networks include a local area network (LAN) and a wide area network (WAN: Wide Area Network), the Internet (WWW: World Wide Web), wired and wireless data communications networks, telephone networks, and wired and wireless television communications networks. Examples of wireless data communication networks include 3G, 4G, 5G, 3rd Generation Partnership Project (3GPP), 5th Generation Partnership Project (5GPP), Long Term Evolution (LTE), World Interoperability for Microwave Access (WIMAX), Wi-Fi , Internet (Internet), LAN (Local Area Network), Wireless LAN (Wireless Local Area Network), WAN (Wide Area Network), PAN (Personal Area Network), RF (Radio Frequency), Bluetooth (Bluetooth) network, NFC ( A Near-Field Communication (Near-Field Communication) network, a satellite broadcasting network, an analog broadcasting network, a Digital Multimedia Broadcasting (DMB) network, etc. are included, but not limited thereto.

In the following, the term at least one is defined as a term including singular and plural, and even if at least one term does not exist, each component may exist in singular or plural, and may mean singular or plural. It will be self-evident. In addition, the singular or plural number of each component may be changed according to embodiments.

At least one user terminal 100 may be a terminal of a user who requests perfume recommendation using a web page, app page, program, or application related to an image-based personal perfume recommendation service. To this end, the user terminal 100 may be a terminal of a user who selects one of a plurality of images associated with a fragrance and receives a recommendation for a perfume by inputting an image adjective describing the fragrance.

Here, at least one user terminal 100 may be implemented as a computer capable of accessing a remote server or terminal through a network. Here, the computer may include, for example, a laptop, a desktop, a laptop, and the like equipped with a navigation system and a web browser. In this case, at least one user terminal 100 may be implemented as a terminal capable of accessing a remote server or terminal through a network. At least one user terminal 100 is, for example, a wireless communication device that ensures portability and mobility, navigation, PCS (Personal Communication System), GSM (Global System for Mobile communications), PDC (Personal Digital Cellular), PHS (Personal Handyphone System), PDA (Personal Digital Assistant), IMT (International Mobile Telecommunication)-2000, CDMA (Code Division Multiple Access)-2000, W-CDMA (W-Code Division Multiple Access), Wibro (Wireless Broadband Internet ) may include all types of handheld-based wireless communication devices such as terminals, smartphones, smart pads, tablet PCs, and the like.

The recommendation service providing server 300 may be a server that provides an image-based personal perfume recommendation service web page, app page, program or application. Also, the recommendation service providing server 300 may be a server that builds a database by mapping and storing at least one image, scent, and perfume. When constructing the database, the recommendation service providing server 300 maps and stores each image, scent, and perfume from at least one perfumer terminal 400, and tags adjectives expressing each image or scent to form a corpus (Corpus). ) or a terminal that creates a word cloud. When the perfumer's dataset construction is completed, the recommendation service providing server 300 selects an image corresponding to a desired scent from the user terminal 100 or extracts the corresponding scent and perfume when a word is input to the user terminal 100. ) may be a recommended server. In addition, when the user terminal 100 purchases and pays for the recommended perfume, the recommendation service providing server 300 recollects the testimonial reviews, durability, preference, or association with the image after a predetermined period of time, so that the user's data is stored. If there is an error, it may be a server that corrects it with feedback. In this way, a series of processes of image selection, word input, recommendation, and feedback for each user are stored as a history log, and the recommendation service providing server 300 builds big data when sufficient datasets are accumulated and at least one artificial intelligence It may be a server that performs modeling by learning and testing an algorithm.

Here, the recommendation service providing server 300 may be implemented as a computer capable of accessing a remote server or terminal through a network. Here, the computer may include, for example, a laptop, a desktop, a laptop, and the like equipped with a navigation system and a web browser.

At least one perfumer terminal 400 may be a terminal of a professional perfumer using an image-based personal perfume recommendation service-related web page, app page, program, or application. At this time, in order to be registered as the perfumer terminal 400, related qualifications, history, career, education, etc. may be transmitted to the recommendation service providing server 300, and the recommendation service providing server 300 verifies the related history, and then the perfumer terminal ( 400) can be databased as a perfumer manpower pool. The perfumer terminal 400 may be a terminal that recommends scents and perfumes suitable for the user in consideration of an image selected by the user terminal 100 and an input word. In addition, the perfumer terminal 400 may be a terminal that performs an operation of constructing a dataset by mapping each image, scent, and perfume when building a database in the recommendation service providing server 300 .

Here, at least one perfumer terminal 400 may be implemented as a computer capable of accessing a remote server or terminal through a network. Here, the computer may include, for example, a laptop, a desktop, a laptop, and the like equipped with a navigation system and a web browser. In this case, at least one perfumer terminal 400 may be implemented as a terminal capable of accessing a remote server or terminal through a network. At least one perfumer terminal 400 is, for example, a wireless communication device that ensures portability and mobility, navigation, PCS (Personal Communication System), GSM (Global System for Mobile communications), PDC (Personal Digital Cellular), PHS (Personal Handyphone System), PDA (Personal Digital Assistant), IMT (International Mobile Telecommunication)-2000, CDMA (Code Division Multiple Access)-2000, W-CDMA (W-Code Division Multiple Access), Wibro (Wireless Broadband Internet ) may include all types of handheld-based wireless communication devices such as terminals, smartphones, smart pads, tablet PCs, and the like.

At least one company server 500 may be a terminal of a company that imports, manufactures, and sells perfume using a web page, app page, program, or application related to an image-based personal perfume recommendation service. At this time, the company server 500 transmits perfume data to the recommendation service providing server 300, and a server that receives purchase statistical data or feedback data of the buyer (user) as a marketing report via the recommendation service providing server 300. can be

Here, at least one enterprise server 500 may be implemented as a computer capable of accessing a remote server or terminal through a network. Here, the computer may include, for example, a laptop, a desktop, a laptop, and the like equipped with a navigation system and a web browser. In this case, at least one enterprise server 500 may be implemented as a terminal capable of accessing a remote server or terminal through a network. At least one enterprise server 500 is, for example, a wireless communication device that ensures portability and mobility, and includes navigation, PCS (Personal Communication System), GSM (Global System for Mobile communications), PDC (Personal Digital Cellular), PHS (Personal Handyphone System), PDA (Personal Digital Assistant), IMT (International Mobile Telecommunication)-2000, CDMA (Code Division Multiple Access)-2000, W-CDMA (W-Code Division Multiple Access), Wibro (Wireless Broadband Internet ) may include all types of handheld-based wireless communication devices such as terminals, smartphones, smart pads, tablet PCs, and the like.

2 is a block diagram for explaining a recommendation service providing server included in the system of FIG. 1, and FIG. 3 is a diagram for explaining an embodiment in which an image-based personal perfume recommendation service according to an embodiment of the present invention is implemented. it is a drawing

Referring to FIG. 2, the recommendation service providing server 300 includes a database conversion unit 310, a perfume storage unit 320, a recommendation unit 330, a big data conversion unit 340, an artificial intelligence unit 350, and a related unit. 360, an automatic recognition unit 370, and a feedback unit 380 may be included.

The recommendation service providing server 300 according to an embodiment of the present invention or another server (not shown) operating in conjunction with at least one user terminal 100, at least one perfumer terminal 400, and at least one corporate server When an image-based personal perfume recommendation service application, program, app page, web page, etc. is transmitted to 500, at least one user terminal 100, at least one perfumer terminal 400, and at least one corporate server 500 ) may install or open an image-based personal perfume recommendation service application, program, app page, web page, or the like. In addition, the service program may be driven in at least one user terminal 100, at least one perfumer terminal 400, and at least one corporate server 500 using a script executed in a web browser. Here, the web browser is a program that allows users to use the web (WWW: World Wide Web) service, and means a program that receives and displays hypertext described in HTML (Hyper Text Mark-up Language). For example, Netscape , Explorer, Chrome, and the like. In addition, an application means an application on a terminal, and includes, for example, an app running on a mobile terminal (smart phone).

Referring to FIG. 2 , the database builder 310 may map and store at least one scent and at least one image. The perfumer terminal 400 may provide a DataSet for constructing at least one artificial intelligence algorithm by mapping at least one image, at least one scent, and at least one perfume. At this time, initially, the image, scent, and perfume are input from the perfumer terminal 400, but, for example, the perfumer may refer to the fragrance with expressions such as warm, while the warm side in the image perception of the general public When perception rarely appears, differences in text or images may occur due to differences between experts and ordinary people. In this case, in order to accumulate basic data, it is desirable to accumulate objective data by evaluating multiple images of the same perfume by a number of samples who have received a certain amount of training on fragrance.

At this time, association (Association) is to relate two objects. For example, recalling objects, events, or experiences related to it through the impression of color can be referred to as association of color. In addition, for scent or perfume, it includes emotional association, color association of scent, and scent association of color. For example, if three colors associated with quince scent, jasmine scent, lavender scent, and floral scent are selected out of 120 colors, the floral scent is the B, V, and P tones of the R series, and the jasmine scent is the YR series. The B, Vp tones and lavender scent are similar to those associated with the G series' L, V, S tones, and the quince scent to the RP series' P, V, Vp tones. In addition, it is possible to associate the intensity of fragrance from color. For example, in the case of evaluating the degree of association of fresh, floral, oriental, and woody fragrances with respect to color stimulation, the tone and fragrance of the remaining three fragrances except for oriental series are evaluated. The cross-modal association relationship between the colors is systematically changed, so that, for example, a brighter warm color is associated with a strong floral scent, whereas a woody scent is associated with a low degree, and a fresh scent can be strongly associated with a vivid or bright cold color. Accordingly, the intensity and type of fragrance may be mapped according to the color of each image or the dynamics of the subject.

In this case, at least one image may include at least one adjective image scale. Here, since the terms or words used by each individual are different, corpus construction may be performed using words used by the user for the same direction as a word cloud or word corpus. Recently, the government's Sejong Corpus project was resumed after 10 years, and the National Institute of the Korean Language released 1.8 billion words of Korean big data to teach AI. Terms used in various generations and media, such as newspaper corpus, written corpus, spoken language corpus, and messenger corpus, have been redefined. At this time, texts actually input by the user can also be gathered together to build a new corpus and redefine the dictionary. In this case, the user's image adjectives can be collected as much as possible, and the discrepancy rate between adjectives and images can be reduced when adjectives are collected together with images.

In addition to selecting an image, the user may also express the scent he or she desires with words including adjectives. Accordingly, in order to construct the [perfume-perfume-color-image-adjective], the following image scale is used to make the [color-image-adjective] part. At this time, since the part of [fragrance-perfume-image] or [fragrance-perfume-color] is a part that a professional perfumer will connect when building a dataset, the [color-image-adjective] part can be constructed below. Of course, it is not excluded to build all of [perfume-perfume-color-image-adjective] with the following concept.

When it comes to conveying emotional feelings, humans sometimes find it difficult to explain objectively. Emotional language is not a concrete form or shape, but an abstract concept that has been felt in our minds. It is possible to secure objectivity by analyzing symbols or images into distinct semantic spaces. In particular, monochromatic image space, color image space, and adjective image space used in color image space are soft, hard, dynamic, and static. It is divided into four basic axes and used as various image data. As a result of research by the IRI Color Research Institute and the Ministry of Knowledge Economy, a monochromatic image space has already been researched and developed. ) means that it is located in space through

Color images can be expressed in language, and conversely, language expressing emotions can be directed again in color images. In other words, emotional words are conveyed with one color image. Sensibility plays an important role in design because it is an emotional state of individuals that allows them to feel various moods and emotions. These emotional images can be expressed in language or color. Color is an emotional image that reflects a person's mind. No. C-2001-001388 of the IRI Color Research Institute is a color matching image space, which organizes the feeling of color created when colors are combined using adjective language, and then organizes associations collected in the monochromatic image space into an adjective image space. put The adjective image scale expressed in perfume consists of emotional image types and emotional groups. The images people see and feel have a lot in common, but it is difficult to explain scientifically. Table 1 summarizes the emotional groups according to the emotional images announced by the I.R.I Color Research Institute. It is a category created by binding the difference in meaning of adjectives that play the role of a bridge that connects the relationship between color and image.

clear Light, refreshing, shallow, neat, soft, transparent, delicate gentle Pure, supple, smooth, rustic, calm, stable, weak soft Neat, tidy, orderly, profound, simple, static natural Friendly, natural, friendly, sweet, emotional, cozy, idyllic, simple elegant Sensual, convenient, stylish, oriental, classic lofty quiet, calm, aged, gloomy, old, muddy nice Refined, Formal, Intelligent, Solid, Traditional, Conservative fancy Relaxed, decorative, Korean, cool, cool, generous modern Practical progressive, urban, masculine, rugged, functional, heavy, dark cute Cute, fresh, fresh, enjoyable, interesting, lovely, sweet light New, free, dynamic, rhythmic, bright, diverse, outstanding, young dynamic Outstanding, distinctive, innovative, strong, tough, wild

At this time, initially, the text and scent in Table 1 will be mapped to the four quadrants consisting of soft, hard, dynamic, and static with the color and motion of the subject for each image with human intervention, but after that, each image is extracted It may be possible to analyze the color and the movement of the subject through image recognition and map the scent and scent. Analysis of whether a subject is dynamic or not is already possible, and colors can already be extracted as HUE and RGB of pixels. The process of collecting, analyzing images with image analysis algorithms, and mapping scents can be automated. In this case, since a person does not have to search for and map images one by one, the cost and time for constructing a database can be reduced while the quantity and quality of the image database can be increased.

In addition, in one embodiment of the present invention, fragrance is considered as a priority when recommending a product, but marketing and design factors such as product concept or package design may also be considered. In other words, it is possible to introduce not only the fragrance but also all the elements constituting the product, and a user-customized product can be recommended by considering the package design, target customer group, and perfume design. For example, when a user inputs images and text, marketing elements, product concepts, target elements, product designs, etc. can be databased together so that a recommendation can be selected in consideration of product concept, story, and design. .

The perfume storage unit 320 may map and store at least one scent and at least one perfume. At this time, since at least one scent and at least one image are mapped, if at least one perfume is mapped to a scent, it can be directly connected to the image, but perfume is not a mixture of single scents, but each scent is Since it is close to a compound that emits a unique scent while being harmoniously mixed, when mapping an image for a single fragrance, a discrepancy may occur when a perfume is actually recommended, and the perfume and image can be matched directly. This may use cross-modal associations between image and nostalgia, which will be described later in the association unit 360.

The recommendation unit 330 may recommend the user information input from the user terminal 100 to the user terminal 100 by extracting and pre-mapped to at least one image selected from the user terminal 100 and storing the scent and perfume. . The user terminal 100 may input user information including age and gender, select an image reminiscent of a fragrance desired by the user from among at least one image, and receive a perfume recommendation based on the user information and the image. The company server 500 may provide perfume data of at least one perfume to the recommendation service providing server 300 and receive a marketing report from the recommendation service providing server 300 .

The big data generator 340 stores raw data including at least one image, at least one scent, and at least one perfume in parallel and distributedly, and includes unstructured data included in the stored raw data, Structured data and semi-structured data can be refined, pre-processing including classification into meta data can be performed, and analysis including data mining can be performed on the pre-processed data.

The artificial intelligence unit 350 may learn and test at least one artificial intelligence algorithm by receiving a dataset in which at least one image, at least one scent, and at least one perfume are mapped.

When mapping at least one scent to at least one image, the association unit 360 uses a cross-modal association between the color of the at least one image and the object and the at least one scent to at least one scent. After analyzing one image, at least one scent can be set to be automatically mapped. Numerous studies have revealed consistent cross-modality associations between color and scent, for example, light herbal scents do not evoke passionate reds or strong blacks. At this time, for association of association, for example, the Apriori algorithm may be used. This is a representative technique of an association rule (association) search algorithm, and allows to search for frequent association rules between data based on the frequency of occurrence. When defining an itemset having more than the minimum support as a frequent itemset, the Apriori algorithm finds only the frequent itemsets with more than the minimum support and calculates the association rule instead of calculating the support for all itemsets. The Apriori algorithm proceeds in two steps. In the first step, the combination of all frequent items with support greater than or equal to the minimum support is calculated, and in the second step, the maximum size of frequent items is obtained and the reliability between subsets of the corresponding items is calculated. .

The automatic recognition unit 370 may automatically extract and classify objects and colors included in at least one image through image recognition deep learning. At this time, the dynamism of the object must also be judged. For example, if the color is a mild color scheme but the object has the dynamism of running, the perfume mapped to this image should not be simply floral. In order to extract the subject, the background must be removed and the foreground separated, and the dynamics or movement of the subject in the foreground must also be grasped. If the subject is not specified as one, the motion is determined by the background itself. When the subject is specified as one, it is possible to extract what posture it is taking, what place it is located in, and the like.

In this case, content-based image retrieval (CBIR) may be used, and color, shape, texture, or motion information may be used as attributes. Color information accounts for the largest portion of image information and is a widely used feature. Texture information is a visual pattern with similar properties that cannot be produced only by the presence of color or contrast. The use of shape information is limited rather than color or texture due to the difficulty of expressing it, but it is based on the fact that human first sight depends on the outline of a quality image. For example, when performing a search using the statistical characteristics of color histograms, classify the three RGB histograms of color characteristics into several bins and classify features based on statistical values of deviation, skewness, and kurtosis algorithms can be implemented. Alternatively, CCV (Color Coherence Vector) may be used to solve the lack of spatial information in the color histogram method. Alternatively, a method of extracting the global characteristics of the shape based on shape template matching and matching the shape of the outline through horizontal analysis of this data may be used. HSV color and uniform local binary patterns can also be used. This converts the image to the HSV color model to extract Hue, converts the image to grayscale to obtain ULBP (Uniform Local Binary Patterns), and A histogram created by classifying can also be used as a feature.

Alternatively, an adaptive weight CBIR algorithm may be used, which is largely divided into three steps. First, it is a preprocessing and size filtering process to determine the object of the image. The second is the process of extracting features such as labeling, area, corner points, color, and density. .

<Object discrimination and feature extraction>

In order to classify images according to characteristics, a process of removing unnecessary information is required. As a preprocessing process to have optimal labeling for each image, the Otsu method can be used. In addition, a process of assigning numbers to labeling using pixel connectivity and rearranging them in order of area size can be added. There may be labeling with only small information of the image, or there may be unwanted noise embedded in the image information. In this case, size filtering can be performed so that components of a specific size or less can be ignored, so that objects can be identified. Object discrimination in an image can be determined by distinguishing whether the background is a main image or an image capable of discriminating an object. Images that meet the condition of Equation 1 are classified, and images in which the difference in normalized area is larger than the threshold of Th are found. Th has a value of a point that is the boundary of large groups.

Using the normalized area feature, images are divided into background images and images with objects, and the images with objects identified go through the process of extracting the features of the objects. Density determines how densely the regions of an object are clustered in an image. In order to quickly extract the feature vector of an image, it is possible to determine the degree of clumping of objects rather than accurately determining the shape of each object. Density can be expressed as the amount of area per unit perimeter. The corner point is one of the important shape information along with the contours of the image, and the Harris corner operator can be used. In order to reduce processing time, location information with a large amount of data is not used, and the amount of data can be reduced by using only the number of corner points. The color pixel value can be used as a feature vector by creating an indexed image using the RGB color model and calculating the average color of each labeling in order to determine the basic color.

<Adaptive Weighting>

It is necessary to analyze the features of the query image objects using the feature vector calculated from the labeled images and adaptively assign weights according to the values. In the object discrimination process, if one labeling occupies more than half of the image information, it is classified as a background image, so the normalized area has values between 0 and 0.5, and a single object image has a value of 1 exceptionally. Corner points are important information in determining the complexity of morphological information. The number of corner points can be used to determine the degree of complexity of the object shape with the number of corner points. That is, the larger the number of corner points, the more complex the shape of the object. Conversely, the smaller the number of corner points, the closer the shape of the object is to that of a circle with fewer angles. Since the absolute total amount of the number of corner points is not determined, the value of the number of corner points can be analyzed by extracting feature vectors of each image.

Similar to the number of corner points, the feature of density is important information in determining the complexity of morphological information. In comparing images, it is better to distinguish the density of each object for quick comparison of feature vectors. Color uses the indexing number distribution of the index image used when extracting color information, unlike feature vectors used to determine area or morphological information. Equation 2 below may be used to determine the color distribution.

IDmain means the main number of an object in the index data matrix, and Atotal means the total amount of object area. In this way, the process of assigning weights according to the characteristics of the query image and adaptively applying them is summarized in Equation 3 below.

Here, Awtotal means the adaptive weight according to the query image, and means the characteristics of the normalized area, density, number of corner points, and color, respectively. n has a maximum value of 4, and w means the weight of each feature vector, which is a value determined through object analysis. Color, corner point, area, and density were used as attribute elements, but extension of additional attributes is possible. At this time, CBIR may be driven in such a way as to return a perfume having a feature located at the shortest distance by comparing a distance between a feature extracted from a subject in an image given by a query and a feature of perfume. A distance based on Euclidean distance or cosine similarity may be used.

After the user terminal 100 selects, purchases, and pays for the perfume, the feedback unit 380 receives a matching rate between the image selected in the user terminal 100 and the fragrance of the perfume as feedback, and the user terminal 100 ) If the association relationship between the user's image and perfume is different from the pre-stored association relationship between the image and perfume, the association relationship of the user terminal 100 may be reconstructed through correction. Compared to the visual image in perfume, the olfactory image has a lot of preference differences depending on the individual's environment and memory, and the distribution that can be defined in the language of emotional adjectives is not large. Therefore, even if the perfumer establishes the [perfume-image] relationship as [A-B], if the user feels [A-C], [A-C] is correct from the user's point of view, so the perfumer's [A-B] cannot be forced. There is no Accordingly, if the user gives feedback as [A-C], it is corrected, that is, a pointer to re-establish the [B-C] relationship in order not to change the database while leaving the user's taste as it is, such as [A-B (C)] ( Pointer) to redefine the label relationship in the DB is called calibration.

In addition, in one embodiment of the present invention, if you are new to perfume or you are not an expert such as a perfumer, you often have no idea what kind of fragrance it is, no matter how much images and adjectives are used to explain it. At this time, if you test once, you can immediately recognize what kind of scent it is by smell, so the location, address, perfume you have, type, number and remaining capacity of at least one affiliated store terminal (not shown) are updated in real time, Recommended perfumes can tell you that you can test them at a branch store. Since it can be troublesome to go outside to test only perfume, if the user terminal 100 has put it in a shopping cart or has a recommended perfume, the location of the user terminal 100 is collected, and the time to go to work or go to school Check whether it is time to leave work or dismissal time rather than busy time, and the location of the user terminal 100 is similar to the location of the affiliated store terminal 400 or exists on the path of the movement pattern, and the corresponding affiliated store terminal 400 enters the shopping cart Tester of stored perfume If perfume exists, a message can be delivered so that a test can be performed on the way by passing it as a notification to the user terminal 100 .

Alternatively, there is a case where an unknown person passes by and wants to ask what kind of perfume it is, but is embarrassed or difficult to talk to. When perfume is sprayed on the user terminal 100, an event is collected, and the current location of the user terminal 100 is collected by GPS, and then the perfume sprayed by the user of the user terminal 100 is curious about another user terminal (not shown). A configuration for providing a guide message is also possible. At this time, since there may be indoors where the exact location of the user terminal 100 is not displayed or an error may occur on the GPS, the message propagation function of the user terminal 100 may be used. In this case, the process may be configured in a scenario in which a relay terminal forms a P2P group with adjacent P2P terminals and voluntarily spreads a perfume advertisement. Relay terminals must form a P2P group to immediately transmit a perfume advertisement, and must perform a procedure for selecting a new relay terminal to spread the advertisement.

Therefore, the relay terminal must be a GO, and among the group formation methods (Standard, Autonomous, Persistent), the P2P group formation independently and the Autonomous group formation method that becomes the GO can be used. When the relay terminal forms a P2P group, it can exchange data frames necessary for advertisement product information and relay selection with P2P clients. Relay terminals selected through data exchange disconnect the existing P2P group that has received the advertisement as a P2P client and form a group continuously through an iterative process of forming a new P2P group as a P2P GO for advertisement transmission.

A method for spreading an advertisement divides the communication range of a terminal starting an advertisement by an arbitrary number of sectors according to an angle. Since one relay terminal is selected for each sector, the number of simultaneous advertisements spread is determined according to the divided %. Then, target areas included in the angular range of the divided sectors are classified, and an advertisement diffusion route is determined in order of target areas having the closest distance to the previously selected target area. In addition, a terminal closest to the target area of the advertisement spreading route determined for each sector is selected as a relay terminal, and if the target area is located within a transmission radius of the relay terminal, a terminal closest to the next target area is selected as a relay terminal. By repeating this process, the advertisement can be spread to the final target area. Although the above-described method is a method for spreading advertisements between terminals adjacent to a physical distance, it can be equally applied to movement and diffusion between human networks.

Hereinafter, an operation process according to the configuration of the above-described recommendation service providing server of FIG. 2 will be described in detail with reference to FIG. 3 as an example. However, it will be apparent that the embodiment is only any one of various embodiments of the present invention, and is not limited thereto.

Referring to FIG. 3, (a) the recommendation service providing server 300 collects at least one image, and (b) associates and associations from at least one perfumer terminal 400 so that fragrance, perfume, and image are mapped. Collect a dataset based on At this time, (c) when entering a name, gender, age, etc. in the user terminal 100 and selecting an image corresponding to a desired scent, the previously mapped and stored scent and perfume are extracted and recommended to the user terminal 100. and receive feedback on recommendations. And, (d) based on the history and feedback of the recommendation, the data set can be rebuilt and at least one artificial intelligence algorithm can be retrained. At this time, each individual may have established different concepts due to various environmental factors such as preference or background, but it is impossible to train with an artificial intelligence algorithm using a dataset with different rules. Accordingly, the artificial intelligence algorithm is trained using the perfumer's dataset, but the method of correcting only the above-mentioned pointers can be used for the parts that are different or different due to individual differences, tastes, growth backgrounds, occupations, etc.

4a shows a tentative name (Your Madeleine) of a service according to an embodiment of the present invention. In Marsel Proust 'In Search of Lost Time', the narrator drank tea with Madeleine melted and was full of colorful sensations at Combray. It is a name that was conceived from reminiscing about childhood. In the service according to an embodiment of the present invention, the language of perfume learned by experts such as perfumers and marketers may be far from the scent that Korean consumers actually want and feel, and actual customers are influenced by the inspiration and advertisements received by foreign perfumers. Purchases are greatly influenced by the image contained in it, and individual consumers make or layer perfumes with limited raw materials to find their own fragrance, but the actual perfume market is so high-quality that more than 2,000 perfume products were released in 2016 alone. It starts with the fact that there are many products of

In addition, in a market situation where artificial intelligence technology is increasing the number of cases in which scents that will sell well by region, gender, and age group are increasing, the present invention can help find well-made perfume products that reflect the inspiration and taste of unique customers. project was derived. While recommending products, it is possible to secure sensory data on various prototypes and individual consumers for perfumes, and furthermore, provide image-based perfume recommendation services through the introduction of machine learning and artificial intelligence technologies. You can find a perfume that contains the moments and emotions you want to keep for everyone who has difficulties with perfume or who knows you well, and if you express the scent that will become each individual Madeleine in pictures and words like the narrator of Marcel Proust, you can find a professional perfumer's analysis. You can get recommendations for good products that are available in the market.

When an image or text is entered as shown in FIGS. 4B to 4D, the perfumer sees this and recommends an appropriate perfume. As described above, when a dataset is initially secured, it is automatically performed by artificial intelligence afterwards, and the perfumer confirms the result. It can be designed to do just that. 4e shows the current status of similar services. The first service is based on words, so research and association relationships are not clearly established, such as between fragrance and image (color), so accuracy is low. In the second case, consistency is insufficient. In the case of the second, low-quality products are inevitable due to lack of expertise. Accordingly, as shown in FIG. 4e, an embodiment of the present invention is configured to recommend a perfume that best suits one's taste among products sold on the market. 4f to 4k are example screens implemented according to an embodiment of the present invention. [Photo+Text] is the query, and the result of analyzing and mapping it to perfume (Output) is the perfume list at the bottom. Of course, the screen design according to an embodiment of the present invention may not be the same and may be changed, modified, and edited according to the embodiment.

Matters not described for the method of providing the image-based personal perfume recommendation service of FIGS. 2 to 4 are the same as those described for the method of providing the image-based personal perfume recommendation service through FIG. Since it can be inferred, the following description will be omitted.

5 is a diagram illustrating a process of transmitting and receiving data between components included in the image-based personal perfume recommendation service providing system of FIG. 1 according to an embodiment of the present invention. Hereinafter, an example of a process of transmitting and receiving data between each component will be described through FIG. 5, but the present application is not limited to such an embodiment, and according to various embodiments described above, It is obvious to those skilled in the art that a process of transmitting and receiving data may be changed.

Referring to FIG. 5 , the recommendation service providing server maps and stores at least one fragrance and at least one image (S5100), maps and stores at least one fragrance and at least one perfume (S5200), Pre-mapped to the user information input in and at least one image selected in the user terminal and stored are extracted and recommended to the user terminal (S5300).

The order between the above-described steps (S5100 to S5300) is only an example, and is not limited thereto. That is, the order of the above-described steps (S5100 to S5300) may be mutually changed, and some of the steps may be simultaneously executed or deleted.

Matters not described for the image-based personal perfume recommendation service provision method of FIG. 5 are the same as those described for the image-based personal perfume recommendation service provision method through FIGS. 1 to 4 above, or can be easily obtained from the description Since it can be inferred, the following description will be omitted.

The method of providing an image-based personal perfume recommendation service according to an embodiment described with reference to FIG. 5 may be implemented in the form of a recording medium including instructions executable by a computer such as an application or program module executed by a computer. there is. Computer readable media can be any available media that can be accessed by a computer and includes both volatile and nonvolatile media, removable and non-removable media. Also, computer readable media may include all computer storage media. Computer storage media includes both volatile and nonvolatile, removable and non-removable media implemented in any method or technology for storage of information such as computer readable instructions, data structures, program modules or other data.

The above-described method for providing an image-based personal perfume recommendation service according to an embodiment of the present invention is executed by an application basically installed in a terminal (this may include a program included in a platform or operating system, etc. It may be executed by an application (ie, a program) directly installed in the master terminal by a user through an application providing server such as an application store server, a web server related to an application or a corresponding service. In this sense, the above-described method for providing an image-based personal perfume recommendation service according to an embodiment of the present invention is implemented as an application (ie, a program) that is basically installed in a terminal or directly installed by a user, and is readable by a computer such as a terminal. can be recorded on a recordable medium.

The above description of the present invention is for illustrative purposes, and those skilled in the art can understand that it can be easily modified into other specific forms without changing the technical spirit or essential features of the present invention. will be. Therefore, the embodiments described above should be understood as illustrative in all respects and not limiting. For example, each component described as a single type may be implemented in a distributed manner, and similarly, components described as distributed may be implemented in a combined form.

The scope of the present invention is indicated by the following claims rather than the detailed description above, and all changes or modifications derived from the meaning and scope of the claims and equivalent concepts should be construed as being included in the scope of the present invention. do.

Claims (11)

a user terminal that inputs user information including age and gender, selects an image reminiscent of a fragrance desired by the user from among at least one image, and receives a perfume recommendation based on the user information and image; and
A database unit for mapping and storing at least one fragrance and at least one image, a perfume storage unit for mapping and storing the at least one fragrance and at least one perfume, user information input from the user terminal and information selected from the user terminal a recommendation service providing server including a recommender for extracting scents and perfumes previously mapped to and stored in at least one image and recommending the same to the user terminal;
An image-based personal perfume recommendation service providing system comprising a.
According to claim 1,
The image-based personal perfume recommendation service providing system,
A perfumer terminal providing a data set for constructing at least one artificial intelligence algorithm by mapping the at least one image, at least one scent, and at least one perfume;
Image-based personal perfume recommendation service providing system further comprising a.
According to claim 1,
The recommendation service providing server,
Raw data including at least one image, at least one fragrance, and at least one perfume is stored in parallel and distributed, and unstructured data, structured data and semi-structured data included in the stored raw data A big data conversion unit that refines semi-structured data, performs preprocessing including classification into meta data, and analyzes the preprocessed data including data mining;
Image-based personal perfume recommendation service providing system further comprising a.
According to claim 1,
The recommendation service providing server,
an artificial intelligence unit that learns and tests at least one artificial intelligence algorithm using a data set in which at least one image, at least one scent, and at least one perfume are mapped;
Image-based personal perfume recommendation service providing system further comprising a.
According to claim 1,
The recommendation service providing server,
When mapping the at least one scent to the at least one image, the color of the at least one image and a cross-modal association between an object and the at least one scent are used to map the at least one scent to the at least one image. an association unit that analyzes and sets the at least one scent to be automatically mapped;
Image-based personal perfume recommendation service providing system further comprising a.
According to claim 5,
The recommendation service providing server,
an automatic recognition unit for automatically extracting and classifying objects and colors included in the at least one image through image recognition deep learning;
Image-based personal perfume recommendation service providing system further comprising a.
According to claim 1,
The image-based personal perfume recommendation service providing system, characterized in that the at least one fragrance is a raw material of a fragrance including a single fragrance or complex fragrances.
According to claim 1,
The image-based personal perfume recommendation service providing system,
an enterprise server providing perfume data of the at least one perfume to the recommendation service providing server and receiving a marketing report from the recommendation service providing server;
Image-based personal perfume recommendation service providing system further comprising a.
According to claim 1,
The image-based personal perfume recommendation service providing system, characterized in that the at least one image includes at least one adjective image scale.
According to claim 1,
The image-based personal perfume recommendation service providing system, characterized in that the at least one image is replaced with at least one adjective image scale.
According to claim 1,
The recommendation service providing server,
After selecting, purchasing, and paying for perfume in the user terminal, a matching rate between the image selected in the user terminal and the fragrance of the perfume is received as feedback, and the association relationship between the user's image and the perfume in the user terminal is pre-stored. a feedback unit for reconstructing an association relationship of the user terminal through correction when different from an association relationship between an image and a fragrance;
Image-based personal perfume recommendation service providing system further comprising a.

Images