WO2016186457A1 - Information providing system and method using image recognition technology - Google Patents

Information providing system and method using image recognition technology Download PDF

Info

Publication number
WO2016186457A1
WO2016186457A1 PCT/KR2016/005283 KR2016005283W WO2016186457A1 WO 2016186457 A1 WO2016186457 A1 WO 2016186457A1 KR 2016005283 W KR2016005283 W KR 2016005283W WO 2016186457 A1 WO2016186457 A1 WO 2016186457A1
Authority
WO
WIPO (PCT)
Prior art keywords
image
registered
value
information
server
Prior art date
Application number
PCT/KR2016/005283
Other languages
French (fr)
Korean (ko)
Inventor
김동주
Original Assignee
주식회사 넥스비즈코리아
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Priority claimed from KR1020160001532A external-priority patent/KR20160135641A/en
Application filed by 주식회사 넥스비즈코리아 filed Critical 주식회사 넥스비즈코리아
Publication of WO2016186457A1 publication Critical patent/WO2016186457A1/en

Links

Images

Classifications

    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06KGRAPHICAL DATA READING; PRESENTATION OF DATA; RECORD CARRIERS; HANDLING RECORD CARRIERS
    • G06K19/00Record carriers for use with machines and with at least a part designed to carry digital markings
    • G06K19/06Record carriers for use with machines and with at least a part designed to carry digital markings characterised by the kind of the digital marking, e.g. shape, nature, code
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00Information retrieval; Database structures therefor; File system structures therefor
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00Pattern recognition
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06QINFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q30/00Commerce
    • G06Q30/02Marketing; Price estimation or determination; Fundraising
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06QINFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q50/00Systems or methods specially adapted for specific business sectors, e.g. utilities or tourism
    • G06Q50/10Services
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06VIMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00Arrangements for image or video recognition or understanding
    • G06V10/10Image acquisition
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06VIMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00Arrangements for image or video recognition or understanding
    • G06V10/70Arrangements for image or video recognition or understanding using pattern recognition or machine learning
    • G06V10/74Image or video pattern matching; Proximity measures in feature spaces
    • G06V10/75Organisation of the matching processes, e.g. simultaneous or sequential comparisons of image or video features; Coarse-fine approaches, e.g. multi-scale approaches; using context analysis; Selection of dictionaries

Definitions

  • the present invention relates to a method for providing related information registered by multiple information providers in the same image using image recognition technology in a mobile environment.
  • Image recognition technology is applied in various ways according to the field of use, one of which is to find the same image among multiple images.
  • a mobile app Application
  • a mobile device equipped with a camera such as a smartphone
  • a user photographs an advertisement image or a label of a product using a smartphone
  • the photographed image is the image.
  • It is a technology that delivers the information registered in the image by finding the same image or the most similar image that is delivered to the recognition server.
  • the present invention has been made to solve the above-described problems, the present invention is a mobile user acquires an image from his mobile terminal, the seller of the image through the APP and the server using the image recognition technology registered by the seller
  • An object of the present invention is to provide a means for connecting a user terminal to a purchase site (such as an internet shopping mall) of a product.
  • an aspect of the present invention is connected to a mobile terminal via a network, and in the information providing method using an image recognition technology comprising an image recognition server, a web server, and a DB server Transmitting the image pattern value extracted from the acquired image or the acquired image to the image recognition server; Extracting an image pattern value when the image recognition server receives an image from the mobile terminal, and retrieving an image pattern value similar to a registered image pattern value when the image pattern value is directly transmitted; The image recognition server transmitting an image code value, which is a search result value, through the mobile terminal or directly to a web server; The web server finds a promotional code value and a shopping code value registered in the received image code value, and transmits the promotional material information registered in the promotional code value or the shopping mall URL information registered in the shopping mall code value to a mobile terminal. It provides an information providing method using an image recognition technology comprising a.
  • the image and the corresponding image code value is registered in advance in the DB server, the registration method is to register a new image that has not been previously registered or to find a desired image from the pre-registered image to the desired image You can register additional information.
  • a method of finding a desired image from the previously registered image is a method of searching through a keyword search by registering a keyword corresponding to a registered image when registering an image in advance, or by searching the same or most through an image search.
  • the same image may be regarded as a method of searching for and selecting a similar image, or registering an image used for the search, or having an extremely small difference between the pattern value of an image to be registered and an image to be registered and the same as the registered pattern value. .
  • Another aspect of the present invention is an information providing system using an image recognition technology connected through a mobile terminal and a network, the image recognition to search for a similar image with respect to the received image, the image code value corresponding to the retrieved similar image server; And a DB server connected to the image recognition server, the DB server storing a plurality of promotional and / or shopping related information in an image code value corresponding to the image in order to provide a plurality of information in one image code value.
  • the mobile terminal wants to provide information through the acquired image, it is configured to connect to select from a plurality of promotional and / or shopping-related information for the image.
  • the number of images registered on the server is greatly reduced.
  • the server searches in a state where all image pattern values are resident in RAM. Instead of registering the same number of images, only one image is registered and multiple link addresses or related information are searched. Linking information not only reduces the data capacity of the server, but also speeds up the search.
  • FIG. 1 illustrates an information providing system using an image recognition technology according to an exemplary embodiment of the present invention.
  • FIG. 2 is a reference diagram for explaining a process of identifying whether or not a similar image is identical by using an image recognition technology according to an embodiment of the present invention.
  • Figure 3 relates to a method for automatically finding the same image using image recognition technology in accordance with an embodiment of the present invention.
  • FIG. 4 is an example of an image connection DB and an information registration DB table for each promotion and shopping information provider according to an exemplary embodiment of the present invention.
  • FIG. 5 is a flowchart illustrating a method of finding an image to be registered from a pre-registered image through keyword search according to an embodiment of the present invention.
  • FIG. 6 is a flowchart illustrating a method of determining whether an image to be registered is a registered image through a search according to an embodiment of the present invention, and selecting and registering an image or a newly registered image according to the result. to be.
  • FIG. 7 is a flowchart for providing related information registered by a plurality of information providers using image recognition technology according to an exemplary embodiment of the present invention.
  • FIG 1 illustrates an information providing system using an image recognition technology according to an embodiment of the present invention.
  • the information providing system using the image recognition technology of FIG. 1 is connected to the mobile terminal 100 through a network, and includes a web server 200, an image recognition server 300, and the like.
  • the image recognition server 300 is connected to the image server connection DB 310, information registration DB 320 for each information provider.
  • the information provider DB information for each provider 320 may be configured to include a promotion DB (320A), shopping DB (320B).
  • Various DBs are classified for convenience of description and may be collectively named as DB server 305.
  • the image code connected to the image recognition server 300 may be configured as a single code, but the information provided by a plurality of information providers is provided in one image. In order to do this, it is preferable that there is a separate information registration DB for each information provider separately from the image server connection DB 310 of the image recognition server 300.
  • the image server connection DB 310 stores a code value connected to at least one information provider-specific information registration DB 320 in one image information.
  • Each image may be registered by the mobile terminal 100 or registered by the system operator.
  • a method of registering an image may be classified into a promotional material registration and a shopping mall registration, and may be registered in the promotion DB 320A and the shopping DB 320B.
  • public relations refers to various types of promotional materials such as introduction articles, photo albums, and videos
  • a purchase site for sale refers to an Internet shopping mall.
  • the mobile terminal 100 installs and holds an application APP for providing information using the image recognition technology of the present invention. Looking at the function of the mobile terminal 100, and transmits the image or image pattern value obtained from the mobile terminal to the image recognition server (300).
  • This image may be an image taken through an APP or an internal camera function installed in a mobile terminal, may be a recognized image, or may be an image secured by a method such as down, scan, or the like separately. The preferred method is camera shot or recognized image.
  • the mobile terminal 100 receives the search result value (image code value) from the image recognition server 300 and transmits it to the web server 200.
  • the image recognition server 300 extracts an image pattern value when the image is received from the mobile terminal 100, and retrieves the most similar image pattern value among the registered image pattern values when the image pattern value is directly transmitted.
  • the search result value (image code value) is transmitted to the mobile terminal 100.
  • the search result value may be transmitted to the web server 200 without transmitting the search result value to the mobile terminal 100.
  • the web server 200 finds the promotional code value and the shopping code value registered in the image code value received from the information registration DB 320 for each information provider, and the promotional material (introduction, photo album, video) registered in the promotional code value. ) Or the shopping mall URL information registered in the shopping mall code value to the mobile terminal 100, and the mobile terminal 1000 receiving the received information is displayed on the mobile screen. Icon), one or more shopping mall information appears, and a user selects a shopping mall among them.
  • the image recognition server 300 plays a role of delivering the image code value by finding the same (most similar) image, and the web server 200 stores the promotional code value and the shopping mall in the DB server where the image code value is found. It is described as playing an intermediary role to find the code value and again to find the information registered in the promotional code value and the shopping mall code value.
  • the image recognition server 300 and the web server 200 may be configured as a single server.
  • the image registration method may include i) a method of registering a new image not previously registered, and ii) a method of finding a desired image from the registered image and additionally registering information desired by the user.
  • ii-1 As a method of searching for a desired image from a registered image, ii-1) when registering an image in advance, the keyword corresponding to the registered image is also registered, so that the keyword can be found through keyword search.
  • Ii-2) There is a method of searching for and selecting the same or most similar image through image search, or registering an image used in the search.
  • ii-3 there may be a method in which the image having the same or identical error value between the registered pattern value and the pattern value of the registered image may be regarded as the same image.
  • a completely different image may be regarded as the same image depending on the image. Therefore, when using this method, it may be effective to configure a standard.
  • the pattern value of the image to be newly registered is equal to or larger than a predetermined value with the pattern value of the previously registered image, it is regarded as the same image. In other words, if more pattern values are extracted, the higher the value of Query Image Score compared to the comparison image, the less likely it is to consider another image as the same. If it is high, it is regarded as the same image. In the above method, it is necessary to additionally find the same image below a certain level.
  • a reference value may be extracted from a pattern value of a pre-registered image to calculate a comparison value (Query Image Score) and a comparison rate of the comparison image.
  • the method of calculating the reference value from the pattern value of the pre-registered image may be based on the number of pattern values extracted from the image engine server 300, but the same image as the pre-registered image is searched and compared with the pre-registered image.
  • the comparison value calculated by reference may be used as a reference value. Similar to the first method, the above method can regard different images as the same image if the comparison value is below a certain level.
  • a method of borrowing human judgment may be effective to determine whether or not the same image is determined below a predetermined comparison value.
  • the reason for this is, for example, when an image of a specific product logo is registered and the same product logo is searched, and when an image including a product name is included in the product logo, both comparison values may be identical. Because it can.
  • FIG. 2 is a reference diagram for explaining a process of identifying whether or not a similar image is identical by using an image recognition technology according to an embodiment of the present invention.
  • Figure (a) and Figure (b) in the product except for the letters G1, G1FE are all the same, but the two products are different model name.
  • Figure (c) is the same product as Figure (a), with the letters GS Caltex engraved at the bottom.
  • the pattern value is much larger than the difference between Fig. (A) and Fig. (B), and the error of Fig. (A) and Fig. (C) is much larger. ) Is more likely to be perceived as an image similar to Figure (b). For this reason, the method of using the human cognitive ability may be effective to solve this problem. In this case, borrowing the human cognitive ability does not apply the human cognitive ability as part of the system, but means that a step of selecting one of the images displayed on the terminal screen is added.
  • an information provider when it registers an image, it searches for whether an image similar to that image is registered in the existing image recognition server 300, and considers the same image if the comparison image (Query Image Score) of the comparison image is above a certain level. It may be desirable to enable automatic registration, and below a certain level, to expose at least one or more of the most similar images on the screen, and to allow the information provider to select which image to use between the pre-registered image and the image that he or she would like to initially register. have.
  • the user selects and registers an image to be registered as the same image even though the same image is extracted.
  • two identical images are registered in the image recognition server 100, but a solution in this case is also required.
  • the solution is that the comparison values of the comparison images are searched for a plurality of comparison images in ascending order, so that the comparison value (Query Image Score) of the highest comparison image or the reference value (Result Image Score) of the registered image and the plurality of retrieved images are compared.
  • the comparison rate is displayed to a number of images or simple related information of a certain level or more, and then the service user is encouraged to view the image or simple related information and select the desired information. You can do it.
  • the method of extracting a plurality of image codes having a predetermined level or more may be extracted from the image recognition server 300 and sent to the mobile terminal 100, or the image recognition server 300 in the mobile terminal 100. There is a method of selecting only image codes having a certain level of comparison rate among a predetermined number of image codes received from.
  • a plurality of pieces of information registered in a corresponding image are arranged first, and the remaining ones are arranged at the rear, rather than having a high comparison ratio of image information.
  • the reason is that most information providers select the same image and register the information, but even though the information is the same image, the information that the information provider wants to register has already been subtracted from the information of the information provider who selected the image that they wanted to register. If it is registered, it is effective to induce selection of the image.
  • FIG. 3 is a reference diagram for explaining a method of automatically finding the same image using an image recognition technique according to an exemplary embodiment of the present invention.
  • the reference value of an image previously registered in an image recognition server, a reference value of a new image, and a comparison value with the previously registered image are identical through an image search. It's a way of thinking of images.
  • the method extracts a comparison value and determines the same value between the reference value of the pre-registered image, the reference value of the new image, and the comparison value of comparing the new image to the pre-registered image.
  • the reference value of the image is extracted through the image recognition server or the test image server to register the image and the image is registered in the image recognition server.
  • the reference value is extracted from the new image through the image recognition server or the test image recognition server.
  • the new image is extracted by the image recognition server.
  • the new image is registered as a new image in the image recognition server when the comparison value for the new image and the reference value of the pre-registered image that is the target of the comparison value are out of a specific error range. If the comparison value of the new image is within the same or a specific error range as the reference value of the previously registered image, and the comparison value of the new image is within the same or specific error range compared to the reference value of the new image, the same image is processed. When the comparison value of the new image is out of a specific error range compared to the reference value of the new image, it is possible to configure to register as a new image.
  • the procedure and method of the steps can be variously modified, but the basic concept is that the reference value of the image registered in the image recognition server, the reference value of the new image, and the comparison value with the previously registered image through the new image search are the same. Or, it is based on a method that is regarded as the same image when it is within an error range, and various methods that can be generated therein are all included in the method.
  • the method of extracting the reference value is not a problem if the reference value can be extracted immediately when registering the image to the image recognition server, but if the image recognition server extracts and stores only the image pattern value of the image
  • the comparison value of the image is a reference value by searching the image recognition server again with the same image.
  • FIG. 4 is an example of an image server connection DB and an information registration DB table for each promotion and shopping information provider according to an exemplary embodiment of the present invention.
  • the promotional code and the shopping mall code are matched for each image code. For example, if an image is registered, an image code of R1512224000100 is generated, and is matched with the image and stored. Promotional code for each information provider 1, 2, 3 is registered in the information registration DB table for promotional information providers, and shopping code is registered in the shopping mall DB.
  • the promotional code is provided in various forms such as introduction, video, photo album, etc.
  • the shopping code is linked to one or more shopping mall link addresses. For example, if an information provider 1 registers a promotion and registers a shopping mall, a promotional code may be assigned and a shopping mall code may also be assigned.
  • promotional code is to register the information you want to register in the introduction, videos, photo albums, etc. If you have one shopping mall in the code, only one URL can be registered, multiple shopping malls can be registered.
  • FIG. 5 is a flowchart illustrating a method of finding an image to be registered from a pre-registered image through keyword search according to an embodiment of the present invention.
  • a PC or mobile terminal 100 connected to the Internet network accesses a web server 200 that provides a service using image recognition technology, and performs keyword search for a specific image (A01). At least one image registered with the keyword is transmitted from the web server 200 and displayed on the screen of the PC or mobile terminal (A02). One of the images displayed on the screen of the PC or mobile terminal 100 is selected and the related information is registered (A03). The selected image and the related information are transmitted to the web server 200 (A04).
  • the step of registering a new image through the PC or mobile terminal 100 is added, or another process Can be added.
  • FIG. 6 is a flowchart illustrating a method of determining whether an image to be registered is a registered image through a search according to an embodiment of the present invention, and selecting and registering an image or a newly registered image according to the result. to be.
  • a PC or mobile terminal 100 connected to the Internet network accesses a web server 200 that provides a service using image recognition technology, uploads and searches a specific image (B01).
  • the web server 200 transmits the received image to the image recognition server 300 (B02).
  • the image recognition server 300 analyzes characteristic pattern values of the transmitted image (B03).
  • the image recognition server 300 compares the pattern value of the transmitted image with a pre-registered image pattern value (B04).
  • the image recognition server 300 transmits at least one or more image information corresponding to the same or most similar pattern value to the web server 200 (B05).
  • the web server 200 transmits the received image information to the PC or the mobile terminal 100 and appears on the screen of the PC or the mobile terminal 100 (B06). At least one of the image or images displayed on the screen of the PC or the mobile terminal 100 and the image used for the search is selected and related information is registered (B07).
  • the selected image and related information are transmitted to the web server 100 (B08).
  • a comparison value (Query Image Score) of the image comparing with the previously registered image pattern value is a predetermined level or more may be added to the step is automatically considered to be the same image.
  • the step of storing the comparison image score (Resxult Image Score), which is obtained by searching for the same image as the registered image instead of the above step, as the reference image score or storing a reference value of a predetermined level or more, and If the comparison value is equal to or higher than the reference value or more than a predetermined comparison rate, the step of automatically registering the same image may be added.
  • the image transmitted from the PC or mobile terminal 100 has a structure that is transmitted to the image recognition server 300 via the web server 200, but the image recognition directly from the PC or mobile terminal 100
  • the server 300 may be configured to transmit or transmit the result to the web server 200 after the result is transmitted from the image recognition server 300 to the PC or the mobile terminal 100.
  • analyzing the characteristic pattern value of the transmitted image may be transmitted to the image recognition server 300 and the image recognition server 300, but it may be analyzed, but the image on the PC or mobile terminal 100 or a separate server Contains a logic to analyze the pattern value of the they may be made to the structure to transmit the pattern value of the image to the image recognition server 300.
  • FIG. 7 is a flowchart for providing related information registered by a plurality of information providers using image recognition technology according to an exemplary embodiment of the present invention.
  • an app that provides a service using the image recognition technology of the mobile terminal 100 is driven to capture a specific image (C01).
  • this image may be an image taken through an APP or an internal camera function installed in a mobile terminal, may be a recognized image, or may be an image secured by a method such as down, scan, or the like.
  • the image is transmitted to the image recognition server 300 (C02).
  • the image recognition server 300 analyzes characteristic pattern values of the transmitted image (C03).
  • the image recognition server 300 compares the pattern value of the transmitted image and the registered image pattern value (C04). At least one image information corresponding to the same or most similar pattern value is transmitted to the mobile terminal 100 (C05).
  • the mobile terminal 100 transmits the received at least one image information to the web server (C06).
  • the web server 200 searches for relevant information registered by at least one or more information providers on the at least one or more image information through a DB server (C07).
  • the web server 200 transmits the at least one or more related information respectively retrieved to the at least one or more image information to the mobile terminal 100 (C08).
  • the searched list of a plurality of related information is displayed on the screen of the mobile device (C09).
  • Promotional information or shopping mall information may be displayed on the mobile screen.
  • the shopping mall UI icon
  • one or more shopping mall information appears, and the shopping mall information may be implemented to select a shopping mall desired by the user.
  • the image transmitted from the PC or the mobile device is transmitted to the image recognition server, but the mobile device is transmitted to the image recognition server via a web server, and the resulting image information is transmitted from the image recognition server to the web server.
  • the web server may be configured to search for relevant information registered by a plurality of information providers corresponding to the image information and transmit the related information to the mobile device.
  • analyzing the characteristic pattern value of the transmitted image may be transmitted to the image recognition server and the image recognition server may be analyzed, but the mobile device or a separate server contains the logic to analyze the pattern value of the image You can also make a structure that analyzes the pattern value of the image and sends it to the image recognition server.
  • the above-described embodiment may be provided by additionally classifying into promotion and shopping according to the information type provided by the information provider.
  • the relevant information corresponding to the image information only the information related to the promotion is initially provided, and the UI (icon) of shopping related information is selected on the screen of the mobile device, and the mobile device transmits the information to a web server.
  • the web server searches for shopping related information registered by at least one information provider through the DB server, and the web server transmits the searched at least one shopping related information to the mobile device.
  • a step of displaying the searched plurality of related information lists on the screen of the mobile device may be further provided.

Abstract

The present invention relates to an information providing method using an image recognition technology, which is connected through a mobile terminal and a network and includes an image recognition server, a webserver, and a DB server. The method comprises the steps of: transmitting, by the mobile terminal, an obtained image or an image pattern value extracted from the obtained image to the image recognition server; extracting the image pattern value when having received the image transmitted from the mobile terminal, and searching for an image pattern value similar to a pre-registered image pattern value when having directly received the image pattern value transmitted thereto, by the image recognition server; transmitting an image code value, which is a searching result value, either through the mobile terminal or directly to a webserver by the image recognition server; and searching for a promotional code value and a shopping code value, which are registered in the received image code value, and transmitting promotion information registered in the promotional code value, shopping mall URL information registered in the shopping mall code value, or the like to the mobile terminal by the webserver.

Description

이미지인식기술을 이용한 정보제공시스템 및 방법Information provision system and method using image recognition technology
본 발명은 모바일 환경에서 이미지인식기술을 이용하여 동일한 이미지에 다수의 정보제공자가 등록한 관련정보를 제공하는 방법에 관한 것이다.The present invention relates to a method for providing related information registered by multiple information providers in the same image using image recognition technology in a mobile environment.
이미지인식기술은 사용 분야에 따라 다양하게 적용되고 있는데, 그 중 하나는 다수의 이미지 중에서 동일한 이미지를 찾아내는 기술이다. 그 일 예를 보면, 스마트폰 등의 카메라가 부착된 모바일기기에서 모바일 APP(Application)을 구동하고, 사용자가 스마트폰을 이용하여 제품의 광고 이미지나 상품의 라벨을 촬영하면 촬영된 이미지는 상기 이미지인식서버로 전달되어 동일한 이미지 또는 가장 유사한 이미지를 찾아내어 해당 이미지에 등록된 정보를 제공하는 기술이다.Image recognition technology is applied in various ways according to the field of use, one of which is to find the same image among multiple images. As an example, when a mobile app (Application) is driven in a mobile device equipped with a camera such as a smartphone, and a user photographs an advertisement image or a label of a product using a smartphone, the photographed image is the image. It is a technology that delivers the information registered in the image by finding the same image or the most similar image that is delivered to the recognition server.
그러나, 이러한 기술을 상품 홍보 판매에 이용할 경우, 부족한 점이 있어 왔다. 예를 들어 광고의 경우, 광고주는 자신의 제품이 어느 쇼핑몰을 통해 판매되든 오프라인을 통해 판매되든 상관없기 때문에 특정 구매사이트로 연결시키기 보다는 제품 홍보에 역점을 둔다. 이에 반하여 상기 제품의 대리점 등 직접 판매업자들의 경우는 고객들이 자신의 판매사이트에서 그 제품을 사게 하기 위하여 자신의 제품 구매사이트로 연결시키기를 원한다. However, there has been a shortcoming when such a technique is used for product promotion sales. For example, in advertising, an advertiser focuses on product promotion rather than linking to a specific purchase site because it does not matter which shopping mall is sold through or offline. In contrast, direct sellers, such as distributors of the product, want customers to connect to their product purchase site in order to have their products buy from their sales site.
상기와 같은 경우, 다수의 판매업자가 동일한 이미지를 자신의 구매사이트로 연결시키려는 현상이 발생하고 있는데, 이를 모두 충족시킬 방법이 현재로서는 존재하지 않는 문제점이 있었다. In this case, a number of vendors are trying to link the same image to their purchase site, but there is a problem that does not exist at present.
본 발명은 상술한 문제점을 해결하기 위하여 안출된 것으로서, 본 발명은 모바일 사용자가 자신의 모바일 단말로부터 이미지를 획득하고, 이미지 인식 기술을 이용하는 APP과 서버 등을 통해 해당 이미지의 판매자들이 자신이 등록한 해당 제품의 구매사이트(인터넷 쇼핑몰 등)로 사용자의 단말을 접속시킬 수 있는 수단을 제공하기 위한 것이다.The present invention has been made to solve the above-described problems, the present invention is a mobile user acquires an image from his mobile terminal, the seller of the image through the APP and the server using the image recognition technology registered by the seller An object of the present invention is to provide a means for connecting a user terminal to a purchase site (such as an internet shopping mall) of a product.
상술한 문제점을 해결하기 위한 기술적 수단으로서, 본 발명의 일측면은 모바일 단말과 네트워크를 통해 접속되며, 이미지인식서버, 웹서버, 및 DB서버를 포함하여 구성된 이미지인식기술을 이용한 정보제공방법에 있어서, 상기 모바일단말에서 획득한 이미지 또는 획득한 이미지로부터 추출한 이미지 패턴값을 상기 이미지인식서버로 전송하는 단계; 상기 이미지인식서버에서 상기 모바일단말로부터 이미지를 전송받았으면 이미지 패턴값을 추출하고, 이미지 패턴값을 직접 전송받았으면 기등록된 이미지 패턴값과 유사한 이미지 패턴값을 검색하는 단계; 상기 이미지인식서버는 검색 결과값인 이미지코드값을 상기 모바일단말을 거쳐 또는 직접 웹서버로 전송하는 단계; 상기 웹서버는 전송받은 이미지코드값에 등록된 홍보용코드값과 쇼핑용코드값을 찾아내고, 홍보용코드값에 등록된 홍보물 정보 또는 쇼핑몰코드값에 등록된 쇼핑몰 URL정보 등을 모바일단말로 전송하는 단계를 포함하는 이미지인식기술을 이용한 정보제공방법을 제공한다.As a technical means for solving the above problems, an aspect of the present invention is connected to a mobile terminal via a network, and in the information providing method using an image recognition technology comprising an image recognition server, a web server, and a DB server Transmitting the image pattern value extracted from the acquired image or the acquired image to the image recognition server; Extracting an image pattern value when the image recognition server receives an image from the mobile terminal, and retrieving an image pattern value similar to a registered image pattern value when the image pattern value is directly transmitted; The image recognition server transmitting an image code value, which is a search result value, through the mobile terminal or directly to a web server; The web server finds a promotional code value and a shopping code value registered in the received image code value, and transmits the promotional material information registered in the promotional code value or the shopping mall URL information registered in the shopping mall code value to a mobile terminal. It provides an information providing method using an image recognition technology comprising a.
바람직하게는, 상기 이미지와 해당 이미지코드값은 사전에 DB서버에 등록되며, 등록 방법은 기존에 등록되지 않은 새로운 이미지를 등록하는 방법 또는 기등록된 이미지에서 원하는 이미지를 찾아 해당 이미지에 자신이 원하는 정보를 추가로 등록할 수 있다.Preferably, the image and the corresponding image code value is registered in advance in the DB server, the registration method is to register a new image that has not been previously registered or to find a desired image from the pre-registered image to the desired image You can register additional information.
또한, 상기 기 등록된 이미지에서 원하는 이미지를 찾는 방법은, 사전에 이미지를 등록할 때, 등록되는 이미지에 해당되는 키워드도 함께 등록하도록 하여 키워드 검색을 통해 찾는 방법, 또는 이미지 검색을 통해 동일 또는 가장 유사한 이미지를 검색하여 선택하거나 아니면 검색에 사용한 이미지를 등록하는 방법, 또는 기등록된 패턴값과 동일하거나 등록하려는 이미지와 기등록된 이미지의 패턴값의 오차가 극히 적은 것을 동일이미지로 간주할 수도 있다.In addition, a method of finding a desired image from the previously registered image is a method of searching through a keyword search by registering a keyword corresponding to a registered image when registering an image in advance, or by searching the same or most through an image search. The same image may be regarded as a method of searching for and selecting a similar image, or registering an image used for the search, or having an extremely small difference between the pattern value of an image to be registered and an image to be registered and the same as the registered pattern value. .
본 발명의 다른 측면은 모바일 단말과 네트워크를 통해 접속된 이미지인식기술을 이용한 정보제공시스템에 있어서, 전달받은 이미지에 대해서 유사한 이미지를 검색하고, 검색된 유사 이미지에 대응된 이미지코드값을 검색하는 이미지인식서버; 및 상기 이미지인식서버에 접속된 DB서버로서, 하나의 이미지코드값에 다수의 정보를 제공하기 위하여, 이미지에 대응되는 이미지코드값에 다수의 홍보 및/또는 쇼핑 관련 정보를 저장하는 DB서버;를 포함하되, 상기 모바일 단말이 획득한 이미지를 통해 정보제공을 원하는 경우, 해당 이미지에 대한 다수의 홍보 및/또는 쇼핑 관련 정보 중에서 선택하여 연결할 수 있도록 구성된다. Another aspect of the present invention is an information providing system using an image recognition technology connected through a mobile terminal and a network, the image recognition to search for a similar image with respect to the received image, the image code value corresponding to the retrieved similar image server; And a DB server connected to the image recognition server, the DB server storing a plurality of promotional and / or shopping related information in an image code value corresponding to the image in order to provide a plurality of information in one image code value. Including, but if the mobile terminal wants to provide information through the acquired image, it is configured to connect to select from a plurality of promotional and / or shopping-related information for the image.
본 발명에 의한 효과는 다음과 같다.The effect by the present invention is as follows.
첫째, 서버에 등록되는 이미지 개수가 대폭 줄어드는 효과가 있다. 서버는 검색 속도를 높이기 위하여 모든 이미지 패턴값을 램(RAM)에 상주시킨 상태에서 검색을 하는데, 동일한 이미지를 등록하는 수만큼 모두 등록을 받는 것 보다는 하나의 이미지만을 등록받고 다수개의 링크주소나 관련정보를 연결하는 것이 서버의 데이터 용량을 줄일 뿐만 아니라, 검색속도 또한 빠르게 하는 효과가 있는 것이다.First, the number of images registered on the server is greatly reduced. In order to speed up the search, the server searches in a state where all image pattern values are resident in RAM. Instead of registering the same number of images, only one image is registered and multiple link addresses or related information are searched. Linking information not only reduces the data capacity of the server, but also speeds up the search.
둘째, 하나의 이미지에 다수개의 구매사이트 링크주소를 연결하게 되면 소비자는 보다 저렴한 가격을 제시하는 구매사이트에서 제품을 구매할 수 있고, 인터넷 포털의 프리미엄 링크와 같이 상단에 표시하는 정보를 입찰을 통해 보다 비싼 가격에 제공할 수가 있어 보다 높은 수익창출이 가능하다.Second, by linking multiple purchase site link addresses to a single image, consumers can purchase products at a lower price purchase site, and the information displayed at the top, such as a premium link on an Internet portal, can be viewed through auctions. Higher profits are possible because they can be provided at higher prices.
도 1은 본 발명의 실시예에 따라서 이미지 인식기술을 이용한 정보제공시스템을 도시하고 있다.1 illustrates an information providing system using an image recognition technology according to an exemplary embodiment of the present invention.
도 2는 본 발명의 실시예에 따라서 이미지인식 기술을 이용하여 유사이미지의 동일여부를 식별하는 과정을 설명하기 위한 참고도이다. 2 is a reference diagram for explaining a process of identifying whether or not a similar image is identical by using an image recognition technology according to an embodiment of the present invention.
도 3은 본 발명의 실시예에 따라서 이미지인식 기술을 이용하여 동일이미지를 자동으로 찾아내는 방법에 관한 것이다.Figure 3 relates to a method for automatically finding the same image using image recognition technology in accordance with an embodiment of the present invention.
도 4는 본 발명의 실시예에 따라서 이미지 연결 DB와 홍보, 쇼핑 정보제공자별 정보등록 DB 테이블의 일 예이다.4 is an example of an image connection DB and an information registration DB table for each promotion and shopping information provider according to an exemplary embodiment of the present invention.
도 5는 본 발명의 실시예에 따라서 키워드 검색을 통해 기 등록된 이미지에서 등록하고자 하는 이미지를 찾아내는 방법에 대한 흐름도이다.5 is a flowchart illustrating a method of finding an image to be registered from a pre-registered image through keyword search according to an embodiment of the present invention.
도 6은 본 발명의 실시예에 따라서 등록하고자 하는 이미지를 검색을 통해 기등록된 이미지인지 유무를 판단하고, 그 결과에 따라 기등록된 이미지 또는 새로 등록하는 이미지를 선택하여 등록하는 방법에 대한 흐름도이다.6 is a flowchart illustrating a method of determining whether an image to be registered is a registered image through a search according to an embodiment of the present invention, and selecting and registering an image or a newly registered image according to the result. to be.
도 7은 본 발명의 실시예에 따라서 다수의 정보제공자가 등록한 관련정보를 이미지 인식기술을 이용하여 정보하는 제공하는 흐름도이다.FIG. 7 is a flowchart for providing related information registered by a plurality of information providers using image recognition technology according to an exemplary embodiment of the present invention.
이하, 본 발명의 이점 및 특징, 그리고 그것들을 달성하는 방법은 첨부되는 도면과 함께 상세하게 후술되어 있는 실시예들을 참조하면 명확해질 것이다. 그러나 본 발명은 이하에서 개시되는 실시예들에 한정되는 것이 아니라 서로 다른 다양한 형태로 구현될 것이며, 단지 본 실시예들은 본 발명의 개시가 완전하도록 하며, 본 발명이 속하는 기술분야에서 통상의 지식을 가진 자에게 발명의 범주를 완전하게 알려주기 위해 제공되는 것이며, 본 발명은 청구항의 범주에 의해 정의될 뿐이다. Advantages and features of the present invention, and a method of achieving them will be apparent from the following detailed description with reference to the accompanying drawings. However, the present invention is not limited to the embodiments disclosed below, but may be implemented in various forms. It is provided to fully convey the scope of the invention to those skilled in the art, the invention being defined only by the scope of the claims.
도 1은 본 발명의 실시 예에 따라고 이미지 인식기술을 이용한 정보제공시스템을 도시하고 있다. 1 illustrates an information providing system using an image recognition technology according to an embodiment of the present invention.
도 1의 이미지 인식기술을 이용한 정보제공시스템은 모바일 단말(100)과 네트워크를 통해서 접속되어 있고, 웹서버(200), 이미지인식서버(300) 등을 포함하여 구성된다. 이미지인식서버(300)는 이미지서버 연결 DB(310), 정보제공자별 정보등록 DB(320)에 접속된다. 정보제공자별 정보등록 DB(320)는 홍보DB (320A), 쇼핑DB (320B)를 포함하여 구성될 수 있다. 각종 DB들은 설명의 편의상 분류된 것으로 모두 총괄하여 DB서버(305)로 명명할 수도 있다.The information providing system using the image recognition technology of FIG. 1 is connected to the mobile terminal 100 through a network, and includes a web server 200, an image recognition server 300, and the like. The image recognition server 300 is connected to the image server connection DB 310, information registration DB 320 for each information provider. The information provider DB information for each provider 320 may be configured to include a promotion DB (320A), shopping DB (320B). Various DBs are classified for convenience of description and may be collectively named as DB server 305.
본 실시예에 의하면, 하나의 이미지에 하나의 정보만이 제공될 때에는 이미지인식서버(300)와 연결되는 이미지코드는 단일 코드로 구성가능하지만, 하나의 이미지에 다수의 정보제공자가 등록한 정보를 제공하기 위해서는 이미지인식서버(300)의 이미지서버 연결 DB(310)와 별도로 정보제공자별 정보등록 DB가 별도로 존재하는 것이 바람직하다. 이미지서버 연결 DB(310)에는 하나의 이미지 정보에 적어도 하나 이상의 정보제공자별 정보등록 DB(320)와 연결되는 코드값이 저장된다.According to this embodiment, when only one information is provided in one image, the image code connected to the image recognition server 300 may be configured as a single code, but the information provided by a plurality of information providers is provided in one image. In order to do this, it is preferable that there is a separate information registration DB for each information provider separately from the image server connection DB 310 of the image recognition server 300. The image server connection DB 310 stores a code value connected to at least one information provider-specific information registration DB 320 in one image information.
각 이미지는 모바일 단말(100)에 의하여 등록되거나 본 시스템 운영자가 등록하는 것이 가능하다. 또한, 이미지를 등록하는 방법은 홍보물 등록과 쇼핑몰 등록으로 구분하여 홍보DB (320A), 쇼핑DB (320B)에 등록할 수 있다.Each image may be registered by the mobile terminal 100 or registered by the system operator. In addition, a method of registering an image may be classified into a promotional material registration and a shopping mall registration, and may be registered in the promotion DB 320A and the shopping DB 320B.
여기서 홍보라 하면 소개글 또는 사진첩, 동영상 등 여러 형태의 홍보물을 의미하고, 판매를 위한 구매사이트라 함은 인터넷 쇼핑몰 등을 의미한다.Here, the term "public relations" refers to various types of promotional materials such as introduction articles, photo albums, and videos, and a purchase site for sale refers to an Internet shopping mall.
모바일단말(100)은 본 발명의 이미지 인식기술을 이용한 정보를 제공하기 위한 어플리케이션 APP을 설치하여 보유한다. 모바일단말(100)의 기능을 보면, 모바일 단말에서 확보한 이미지 또는 이미지 패턴값을 이미지인식서버(300)로 전송한다. 이 이미지는 모바일단말에 설치된 APP 또는 내부 카메라 기능을 통해서 촬영된 이미지일 수도 있고 인식된 이미지일 수도 있고, 별도로 다운, 스캔 등의 방식으로 확보하고 있는 이미지일 수도 있다. 바람직한 방식은 카메라 촬영 또는 인식된 이미지이다. 이에 대한 응답으로, 모바일 단말(100)은 이미지인식서버(300)로부터 검색결과값(이미지코드값)을 전달받아 이를 웹서버(200)로 전송한다.The mobile terminal 100 installs and holds an application APP for providing information using the image recognition technology of the present invention. Looking at the function of the mobile terminal 100, and transmits the image or image pattern value obtained from the mobile terminal to the image recognition server (300). This image may be an image taken through an APP or an internal camera function installed in a mobile terminal, may be a recognized image, or may be an image secured by a method such as down, scan, or the like separately. The preferred method is camera shot or recognized image. In response, the mobile terminal 100 receives the search result value (image code value) from the image recognition server 300 and transmits it to the web server 200.
이미지인식서버(300)는 모바일 단말(100)로부터 이미지를 전송받았으면 이미지 패턴값을 추출하고, 이미지 패턴값을 직접 전송받았으면 기 등록된 이미지 패턴값 중 가장 유사한 이미지 패턴값을 검색하는 기능을 수행하며, 검색 결과값(이미지코드값)을 모바일 단말(100)로 전송한다. 다른 방식으로는 검색결과값을 모바일 단말(100)로 전송하지 않고 웹서버(200)로 전송하는 것도 가능하다.The image recognition server 300 extracts an image pattern value when the image is received from the mobile terminal 100, and retrieves the most similar image pattern value among the registered image pattern values when the image pattern value is directly transmitted. The search result value (image code value) is transmitted to the mobile terminal 100. Alternatively, the search result value may be transmitted to the web server 200 without transmitting the search result value to the mobile terminal 100.
웹서버(200)는 정보제공자별 정보등록 DB(320)에서 전송받은 이미지코드값에 등록된 홍보용코드값과 쇼핑용코드값을 찾아내고, 홍보용코드값에 등록된 홍보물(소개글, 사진첩, 동영상) 정보 또는 쇼핑몰코드값에 등록된 쇼핑몰 URL정보 등을 모바일 단말(100)로 전송하고 이를 전달받은 모바일 단말(1000은 모바일 화면에 홍보물 정보 또는 쇼핑몰 정보가 표시된다. 특히, 쇼핑몰 정보는 쇼핑몰 UI(아이콘)을 선택하면 1개 또는 다수개의 쇼핑몰 정보가 나타나, 그 중 사용자가 원하는 쇼핑몰을 선택한다. The web server 200 finds the promotional code value and the shopping code value registered in the image code value received from the information registration DB 320 for each information provider, and the promotional material (introduction, photo album, video) registered in the promotional code value. ) Or the shopping mall URL information registered in the shopping mall code value to the mobile terminal 100, and the mobile terminal 1000 receiving the received information is displayed on the mobile screen. Icon), one or more shopping mall information appears, and a user selects a shopping mall among them.
본 실시예에서는 이미지인식서버(300)는 동일한(가장 유사한) 이미지를 찾아 이미지코드값을 전달하는 역할을 하고, 웹서버(200)는 그렇게 찾아낸 이미지코드값이 저장된 DB서버에서 홍보용코드값과 쇼핑몰코드값을 찾아내고, 또 다시 홍보용코드값과 쇼핑몰 코드값에 등록된 정보를 찾아내는 중개자적인 역할을 하는 것으로 기술되어 있다. 그러나, 필요에 따라서는 이미지인식서버(300)와 웹서버(200)는 단일 서버로 구성하는 것도 가능하다. In this embodiment, the image recognition server 300 plays a role of delivering the image code value by finding the same (most similar) image, and the web server 200 stores the promotional code value and the shopping mall in the DB server where the image code value is found. It is described as playing an intermediary role to find the code value and again to find the information registered in the promotional code value and the shopping mall code value. However, if necessary, the image recognition server 300 and the web server 200 may be configured as a single server.
이하, 이미지 등록 방법을 상세히 설명한다.Hereinafter, the image registration method will be described in detail.
이미지 등록 방법은, i)기존에 등록되지 않은 새로운 이미지를 등록하는 방법과 ii)기등록된 이미지에서 원하는 이미지를 찾아 해당 이미지에 자신이 원하는 정보를 추가로 등록하는 방법이 존재할 수 있다.The image registration method may include i) a method of registering a new image not previously registered, and ii) a method of finding a desired image from the registered image and additionally registering information desired by the user.
기 등록된 이미지에서 원하는 이미지를 찾는 방법으로는, ii-1) 사전에 이미지를 등록할 때, 등록되는 이미지에 해당되는 키워드도 함께 등록하도록 하여 키워드 검색을 통해 찾는 방법이 있고, ii-2) 이미지 검색을 통해 동일 또는 가장 유사한 이미지를 검색하여 선택하거나 아니면 검색에 사용한 이미지를 등록하는 방법이 있다. ii-3) 이외에 기등록된 패턴값과 동일하거나 등록하려는 이미지와 기등록된 이미지의 패턴값의 오차가 극히 적은 것을 동일이미지로 간주할 수도 있는 방법이 있다. 그러나, 이 방법에 의하면 이미지에 따라서 전혀 다른 이미지가 동일 이미지로 간주되는 경우가 있다. 따라서 이 방법을 이용할 때는 일정 기준을 가지도록 구성하는 것이 효과적일 수 있다. As a method of searching for a desired image from a registered image, ii-1) when registering an image in advance, the keyword corresponding to the registered image is also registered, so that the keyword can be found through keyword search. Ii-2) There is a method of searching for and selecting the same or most similar image through image search, or registering an image used in the search. In addition to ii-3), there may be a method in which the image having the same or identical error value between the registered pattern value and the pattern value of the registered image may be regarded as the same image. However, according to this method, a completely different image may be regarded as the same image depending on the image. Therefore, when using this method, it may be effective to configure a standard.
이미지를 등록하기 위해 동일 이미지 여부를 판단하는 방법을 좀 더 상세히 설명한다. How to determine whether the same image to register the image will be described in more detail.
첫번째 방법으로는 새로 등록하려는 이미지의 패턴값이 기 등록된 이미지의 패턴값과 일정 수치 이상이면 동일이미지로 간주하는 방법이다. 즉, 패턴값이 많이 추출되면 추출될수록 비교 이미지와의 비교한 결과값인 비교값(Query Image Score)이 높게 나오면 다른 이미지를 동일한 것으로 간주할 가능성이 극히 적을 것이므로 비교 이미지의 비교값이 일정 수준 이상 높게 나오면 동일이미지로 간주하는 방법이다. 위 방법에서는 일정 수준 이하에서 동일이미지를 찾아내는 방법이 추가로 필요하다.In the first method, if the pattern value of the image to be newly registered is equal to or larger than a predetermined value with the pattern value of the previously registered image, it is regarded as the same image. In other words, if more pattern values are extracted, the higher the value of Query Image Score compared to the comparison image, the less likely it is to consider another image as the same. If it is high, it is regarded as the same image. In the above method, it is necessary to additionally find the same image below a certain level.
두번째 방법으로는 기 등록된 이미지와 새로 등록하는 이미지의 일정 수준 이상의 비교율에 따라 판단하는 방법이 있다. 이 방법에 의하면, 매칭률을 계산하기 위해서는 기 등록된 이미지의 패턴값에서 기준값(Resxult Image Score)을 추출하여 비교 이미지의 비교값(Query Image Score)과 비교율을 계산해 낼 수 있을 것이다.As a second method, there is a method of judging based on a comparison ratio of a predetermined level or more between a pre-registered image and a newly registered image. According to this method, in order to calculate a matching rate, a reference value (Resxult Image Score) may be extracted from a pattern value of a pre-registered image to calculate a comparison value (Query Image Score) and a comparison rate of the comparison image.
기등록된 이미지의 패턴값에서 기준값을 산출하는 방법은 이미지엔진서버(300)에서 추출한 패턴값의 갯수를 기준값으로 삼을 수도 있지만, 기 등록된 이미지와 동일한 이미지를 검색하여 기등록된 이미지와 비교하여 산출된 비교값을 기준값으로 삼을 수도 있다. 위 방법도 첫번째 방법과 마찬가지로 비교값이 일정 수준 이하에서는 다른 이미지를 동일 이미지로 간주할 수가 있다.The method of calculating the reference value from the pattern value of the pre-registered image may be based on the number of pattern values extracted from the image engine server 300, but the same image as the pre-registered image is searched and compared with the pre-registered image. The comparison value calculated by reference may be used as a reference value. Similar to the first method, the above method can regard different images as the same image if the comparison value is below a certain level.
결국 비교값으로 비교하든 비교율로 계산하든 결국 일정 비교값 이하에서는 동일 이미지 판단 유무는 인간의 판단을 빌리는 방법이 효과적일 수 있다. 그 이유는 예를 들어 특정 상품로고의 이미지를 등록하고, 동일한 상기 상품로고로 검색했을 때와 상기 상품로고에 상품명이 포함된 이미지를 검색했을 때는 경우에 따라 그 비교값이 2개 다 동일하게 나올 수 있기 때문이다.After all, whether compared with a comparison value or calculated with a comparison rate, a method of borrowing human judgment may be effective to determine whether or not the same image is determined below a predetermined comparison value. The reason for this is, for example, when an image of a specific product logo is registered and the same product logo is searched, and when an image including a product name is included in the product logo, both comparison values may be identical. Because it can.
도 2는 본 발명의 실시예에 따라서 이미지인식 기술을 이용하여 유사이미지의 동일여부를 식별하는 과정을 설명하기 위한 참고도이다. 2 is a reference diagram for explaining a process of identifying whether or not a similar image is identical by using an image recognition technology according to an embodiment of the present invention.
도 2를 참조하면, 그림 (a)와 그림 (b)는 제품 내에 G1, G1FE란 글자만 다르고 모두 동일하지만 두 제품은 모델명이 다른 제품이다. 반면, 그림 (c)는 하단에 GS칼텍스라는 글자를 세겨 넣었을 뿐 모델명이 그림 (a)와 동일한 제품이다.Referring to Figure 2, Figure (a) and Figure (b) in the product, except for the letters G1, G1FE are all the same, but the two products are different model name. On the other hand, Figure (c) is the same product as Figure (a), with the letters GS Caltex engraved at the bottom.
여기에 기존의 이미지인식기술을 결합시켜 보면 그 패턴값은 그림 (a)와 그림 (b)의 차이 보다 그림 (a)와 그림 (c)의 오차가 훨씬 크게 나와 그림 (a)는 그림 (c)보다 그림 (b)와 더 유사한 이미지로 인식될 가능성이 높다. 이러한 이유에서 이 같은 문제점을 해결할 수 있는 방법은 인간의 인지 능력을 이용하는 방식이 효과적일 수 있다. 여기서 인간의 인지 능력을 빌린다는 것은 인간의 인지 능력을 시스템의 일부분으로 적용하는 것이 아니라 단말기 화면에 나타난 이미지 중 선택되어지는 단계가 추가되는 것을 의미하는 것이다.Combined with the existing image recognition technology, the pattern value is much larger than the difference between Fig. (A) and Fig. (B), and the error of Fig. (A) and Fig. (C) is much larger. ) Is more likely to be perceived as an image similar to Figure (b). For this reason, the method of using the human cognitive ability may be effective to solve this problem. In this case, borrowing the human cognitive ability does not apply the human cognitive ability as part of the system, but means that a step of selecting one of the images displayed on the terminal screen is added.
결국 정보제공자가 이미지를 등록할 때, 해당 이미지와 유사한 이미지가 기존 이미지인식서버(300)에 등록되어 있는지를 검색하여 비교 이미지의 비교값(Query Image Score)이 일정 수준 이상이면 동일한 이미지로 간주하여 자동 등록되게끔 하고, 일정 수준 이하에서는 가장 유사한 이미지들을 적어도 1개 이상을 화면에 노출되도록 하고 정보제공자가 기 등록된 이미지와 자신이 애초에 등록하려는 이미지 중 사용할 이미지를 선택하도록 구현하는 것이 바람직할 수 있다. As a result, when an information provider registers an image, it searches for whether an image similar to that image is registered in the existing image recognition server 300, and considers the same image if the comparison image (Query Image Score) of the comparison image is above a certain level. It may be desirable to enable automatic registration, and below a certain level, to expose at least one or more of the most similar images on the screen, and to allow the information provider to select which image to use between the pre-registered image and the image that he or she would like to initially register. have.
다른 방식에 의하면, 동일한 이미지가 추출되어 있음에도 사용자가 자신이 등록하려는 이미지를 동일이미지로 선택하여 등록하도록 구성하는 경우이다. 이 경우 이미지인식서버(100)에는 2개의 동일 이미지가 등록되게 되는데 이 때의 해결 방법도 필요하게 된다. 이의 해결방법은 비교 이미지의 비교값은 높은 순서 대로 다수개의 비교 이미지를 검색하여 가장 높은 비교 이미지의 비교값(Query Image Score) 또는 기 등록된 이미지의 기준값(Result Image Score)과 검색된 다수개의 이미지의 비교값과의 비교율을 계산하여 비교율이 일정 수준 이상의 이미지 또는 간단한 관련 정보를 화면 상에 다수개 나타나도록 한 뒤 서비스 이용자가 해당 이미지 또는 간단한 관련 정보를 보고 자신이 원하는 정보를 선택하도록 유도하는 방식을 취할 수 있다.According to another method, the user selects and registers an image to be registered as the same image even though the same image is extracted. In this case, two identical images are registered in the image recognition server 100, but a solution in this case is also required. The solution is that the comparison values of the comparison images are searched for a plurality of comparison images in ascending order, so that the comparison value (Query Image Score) of the highest comparison image or the reference value (Result Image Score) of the registered image and the plurality of retrieved images are compared. By calculating the comparison rate with the comparison value, the comparison rate is displayed to a number of images or simple related information of a certain level or more, and then the service user is encouraged to view the image or simple related information and select the desired information. You can do it.
상기 방법에서 일정 수준 이상의 비교율을 갖는 다수개의 이미지코드를 추출하는 방법은 이미지인식서버(300)에서부터 추출하여 모바일 단말(100)에 보낼 수도 있고, 모바일 단말(100)에서 이미지인식서버(300)에서 전송받은 일정 갯수 이상의 이미지코드 중 일정 수준 이상의 비교율을 갖는 이미지코드 만을 선택하는 방법이 있다.In the above method, the method of extracting a plurality of image codes having a predetermined level or more may be extracted from the image recognition server 300 and sent to the mobile terminal 100, or the image recognition server 300 in the mobile terminal 100. There is a method of selecting only image codes having a certain level of comparison rate among a predetermined number of image codes received from.
상기 방법에서 이미지 정보의 배열 방법 비교율이 높은 것 보다는 해당 이미지에 등록된 정보가 많은 것을 우선 시하여 다수개 배열하고 나머지는 후미에 배열하는 것이 바람직하다. 그 이유는 거의 대부분의 정보제공자는 동일한 이미지를 선택하여 정보를 등록하겠지만, 굳이 동일한 이미지임에도 자신이 등록하고자 했던 이미지를 선택한 정보제공자의 정보는 후미로 빼어 정보제공자들이 자신이 등록하고자 하는 이미지가 이미 등록되어 있으면 해당 이미지를 선택하게끔 유도하는 효과가 있기 때문이다.In the above method, it is preferable that a plurality of pieces of information registered in a corresponding image are arranged first, and the remaining ones are arranged at the rear, rather than having a high comparison ratio of image information. The reason is that most information providers select the same image and register the information, but even though the information is the same image, the information that the information provider wants to register has already been subtracted from the information of the information provider who selected the image that they wanted to register. If it is registered, it is effective to induce selection of the image.
도 3은 본 발명의 실시예에 따라서 이미지인식 기술을 이용하여 동일이미지를 자동으로 찾아내는 방법을 설명하기 위한 참고도이다. 3 is a reference diagram for explaining a method of automatically finding the same image using an image recognition technique according to an exemplary embodiment of the present invention.
도 3을 참조하면, 상술한 방법을 자동화한 실시 예로서 이미지인식서버에 기 등록된 이미지의 기준값과 신규 이미지의 기준값, 신규 이미지검색을 통해 상기 기 등록된 이미지와의 비교값이 모두 동일한 것을 동일이미지로 간주하는 방식이다.Referring to FIG. 3, the reference value of an image previously registered in an image recognition server, a reference value of a new image, and a comparison value with the previously registered image are identical through an image search. It's a way of thinking of images.
다시 말하면, 비교값을 추출하여 기 등록된 이미지의 기준값과 신규 이미지의 기준값, 신규 이미지를 기 등록된 이미지에 비교한 비교값간의 동일 유무를 판단하여 결정하는 방식이다. In other words, the method extracts a comparison value and determines the same value between the reference value of the pre-registered image, the reference value of the new image, and the comparison value of comparing the new image to the pre-registered image.
보다 구체적으로 설명하면, 등록하고자 하는 이미지를 이미지인식서버 또는 테스트 이미지서버를 통해 상기 이미지의 기준값을 추출하고 이미지인식서버에 상기 이미지를 등록한다. 다음으로, 신규 이미지를 이미지인식서버 또는 테스트이미지인식서버를 통해 기준값을 추출한다. 신규 이미지를 이미지인식서버를 통해 가장 근접한 비교값을 추출한다. 신규 이미지에 대한 비교값과 비교값의 대상이 되는 상기 기 등록된 이미지의 기준값과 비교하여 특정 오차범위를 벗어나면 상기 이미지인식서버에 신규 이미지를 새로운 이미지로 등록한다. 신규 이미지의 비교값이 상기 기 등록된 이미지의 기준값과 동일 또는 특정 오차범위 이내이고, 상기 신규 이미지의 비교값이 상기 신규 이미지의 기준값과 비교하여 동일 또는 특정 오차범위 이내이면 동일 이미지로 처리한다. 상기 신규 이미지의 비교값이 상기 신규 이미지의 기준값과 비교하여 특정 오차범위을 벗어나면 새로운 이미지로 등록되도록 구성하는 것이 가능하다. In more detail, the reference value of the image is extracted through the image recognition server or the test image server to register the image and the image is registered in the image recognition server. Next, the reference value is extracted from the new image through the image recognition server or the test image recognition server. The new image is extracted by the image recognition server. The new image is registered as a new image in the image recognition server when the comparison value for the new image and the reference value of the pre-registered image that is the target of the comparison value are out of a specific error range. If the comparison value of the new image is within the same or a specific error range as the reference value of the previously registered image, and the comparison value of the new image is within the same or specific error range compared to the reference value of the new image, the same image is processed. When the comparison value of the new image is out of a specific error range compared to the reference value of the new image, it is possible to configure to register as a new image.
단계의 수순과 방식은 다양하게 변형이 가능하지만, 그 기본 개념은 이미지인식서버에 기 등록된 이미지의 기준값과 신규 이미지의 기준값, 신규 이미지검색을 통해 상기 기 등록된 이미지와의 비교값이 모두 동일 또는 오차범위 이내일 때 동일이미지로 간주하는 방법을 근간으로 하며 거기서 생성될 수 있는 다양한 방법들은 모두 상기 방법에 포함되는 것으로 한다.The procedure and method of the steps can be variously modified, but the basic concept is that the reference value of the image registered in the image recognition server, the reference value of the new image, and the comparison value with the previously registered image through the new image search are the same. Or, it is based on a method that is regarded as the same image when it is within an error range, and various methods that can be generated therein are all included in the method.
한편, 상기에서 기준값을 추출하는 방법은 이미지인식서버에 이미지를 등록할 시 기준값이 바로 추출될 수 있는 시스템이면 문제가 없지만, 상기 이미지인식서버가 이미지패턴값만 추출하여 저장되는 경우에는 해당 이미지의 기준값 추출을 위하여 동일한 이미지로 상기 이미지인식서버에 다시 검색을 통해 상기 이미지의 비교값을 기준값으로 삼는 방법이 있다.On the other hand, the method of extracting the reference value is not a problem if the reference value can be extracted immediately when registering the image to the image recognition server, but if the image recognition server extracts and stores only the image pattern value of the image In order to extract the reference value, there is a method of using the comparison value of the image as a reference value by searching the image recognition server again with the same image.
또한, 상기 단계에서의 수순은 일부 변경되더라도 그 결과값이 동일하면 동일한 것으로 간주한다.In addition, the procedure in this step is regarded as the same if the result value is the same even if it is partially changed.
도 4는 본 발명의 실시예에 따라서 이미지서버 연결 DB와 홍보, 쇼핑 정보제공자별 정보등록 DB 테이블의 일예이다.4 is an example of an image server connection DB and an information registration DB table for each promotion and shopping information provider according to an exemplary embodiment of the present invention.
도 4을 참조하면, 각각의 이미지 코드에 대비하여 홍보코드와 쇼핑몰 코드가 매칭된다. 예를 들어 임의의 이미지가 등록되면 R1512224000100이라는 이미지 코드가 생성되어 해당 이미지와 매칭되어 저장된다. 홍보 정보제공자별 정보등록 DB 테이블에는 정보제공자1, 2, 3 각각의 홍보용 코드가 등록이 되고, 쇼핑몰 DB에는 쇼핑용 코드가 등록이 됩니다.Referring to FIG. 4, the promotional code and the shopping mall code are matched for each image code. For example, if an image is registered, an image code of R1512224000100 is generated, and is matched with the image and stored. Promotional code for each information provider 1, 2, 3 is registered in the information registration DB table for promotional information providers, and shopping code is registered in the shopping mall DB.
그리고 홍보용 코드는 홍보할 수 있는 여러가지 요소인 소개글, 동영상, 사진첩 등의 여러 형태로 제공된다는 뜻이고, 쇼핑용 코드는 등록한 한개 또는 다수개 쇼핑몰 링크 주소로 연결됩니다. 예를 들어, 정보제공자 1이 홍보도 등록하고 쇼핑몰도 등록하겠다라고 하면 홍보용 코드도 할당 받고, 쇼핑몰 코드도 할당할 수있다. In addition, the promotional code is provided in various forms such as introduction, video, photo album, etc., and the shopping code is linked to one or more shopping mall link addresses. For example, if an information provider 1 registers a promotion and registers a shopping mall, a promotional code may be assigned and a shopping mall code may also be assigned.
 그리고 홍보용 코드에는 소개글, 동영상, 사진첩 등에서 자신이 등록하고 싶은 정보를 등록하는 것이고, 쇼핑몰 코드에는 내가 연결할 쇼핑몰이 1개면 1개의 URL만 등록하고, 다수개의 쇼핑몰도 등록할 수 있다. And the promotional code is to register the information you want to register in the introduction, videos, photo albums, etc. If you have one shopping mall in the code, only one URL can be registered, multiple shopping malls can be registered.
 이하, 본 발명의 실시예에 따른 이미지 인식기술을 이용한 정보제공방법을 상세히 설명한다.Hereinafter, an information providing method using an image recognition technology according to an embodiment of the present invention will be described in detail.
도 5는 본 발명의 실시예에 따라서 키워드 검색을 통해 기 등록된 이미지에서 등록하고자 하는 이미지를 찾아내는 방법에 대한 흐름도이다.5 is a flowchart illustrating a method of finding an image to be registered from a pre-registered image through keyword search according to an embodiment of the present invention.
본 방식은 동일한 이미지에 다수의 정보제공자가 각기 다른 정보를 등록하는 방법이다. 먼저, 인터넷망에 연결된 PC 또는 모바일단말(100)이 이미지인식기술을 활용한 서비스를 제공하는 웹서버(200)에 접속하여 특정 이미지에 대한 키워드 검색을 한다(A01). 웹서버(200)로부터 해당 키워드로 등록된 이미지가 적어도 하나 이상 전송되어 상기 PC 또는 모바일단말의 화면에 나타난다(A02). PC 또는 모바일단말(100)의 화면에 나타난 이미지 중 하나가 선택되어지고 관련정보가 등록된다(A03). 선택된 이미지 및 관련정보가 웹서버(200)로 전송된다(A04). In this method, multiple information providers register different information in the same image. First, a PC or mobile terminal 100 connected to the Internet network accesses a web server 200 that provides a service using image recognition technology, and performs keyword search for a specific image (A01). At least one image registered with the keyword is transmitted from the web server 200 and displayed on the screen of the PC or mobile terminal (A02). One of the images displayed on the screen of the PC or mobile terminal 100 is selected and the related information is registered (A03). The selected image and the related information are transmitted to the web server 200 (A04).
본 실시예에서 검색 키워드에 해당하는 이미지가 없거나 키워드로 검색되어진 이미지들 중 정보제공자가 원하는 이미지가 없을 시에는, PC 또는 모바일단말(100)를 통해 새로운 이미지를 등록하는 단계가 추가되거나, 다른 과정이 추가될 수 있다.In the present embodiment, when there is no image corresponding to the search keyword or the image searched by the keyword does not have the desired image, the step of registering a new image through the PC or mobile terminal 100 is added, or another process Can be added.
도 6은 본 발명의 실시예에 따라서 등록하고자 하는 이미지를 검색을 통해 기등록된 이미지인지 유무를 판단하고, 그 결과에 따라 기등록된 이미지 또는 새로 등록하는 이미지를 선택하여 등록하는 방법에 대한 흐름도이다.6 is a flowchart illustrating a method of determining whether an image to be registered is a registered image through a search according to an embodiment of the present invention, and selecting and registering an image or a newly registered image according to the result. to be.
먼저, 인터넷망에 연결된 PC 또는 모바일단말(100)이 이미지인식기술을 활용한 서비스를 제공하는 웹서버(200)에 접속하여 특정 이미지를 업로드하여 검색한다(B01). 웹서버(200)는 상기 전송받은 이미지를 이미지인식서버(300)로 전송한다(B02). 이미지인식서버(300)는 상기 전송된 이미지의 특징적 패턴값을 분석한다(B03). 이미지인식서버(300)는 상기 전송된 이미지의 패턴값과 기 등록된 이미지 패턴값을 비교한다(B04). 이미지인식서버(300)는 동일 또는 가장 유사한 패턴값에 해당하는 이미지 정보를 적어도 하나 이상을 웹서버(200)로 전송한다(B05). 웹서버(200)는 전송받은 이미지 정보를 PC 또는 모바일단말(100)로 전송하여 상기 PC 또는 모바일단말(100)의 화면에 나타난다(B06). PC 또는 모바일단말(100)에서 화면에 나타난 이미지 또는 이미지들과 상기 검색에 사용된 이미지 중 적어도 하나가 선택되어지고 관련 정보가 등록된다(B07). 상기 선택된 이미지 및 관련정보가 웹서버(100)로 전송된다(B08).First, a PC or mobile terminal 100 connected to the Internet network accesses a web server 200 that provides a service using image recognition technology, uploads and searches a specific image (B01). The web server 200 transmits the received image to the image recognition server 300 (B02). The image recognition server 300 analyzes characteristic pattern values of the transmitted image (B03). The image recognition server 300 compares the pattern value of the transmitted image with a pre-registered image pattern value (B04). The image recognition server 300 transmits at least one or more image information corresponding to the same or most similar pattern value to the web server 200 (B05). The web server 200 transmits the received image information to the PC or the mobile terminal 100 and appears on the screen of the PC or the mobile terminal 100 (B06). At least one of the image or images displayed on the screen of the PC or the mobile terminal 100 and the image used for the search is selected and related information is registered (B07). The selected image and related information are transmitted to the web server 100 (B08).
상기 기 등록된 이미지 패턴값과 비교하는 이미지의 비교값(Query Image Score)이 일정 수준 이상이면 동일한 이미지로 간주되어 자동으로 등록되는 단계가 추가될 수 있다.If a comparison value (Query Image Score) of the image comparing with the previously registered image pattern value is a predetermined level or more may be added to the step is automatically considered to be the same image.
또는 위 단계 대신 기등록된 이미지와 동일한 이미지를 검색하여 나타난 비교 이미지 비교값(Query Image Score)을 기준값(Resxult Image Score)으로 저장하거나 일정 수준 이상의 기준값을 저장하는 단계와, 새로 검색되는 비교 이미지의 비교값이 상기 기준값과 비교하여 동일 또는 일정 비교율 이상이면 동일이미지로 간주하여 자동으로 등록되는 단계가 추가될 수 있다.Alternatively, the step of storing the comparison image score (Resxult Image Score), which is obtained by searching for the same image as the registered image instead of the above step, as the reference image score or storing a reference value of a predetermined level or more, and If the comparison value is equal to or higher than the reference value or more than a predetermined comparison rate, the step of automatically registering the same image may be added.
또한, 상기 실시 예에서는 PC 또는 모바일단말(100)로부터 전송되는 이미지는 웹서버(200)를 거처 이미지인식서버(300)로 전송되는 구조로 되어 있지만, PC 또는 모바일단말(100)에서 직접 이미지인식서버(300)로 전송하거나 그 결과를 이미지인식서버(300)에서 PC 또는 모바일단말(100)로 전송된 뒤 웹서버(200)로 전송하는 구조로 구성할 수도 있다.In addition, in the above embodiment, the image transmitted from the PC or mobile terminal 100 has a structure that is transmitted to the image recognition server 300 via the web server 200, but the image recognition directly from the PC or mobile terminal 100 The server 300 may be configured to transmit or transmit the result to the web server 200 after the result is transmitted from the image recognition server 300 to the PC or the mobile terminal 100.
또, 전송되는 이미지의 특징적 패턴값을 분석하는 것은 이미지인식서버(300)로 이미지가 전송되어 이미지인식서버(300)에서 이를 분석할 수도 있지만, PC 또는 모바일단말(100)이나 별도의 서버에 이미지의 패턴값을 분석하는 로직이 들어있어 이들이 이미지의 패턴값을 분석하여 이미지인식서버(300)로 전송하는 구조로 만들 수도 있다.In addition, analyzing the characteristic pattern value of the transmitted image may be transmitted to the image recognition server 300 and the image recognition server 300, but it may be analyzed, but the image on the PC or mobile terminal 100 or a separate server Contains a logic to analyze the pattern value of the they may be made to the structure to transmit the pattern value of the image to the image recognition server 300.
도 7은 본 발명의 실시예에 따라서 다수의 정보제공자가 등록한 관련정보를 이미지 인식기술을 이용하여 정보하는 제공하는 흐름도이다.FIG. 7 is a flowchart for providing related information registered by a plurality of information providers using image recognition technology according to an exemplary embodiment of the present invention.
먼저, 모바일단말(100)의 이미지인식기술을 활용한 서비스를 제공하는 앱이 구동되어 특정 이미지가 촬영된다(C01). 전술한 바와 같이, 이 이미지는 모바일단말에 설치된 APP 또는 내부 카메라 기능을 통해서 촬영된 이미지일 수도 있고 인식된 이미지일 수도 있고, 별도로 다운, 스캔 등의 방식으로 확보하고 있는 이미지일 수도 있다. 이후, 해당 이미지는 이미지인식서버(300)로 전송된다(C02). 이미지인식서버(300)는 상기 전송된 이미지의 특징적 패턴값을 분석한다(C03). 다음으로, 상기 이미지인식서버(300)는 상기 전송된 이미지의 패턴값과 기 등록된 이미지 패턴값을 비교한다(C04). 동일 또는 가장 유사한 패턴값에 해당하는 이미지 정보를 적어도 하나 이상 상기 모바일단말(100)로 전송한다(C05). First, an app that provides a service using the image recognition technology of the mobile terminal 100 is driven to capture a specific image (C01). As described above, this image may be an image taken through an APP or an internal camera function installed in a mobile terminal, may be a recognized image, or may be an image secured by a method such as down, scan, or the like. Thereafter, the image is transmitted to the image recognition server 300 (C02). The image recognition server 300 analyzes characteristic pattern values of the transmitted image (C03). Next, the image recognition server 300 compares the pattern value of the transmitted image and the registered image pattern value (C04). At least one image information corresponding to the same or most similar pattern value is transmitted to the mobile terminal 100 (C05).
모바일단말(100)은 상기 전송받은 적어도 하나 이상의 이미지 정보를 웹서버로 전송한다(C06). 웹서버(200)는 DB서버를 통해 상기 적어도 하나 이상의 이미지 정보에 적어도 하나 이상의 정보제공자가 각각 등록한 관련정보들을 검색한다(C07). 상기 웹서버(200)는 상기 적어도 하나 이상의 이미지 정보에 상기 각각 검색된 적어도 하나 이상의 관련정보를 상기 모바일단말(100)로 전송한다(C08). 상기 검색된 관련정보가 다수개인 경우, 상기 모바일기기의 화면에 상기 검색된 다수개의 관련정보 리스트가 화면에 나타낸다(C09). 모바일 화면에 홍보물 정보 또는 쇼핑몰 정보가 표시될 수 있다. 특히, 쇼핑몰 정보는 쇼핑몰 UI(아이콘)을 선택하면 1개 또는 다수개의 쇼핑몰 정보가 나타나, 그 중 사용자가 원하는 쇼핑몰을 선택하도록 구현하는 것도 가능하다.The mobile terminal 100 transmits the received at least one image information to the web server (C06). The web server 200 searches for relevant information registered by at least one or more information providers on the at least one or more image information through a DB server (C07). The web server 200 transmits the at least one or more related information respectively retrieved to the at least one or more image information to the mobile terminal 100 (C08). When there are a plurality of searched related information, the searched list of a plurality of related information is displayed on the screen of the mobile device (C09). Promotional information or shopping mall information may be displayed on the mobile screen. In particular, when the shopping mall UI (icon) is selected, one or more shopping mall information appears, and the shopping mall information may be implemented to select a shopping mall desired by the user.
상기 실시 예에서 PC 또는 모바일기기로부터 전송되는 이미지는 이미지인식서버로 전송되는 구조로 되어 있지만, 모바일기기에서 웹서버를 거처 이미지인식서버로 전송되고 그 결과값인 이미지 정보가 이미지인식서버에서 웹서버로 전송되고, 웹서버는 상기 이미지 정보에 해당하는 다수의 정보제공자가 각각 등록한 관련정보를 검색하여 상기 모바일기기로 전송하는 구조로 구성할 수도 있다.In the above embodiment, the image transmitted from the PC or the mobile device is transmitted to the image recognition server, but the mobile device is transmitted to the image recognition server via a web server, and the resulting image information is transmitted from the image recognition server to the web server. The web server may be configured to search for relevant information registered by a plurality of information providers corresponding to the image information and transmit the related information to the mobile device.
또한, 전송되는 이미지의 특징적 패턴값을 분석하는 것은 이미지인식서버로 이미지가 전송되어 이미지인식서버에서 이를 분석할 수도 있지만, 상기 모바일기기나 별도의 서버에서 이미지의 패턴값을 분석하는 로직이 들어있어 이곳에서 이미지의 패턴값을 분석하여 이미지인식서버로 전송하는 구조로 만들 수도 있다.In addition, analyzing the characteristic pattern value of the transmitted image may be transmitted to the image recognition server and the image recognition server may be analyzed, but the mobile device or a separate server contains the logic to analyze the pattern value of the image You can also make a structure that analyzes the pattern value of the image and sends it to the image recognition server.
다른 변형으로는, 상술한 실시 예에서 정보제공자가 제공하는 정보유형에 따라 홍보와 쇼핑으로 분류하는 단계를 추가하여 제공할 수 있다. 상기 이미지정보에 해당하는 관련정보 제공 시 최초에는 홍보와 관련된 정보만을 제공하고, 상기 모바일기기 화면에서 쇼핑 관련정보의 UI(아이콘)가 선택되어지도록 하며, 모바일기기는 상기 정보를 웹서버로 전송하고, 웹서버는 상기 DB서버를 통해 적어도 하나 이상의 정보제공자가 각각 등록한 쇼핑 관련정보들을 검색하며, 웹서버는 검색된 적어도 하나 이상의 쇼핑 관련정보를 상기 모바일기기로 전송한다. 상기 검색된 관련정보가 다수개인 경우, 상기 모바일기기의 화면에 상기 검색된 다수개의 관련정보 리스트가 화면에 나타나는 단계가 추가로 제공될 수 있다. In another variation, the above-described embodiment may be provided by additionally classifying into promotion and shopping according to the information type provided by the information provider. When providing the relevant information corresponding to the image information, only the information related to the promotion is initially provided, and the UI (icon) of shopping related information is selected on the screen of the mobile device, and the mobile device transmits the information to a web server. The web server searches for shopping related information registered by at least one information provider through the DB server, and the web server transmits the searched at least one shopping related information to the mobile device. When there are a plurality of searched related information, a step of displaying the searched plurality of related information lists on the screen of the mobile device may be further provided.
전술한 본 발명은 바람직한 실시예에 대하여 설명하였지만, 본 발명은 이에 한정되는 것이 아니고 특허청구범위와 발명의 상세한 설명 및 첨부한 도면의 범위 안에서 여러 가지로 변형하여 실시하는 것이 가능하고 이 또한 본 발명에 속한다.Although the above-described present invention has been described with respect to the preferred embodiments, the present invention is not limited thereto, and it is possible to carry out various modifications within the scope of the claims and the detailed description of the invention and the appended drawings, and this invention also. Belongs to.

Claims (8)

  1. 모바일 단말과 네트워크를 통해 접속되며, 이미지인식서버, 웹서버, 및 DB서버를 포함하여 구성된 이미지인식기술을 이용한 정보제공방법에 있어서,In the information providing method using an image recognition technology connected to a mobile terminal and a network, comprising an image recognition server, a web server, and a DB server,
    상기 모바일단말에서 획득한 이미지 또는 획득한 이미지로부터 추출한 이미지 패턴값을 상기 이미지인식서버로 전송하는 단계;Transmitting the image pattern value extracted from the image acquired from the mobile terminal or the acquired image to the image recognition server;
    상기 이미지인식서버에서 상기 모바일단말로부터 이미지를 전송받았으면 이미지 패턴값을 추출하고, 이미지 패턴값을 직접 전송받았으면 기등록된 이미지 패턴값과 유사한 이미지 패턴값을 검색하는 단계;Extracting an image pattern value when the image recognition server receives an image from the mobile terminal, and retrieving an image pattern value similar to a registered image pattern value when the image pattern value is directly transmitted;
    상기 이미지인식서버는 검색 결과값인 이미지코드값을 상기 모바일단말을 거쳐 또는 직접 웹서버로 전송하는 단계;The image recognition server transmitting an image code value, which is a search result value, through the mobile terminal or directly to a web server;
    상기 웹서버는 전송받은 이미지코드값에 등록된 홍보용코드값과/또는 쇼핑용코드값을 찾아내고, 홍보용코드값에 등록된 홍보물 정보 또는 쇼핑몰코드값에 등록된 쇼핑몰 URL정보 등을 모바일단말로 전송하는 단계를 포함하는 이미지인식기술을 이용한 정보제공방법.The web server finds a promotional code value and / or a shopping code value registered in the received image code value, and transmits promotional material information registered in the promotional code value or shopping mall URL information registered in the shopping mall code value to a mobile terminal. Information providing method using an image recognition technology comprising the step of.
  2. 제1 항에 있어서,According to claim 1,
    상기 이미지와 해당 이미지코드값은 사전에 DB서버에 등록되며,The image and the corresponding image code value are registered in advance in the DB server,
    등록 방법은 기존에 등록되지 않은 새로운 이미지를 등록하는 방법 또는 기등록된 이미지에서 원하는 이미지를 찾아 해당 이미지에 자신이 원하는 정보를 추가로 등록하는 이미지인식기술을 이용한 정보제공방법.The registration method is a method of registering a new image that has not been previously registered, or an information providing method using image recognition technology that finds a desired image in a pre-registered image and adds the desired information to the image.
  3. 제2 항에 있어서,The method of claim 2,
    상기 기 등록된 이미지에서 원하는 이미지를 찾는 방법은,How to find the desired image from the registered image,
    사전에 이미지를 등록할 때, 등록되는 이미지에 해당되는 키워드도 함께 등록하도록 하여 키워드 검색을 통해 찾는 방법, 또는 이미지 검색을 통해 동일 또는 가장 유사한 이미지를 검색하여 선택하거나 아니면 검색에 사용한 이미지를 등록하는 방법, 또는 기등록된 패턴값과 동일하거나 등록하려는 이미지와 기등록된 이미지의 패턴값의 오차가 극히 적은 것을 동일이미지로 간주할 수도 있는 방법을 이용하는 이미지인식기술을 이용한 정보제공방법.When registering an image in advance, the keyword corresponding to the registered image is also registered so as to find it through keyword search, or by searching for and selecting the same or most similar image through image search, or registering the image used in the search. An information providing method using an image recognition technology using a method or a method in which an error of the pattern value of the registered image and the registered image and the image to be registered and the registered image is extremely small may be regarded as the same image.
  4. 제2 항에 있어서,The method of claim 2,
    상기 이미지인식서버에 기존에 등록되지 않은 새로운 이미지를 등록하는 방법은 기 등록된 이미지의 기준값을 추출하고 그 기준값과 신규 이미지의 기준값, 신규 이미지검색을 통해 기 등록된 이미지와의 비교값이 모두 오차범위 이내일 때 동일이미지로 결정하는 이미지인식기술을 이용한 정보제공방법.In the method of registering a new image not previously registered in the image recognition server, the reference value of the previously registered image is extracted, and the reference value, the reference value of the new image, and the comparison value between the previously registered image through the new image search are all error. Information providing method using image recognition technology to determine the same image within the range.
  5. 모바일 단말과 네트워크를 통해 접속된 이미지인식기술을 이용한 정보제공시스템에 있어서,In the information providing system using the image recognition technology connected via a mobile terminal and a network,
    전달받은 이미지에 대해서 유사한 이미지를 검색하고, 검색된 유사 이미지에 대응된 이미지코드값을 검색하는 이미지인식서버; 및An image recognition server searching for a similar image with respect to the received image and searching for an image code value corresponding to the found similar image; And
    상기 이미지인식서버에 접속된 DB서버로서, 하나의 이미지코드값에 다수의 정보를 제공하기 위하여, 이미지에 대응되는 이미지코드값에 다수의 홍보 및/또는 쇼핑 관련 정보를 저장하는 DB서버;를 포함하되,A DB server connected to the image recognition server, the DB server storing a plurality of promotional and / or shopping related information in an image code value corresponding to an image in order to provide a plurality of information in one image code value; But
    상기 모바일 단말이 획득한 이미지를 통해 정보제공을 원하는 경우, 해당 이미지에 대한 다수의 홍보 및/또는 쇼핑 관련 정보 중에서 선택하여 연결할 수 있도록 구성된 정보제공시스템.When providing information through the image obtained by the mobile terminal, an information providing system configured to select and connect from a plurality of promotional and / or shopping-related information for the image.
  6. 제5 항에 있어서,The method of claim 5,
    상기 이미지와 해당 정보는 사전에 DB서버에 등록되며,The image and the corresponding information are registered in advance in the DB server,
    등록 방법은 기존에 등록되지 않은 새로운 이미지를 등록하는 방법 또는 기등록된 이미지에서 원하는 이미지를 찾아 해당 이미지에 자신이 원하는 정보를 추가로 등록하는 정보제공시스템.The registration method is a method of registering a new image that has not been previously registered or an information providing system that finds a desired image in a pre-registered image and additionally registers desired information in the image.
  7. 제6 항에 있어서,The method of claim 6,
    상기 기 등록된 이미지에서 원하는 이미지를 찾는 방법은,How to find the desired image from the registered image,
    사전에 이미지를 등록할 때, 등록되는 이미지에 해당되는 키워드도 함께 등록하도록 하여 키워드 검색을 통해 찾는 방법, 또는 이미지 검색을 통해 동일 또는 가장 유사한 이미지를 검색하여 선택하거나 아니면 검색에 사용한 이미지를 등록하는 방법, 또는 기등록된 패턴값과 동일하거나 등록하려는 이미지와 기등록된 이미지의 패턴값의 오차가 극히 적은 것을 동일이미지로 간주할 수도 있는 방법을 이용하는 정보제공시스템.When registering an image in advance, the keyword corresponding to the registered image is also registered so as to find it through keyword search, or by searching for and selecting the same or most similar image through image search, or registering the image used in the search. An information providing system using a method or a method in which an error of the pattern value of an image to be registered and a registered image that is identical to or registered with an already registered pattern is extremely small can be regarded as the same image.
  8. 제5 항에 있어서,The method of claim 5,
    웹서버를 더 포함하고,Further includes a web server,
    상기 이미지인식서버는 유사한 이미지를 찾아 유사 이미지에 대한 정보인 이미지코드값을 상기 웹서버에 전달하고, 상기 웹서버는 상기 DB서버에서 상기 이미지코드값에 매칭된 홍보용코드값 또는 쇼핑몰코드값을 검색하는 정보제공시스템.The image recognition server searches for a similar image and transmits an image code value, which is information about a similar image, to the web server, and the web server searches for a promotional code value or a shopping mall code value matching the image code value in the DB server. Information provision system.
PCT/KR2016/005283 2015-05-18 2016-05-18 Information providing system and method using image recognition technology WO2016186457A1 (en)

Applications Claiming Priority (6)

Application Number Priority Date Filing Date Title
KR20150068778 2015-05-18
KR10-2015-0068778 2015-05-18
KR10-2015-0137234 2015-09-30
KR20150137234 2015-09-30
KR1020160001532A KR20160135641A (en) 2015-05-18 2016-01-06 Method of providing information using image-recognition technology
KR10-2016-0001532 2016-01-06

Publications (1)

Publication Number Publication Date
WO2016186457A1 true WO2016186457A1 (en) 2016-11-24

Family

ID=57320585

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/KR2016/005283 WO2016186457A1 (en) 2015-05-18 2016-05-18 Information providing system and method using image recognition technology

Country Status (1)

Country Link
WO (1) WO2016186457A1 (en)

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20030043920A1 (en) * 2001-05-09 2003-03-06 Kozo Akiyoshi Image processing method
WO2005091211A1 (en) * 2004-03-16 2005-09-29 3Vr Security, Inc. Interactive system for recognition analysis of multiple streams of video
US20060227992A1 (en) * 2005-04-08 2006-10-12 Rathus Spencer A System and method for accessing electronic data via an image search engine
US20070159522A1 (en) * 2004-02-20 2007-07-12 Harmut Neven Image-based contextual advertisement method and branded barcodes
KR20120011347A (en) * 2010-07-21 2012-02-08 주식회사 비즈모델라인 Method for Setting Order Value of Store Group, Smart Phone and Program

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20030043920A1 (en) * 2001-05-09 2003-03-06 Kozo Akiyoshi Image processing method
US20070159522A1 (en) * 2004-02-20 2007-07-12 Harmut Neven Image-based contextual advertisement method and branded barcodes
WO2005091211A1 (en) * 2004-03-16 2005-09-29 3Vr Security, Inc. Interactive system for recognition analysis of multiple streams of video
US20060227992A1 (en) * 2005-04-08 2006-10-12 Rathus Spencer A System and method for accessing electronic data via an image search engine
KR20120011347A (en) * 2010-07-21 2012-02-08 주식회사 비즈모델라인 Method for Setting Order Value of Store Group, Smart Phone and Program

Similar Documents

Publication Publication Date Title
WO2016186289A1 (en) System and method for providing information using image recognition technique
WO2015005717A1 (en) User interest-based product information recommendation system
WO2018174637A1 (en) Real time shopping method using video recognition in broadcast, and smart device in which application for implementing same is installed
WO2012020927A1 (en) Integrated image search system and a service method therewith
WO2012108708A2 (en) Local area advertisement system and method for calculating a local area advertisement cost
WO2020085786A1 (en) Style recommendation method, device and computer program
WO2017115892A1 (en) User-customized advertisement providing system through public display and advertisement providing method using same
WO2020251174A1 (en) Method for advertising user-customized fashion item and server executing same
WO2017003015A1 (en) Internet advertisement provision device and method
WO2018043990A1 (en) Method, device and computer program for providing image search information
WO2018225939A1 (en) Method, device, and computer program for providing image-based advertisement
WO2017099364A1 (en) Method, device, and computer program for managing shopping information
WO2014010916A1 (en) System and method for intermediating electronic commerce using offline transaction information
WO2010064849A2 (en) Method, system and computer readable recording medium for posting advertisement in a region with high user preference
US20100131358A1 (en) Information management method, information management device, information management program, and call receiving device
WO2017111287A1 (en) Method, apparatus and computer program for providing commercial contents
WO2018038437A1 (en) Method for providing product information using video and management server used therefor
WO2018043923A1 (en) Display device and control method therefor
WO2013019018A2 (en) System for playing video files and providing video information services
WO2020242089A2 (en) Artificial intelligence-based curating method and device for performing same method
WO2016186457A1 (en) Information providing system and method using image recognition technology
KR101966596B1 (en) Contents processing type marketing system based on SNS, and method thereof
WO2018221988A1 (en) Content information providing method and system
JP7110738B2 (en) Information processing device, program and information processing system
WO2017222226A1 (en) Method for registering advertised product on image content and server for executing same

Legal Events

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

Ref document number: 16796784

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 16796784

Country of ref document: EP

Kind code of ref document: A1