[go: up one dir, main page]

WO2022119136A1 - Procédé et système d'extraction d'informations d'étiquette à partir d'une image de capture d'écran - Google Patents

Procédé et système d'extraction d'informations d'étiquette à partir d'une image de capture d'écran Download PDF

Info

Publication number
WO2022119136A1
WO2022119136A1 PCT/KR2021/015463 KR2021015463W WO2022119136A1 WO 2022119136 A1 WO2022119136 A1 WO 2022119136A1 KR 2021015463 W KR2021015463 W KR 2021015463W WO 2022119136 A1 WO2022119136 A1 WO 2022119136A1
Authority
WO
WIPO (PCT)
Prior art keywords
text
information
area
screenshot image
screenshot
Prior art date
Legal status (The legal status 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 status listed.)
Ceased
Application number
PCT/KR2021/015463
Other languages
English (en)
Korean (ko)
Inventor
정구일
홍건표
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Miner Inc
Original Assignee
Miner Inc
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
Application filed by Miner Inc filed Critical Miner Inc
Publication of WO2022119136A1 publication Critical patent/WO2022119136A1/fr
Anticipated expiration legal-status Critical
Ceased legal-status Critical Current

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/50—Information retrieval; Database structures therefor; File system structures therefor of still image data
    • G06F16/58—Retrieval characterised by using metadata, e.g. metadata not derived from the content or metadata generated manually
    • G06F16/583—Retrieval characterised by using metadata, e.g. metadata not derived from the content or metadata generated manually using metadata automatically derived from the content
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/50—Information retrieval; Database structures therefor; File system structures therefor of still image data
    • G06F16/58—Retrieval characterised by using metadata, e.g. metadata not derived from the content or metadata generated manually
    • G06F16/583—Retrieval characterised by using metadata, e.g. metadata not derived from the content or metadata generated manually using metadata automatically derived from the content
    • G06F16/5846—Retrieval characterised by using metadata, e.g. metadata not derived from the content or metadata generated manually using metadata automatically derived from the content using extracted text
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F40/00—Handling natural language data
    • G06F40/10—Text processing
    • G06F40/103—Formatting, i.e. changing of presentation of documents
    • G06F40/117—Tagging; Marking up; Designating a block; Setting of attributes
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/08—Learning methods
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06Q—INFORMATION 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
    • G06Q10/00—Administration; Management
    • G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06Q—INFORMATION 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
    • G06Q30/00—Commerce
    • G06Q30/02—Marketing; Price estimation or determination; Fundraising
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06Q—INFORMATION 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
    • G06Q30/00—Commerce
    • G06Q30/02—Marketing; Price estimation or determination; Fundraising
    • G06Q30/0241—Advertisements
    • G06Q30/0251—Targeted advertisements
    • G06Q30/0255—Targeted advertisements based on user history
    • G06Q30/0256—User search
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06Q—INFORMATION 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
    • G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
    • G06Q50/10—Services
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V30/00—Character recognition; Recognising digital ink; Document-oriented image-based pattern recognition
    • G06V30/40—Document-oriented image-based pattern recognition
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N5/00—Details of television systems
    • H04N5/76—Television signal recording
    • H04N5/765—Interface circuits between an apparatus for recording and another apparatus
    • H04N5/77—Interface circuits between an apparatus for recording and another apparatus between a recording apparatus and a television camera

Definitions

  • One object of the present invention is to provide a method of extracting tag information from a screenshot image.
  • An object of the present invention is to provide a method of searching for a screenshot image based on tag information mapped to the screenshot image.
  • One object of the present invention relates to a method of providing a screenshot image search result based on tag information mapped to a screenshot image.
  • a system for extracting tag information from a screenshot image disclosed in the present application includes: a user terminal in which at least one screenshot image is stored; and a server for transmitting and receiving information about the screenshot image with the user terminal, wherein the screenshot image includes at least one text, and the user terminal includes text in which the text exists in the screenshot image.
  • the server After obtaining an area and a non-text area in which the text does not exist, extracting text information from the text area through a character reading module, mapping the text information to an identification number assigned to the screenshot image, and then the server After receiving the text information from the user terminal, the server extracts a keyword from the text information, maps the tag information derived based on the keyword to the identification number, and transmits it to the user terminal can
  • tag information is mapped to the screenshot image captured and stored by the user, the user can quickly select a necessary screenshot image from among numerous screenshot images stored in the terminal based on the mapped tag information. and search accurately.
  • FIG. 2 is a diagram for explaining the configuration of a user terminal.
  • FIG. 3 is a diagram illustrating various images stored in a storage unit provided in a user terminal by way of example.
  • FIG. 5 is a diagram for explaining an overall process of a method of extracting tag information from a screenshot image according to an embodiment.
  • FIG. 6 is a diagram for explaining the classification of a screenshot image into a text area or a non-text area through a neural network model.
  • 9 to 11 are diagrams for exemplarily explaining text information extracted from a text area.
  • FIG. 12 is a diagram for explaining a method for extracting image information from a non-text area.
  • FIG. 13 is a diagram for explaining the entire process of extracting a keyword from text information in a server.
  • FIG. 14 is a diagram for explaining a method of acquiring a keyword from text information through a second neural network model according to an embodiment.
  • 15 is a diagram for exemplarily explaining keywords and target information extracted from a screenshot image.
  • 16 is a view for explaining a screen displayed through an output unit of a user terminal by way of example.
  • 17 is a diagram for explaining a method of selecting a recommendation tag and providing it to a user.
  • a system for extracting tag information from a screenshot image includes: a user terminal in which at least one screenshot image is stored; and a server for transmitting and receiving information about the screenshot image with the user terminal, wherein the screenshot image includes at least one text, and the user terminal includes text in which the text exists in the screenshot image.
  • extracting text information from the text area through a character reading module, mapping the text information to the screenshot image and sending it to the server may receive the text information from the user terminal, extract a keyword from the text information, map tag information derived based on the keyword to the screenshot image, and then transmit it to the user terminal.
  • the text area includes a first text area, an area in which text having a first size within an error range, and a second text area, an area in which text having a second size within an error range, among texts in the screenshot image.
  • the first text area is an area of text adjacent within a certain range among texts having the first font in the screenshot image
  • the second text area is the second font in the screenshot image. It may be a text area that is adjacent within a certain range among texts having .
  • the text information includes first text information and second text information
  • the user terminal extracts the first text information from the first text area through the character reading module, and the character from the second text area After extracting the second text information through a reading module, the first text information and the second text information are extracted, wherein the first text area and the second text area can be independently input to the character reading module have.
  • the user terminal obtains the text area from the screenshot image by using a pre-trained neural network model, wherein the pre-trained neural network model divides the text area into the first text area and the second text area to the Nth text area. It can be learned to distinguish text areas.
  • the keywords extracted from the text information may be plural, and the keywords may be words, numbers, sentences, or a combination thereof representing the screenshot image among texts included in the text information.
  • the user terminal receives the first tag information from the server at a first time and receives the second tag information at a second time, and the user terminal determines the first tag information as the recommendation tag;
  • the first time point may be earlier than the second time point.
  • the user terminal may receive a plurality of tag information extracted from the plurality of screenshot images from the server, and the user terminal may determine the recommendation tag based on a frequency at which the plurality of tag information is received from the server. .
  • the user terminal acquires a first text area in which text exists in the first screenshot image, extracts the first text information from the first text area through the character reading module, and the second screen acquire a second text area in which text exists in the shot image, and extract the second text information from the second text area through the character reading module, wherein the first text area and the second text area are It can be independently input to the character reading module.
  • the first text area is an area in which text exists in the first screenshot image, and includes a text area having a first property and a text area having a second property to a text zero having an Nth property,
  • the first to Nth properties may be determined based on at least one of a size and a font of the text.
  • the second text area is an area in which text exists in the second screenshot image, and includes a text area having a first property and a text area having a second property to a text zero having an Nth property,
  • the first to Nth properties may be determined based on at least one of a size and a font of the text.
  • the first text area is an area of text adjacent within a certain range among texts in the first screenshot image
  • the second text area is text adjacent within a certain range of texts in the second screenshot image. may be the area of
  • the server After receiving the first text information from the user terminal, the server extracts a first keyword from the first text information, extracts the first tag information based on the first keyword, and the user terminal After receiving the second text information from , a second keyword may be extracted from the second text information, and the second tag information may be extracted based on the second keyword.
  • the first keyword is at least one word, number, sentence, or a combination thereof representing the first screenshot image among texts included in the first text information
  • the second keyword is included in the second text information. It may be at least one word, number, sentence, or a combination thereof representing the second screenshot image among the included text.
  • the first tag information may include a first representative image determined based on the first keyword
  • the second tag information may include a second representative image determined based on the second keyword
  • Importance is reflected in the first keyword, and the importance is the first screenshot image among words, numbers, sentences, or combinations thereof representing the first screenshot image among texts included in the first text information. is determined based on a probability of representing Among these combinations, it may be determined based on a probability of representing the second screenshot image.
  • the user input includes a first user input and a second user input
  • the input unit receives the first user input at a first time point and inputs the second user input at a second time point different from the first time point.
  • the control unit may determine the matching tag information in consideration of a time point at which the first user input and the second user input are input to the input unit.
  • the control unit determines the matching tag information based on the first user input and the second user input, and determines the matching tag information by giving importance to the first user input, wherein the first time point is earlier than the second time point can
  • the matching tag information may be tag information corresponding to at least one of the first user input and the second user input among the plurality of tag information.
  • the matching tag information may be tag information corresponding to the second user input from among the plurality of tag information corresponding to the first user input.
  • the control unit may control the screen shot image to which the matching tag information is mapped to be output through the display unit, and control so that the screenshot image is displayed on the display unit according to a time sequence stored in the storage unit.
  • FIG. 1 is a diagram for explaining a method of extracting tag information from a screenshot image according to an embodiment
  • FIG. 2 is a diagram for explaining the configuration of a user terminal.
  • a method of extracting tag information from a screenshot image may be performed through a user terminal 1000 and a server 2000 .
  • the user terminal 1000 includes a control unit 100, an image capturing unit 200, a storage unit 300, a user input unit 400, an output unit 500, a power supply unit 600, and a communication unit ( 700) may be included.
  • the user terminal 1000 may include a portable information and communication device, for example, a smart phone or a tablet.
  • the image capturing unit 200 is a digital camera and may include an image sensor and an image processing unit.
  • An image sensor is a device that converts an optical image into an electrical signal, and may be configured as a chip in which a plurality of photodiodes are integrated.
  • the image sensor may include a charge coupled device (CCD), a complementary metal oxide semiconductor (CMOS), or the like.
  • the image processing unit may generate image information by image processing the captured result.
  • the storage unit 300 is a storage means for storing data that can be read by a microprocessor, and includes a hard disk drive (HDD), a solid state disk (SSD), a silicon disk drive (SDD), a ROM, a RAM, CD-ROMs, magnetic tapes, floppy disks, and optical data storage devices may be included.
  • HDD hard disk drive
  • SSD solid state disk
  • SDD silicon disk drive
  • ROM read only memory
  • RAM random access memory
  • data received by the user terminal 1000 may be stored in the storage unit 300 .
  • the storage unit 300 may store an image directly captured by the user through the image capturing unit 200 , and a screenshot image obtained by a user capturing information obtained online and output on the output unit 500 . can be stored.
  • the user input unit 400 receives a user input for the user terminal 1000 .
  • the received input may be transmitted to the controller 100 .
  • the user input unit 400 may receive a user input through a touch display.
  • the user input unit 400 may refer to a user interface screen on which a command is input from a user.
  • the output unit 500 outputs various types of information according to a control command of the control unit 100 .
  • the output unit 500 may output information through the display panel. More specifically, the output unit 500 may output information related to the user's hair loss state through the display panel.
  • the output unit 500 is not limited to a display panel, and may include various means for outputting information, such as a speaker.
  • the power supply unit 600 includes a battery, and the battery may be installed in the user terminal 1000 or may be detachably provided from the outside.
  • the power supply unit 600 may supply power required by each component of the user terminal 1000 .
  • the communication unit 700 may include a wireless communication module and/or a wired communication module.
  • the wireless communication module may include a Wi-Fi communication module, a cellular communication module, and the like.
  • the controller 100 may include at least one processor.
  • each processor may execute a predetermined operation by executing at least one instruction stored in the memory.
  • the controller 100 may control the overall operation of components included in the user terminal 1000 .
  • the user terminal 1000 may be controlled or operated by the controller 100 .
  • the user may perform a web search through the user terminal 1000 and obtain necessary information through the search contents. If necessary, the user may store the acquired information in the terminal, and in this case, the screen shot function provided in the user terminal 1000 may be utilized.
  • the storage unit 300 may include a plurality of images (hereinafter, referred to as captured images) stored by the user directly photographing an object in addition to the screenshot image captured by the user. At this time, the storage unit 300 may store the captured image and the screenshot image accumulated.
  • FIG. 3 is a diagram illustrating various images stored in a storage unit provided in a user terminal by way of example.
  • a screenshot image captured by a user and a photographed image may be stored in a single storage unit 300 .
  • the screenshot image and the captured image may be stored after being classified in different locations of the storage unit 300 .
  • only a screenshot image may be separated from among a plurality of images stored in the storage unit 300 , or only a desired screenshot image may be searched from among a plurality of screenshot images stored in the storage unit 300 . can do.
  • FIG. 4 is a diagram for explaining a screenshot image by way of example.
  • a screenshot image will be described by way of example with reference to FIG. 4 .
  • a plurality of screenshot images may be stored in the storage unit 300 .
  • the screenshot image may mean an image captured by the user of the screen displayed on the output unit 500 of the user terminal 1000 .
  • the screenshot image may be an image in which a part of the web surfing screen (eg, shopping mall, news, etc.) is captured or an image in which a part of the execution screen (eg, gifticon image, etc.) of an application stored in the terminal is captured. have.
  • a part of the web surfing screen eg, shopping mall, news, etc.
  • a part of the execution screen eg, gifticon image, etc.
  • the captured screenshot image includes various information, and the information may be provided in the form of an image or text.
  • the screenshot image may include a plurality of image areas NTA1 and NTA2 and may include a plurality of text areas TA1 to TA8 .
  • the text included in the plurality of text areas TA1 to TA8 may be configured with the same font or may be configured with the same size.
  • texts included in the plurality of text areas TA1 to TA8 may be configured in different fonts or have different sizes.
  • texts included in the plurality of text areas TA1 to TA8 may have the same font but different sizes.
  • texts included in the plurality of text areas TA1 to TA8 may be configured in different fonts but have the same size.
  • the method of extracting tag information from a screenshot image may extract tag information based on a plurality of text areas TA1 to TA8 included in the screenshot image, which will be described in detail below. .
  • FIG. 5 is a diagram for explaining an overall process of a method of extracting tag information from a screenshot image according to an embodiment.
  • the method for extracting tag information from a screenshot image includes the steps of obtaining a screenshot image (S1100), analyzing the screenshot image (S1200), and converting the screenshot image into a plurality of Separating into regions (S1300), extracting text from a text region among a plurality of regions (S1400), mapping the extracted text with a screenshot image and storing it (S1500), sending the extracted text to a server It may include a step (S1600).
  • the step of obtaining a screenshot image includes at least one screenshot image or two or more screenshot images among a plurality of screenshot images captured by the user's action and stored in the storage unit 300 as described above. It may include the step of obtaining.
  • the analysis of the acquired screenshot image may be performed.
  • analyzing the screenshot image ( S1200 ) may include performing image pre-processing so that the obtained screenshot image is suitable for analysis.
  • image preprocessing may be performed so that the obtained screenshot image has a resolution, sharpness, brightness, saturation, etc. suitable for image analysis.
  • the screenshot image may be divided into a plurality of regions. Separating the screenshot image into a plurality of regions ( S1300 ) may include dividing at least some regions of the obtained screenshot image into a text region or a non-text region.
  • FIG. 6 is a diagram for explaining the classification of a screenshot image into a text area or a non-text area through a neural network model.
  • the obtained screenshot image SI may be divided into a text area TA or a non-text area NTA through the first neural network model NN1 .
  • the text area TA may include a first text area, a second text area, and the like, which are divided based on the size (or size) of the text.
  • the text area TA includes a first text area including texts having a first size among texts included in the screenshot image, and a second text area including texts having a second size. can do.
  • the text area TA may include a first text area, a second text area, etc. provided based on the size and font of the text.
  • the text area TA has a first size among texts included in the screenshot image and includes a first text area including texts corresponding to the first font and a second size corresponding to the second font.
  • a second text area including texts may be included.
  • the obtained screenshot image SI may be divided into a first text area having a first property, a second text area having a second property, an N-th text area having an N-th property, and a non-text area (NTA).
  • NTA non-text area
  • the obtained screenshot image SI may be divided into a first text area, a second text area to an N-th text area, and a non-text area through a pre-trained neural network model, but is not limited thereto.
  • the non-text area may mean an area composed of a non-text portion of content included in the screenshot image.
  • the non-text area NTA may mean an area including photos of objects or people included in the screenshot image, logos, trademarks, product photos, and the like.
  • the non-text area NTA may be formed in plurality to correspond to the number of non-text parts of the content included in the screenshot image. That is, the non-text area NTA may include a first non-text area, a second non-text area, and the like.
  • the controller 100 divides a screenshot image SI into a text area TA or a non-text area NTA through the first neural network model NN1 trained in advance.
  • the first neural network model NN1 may be trained to receive a screenshot image and obtain a text area TA and/or a non-text area NTA from the screenshot image.
  • the first neural network model NN1 may be trained to obtain a text area TA and a non-text area NTA based on the screenshot image.
  • the first neural network model NN1 may obtain an output value after receiving a screenshot image. Thereafter, the first neural network model NN1 may be trained by a method of updating the first neural network model NN1 based on an error value calculated in consideration of the difference between the output value and the labeling data.
  • the output value may include a first output value corresponding to the first labeling value and a second output value corresponding to the second labeling value.
  • the controller 100 may acquire a plurality of text areas TA by using the first neural network model NN1 .
  • the controller 100 may obtain the first text area and the second text area by using the first neural network model NN1 .
  • the controller 100 may obtain a text area (eg, a first text area) having a first size among texts included in the screenshot image by using the first neural network model NN1, A text area (eg, a second text area) having 2 sizes may be obtained.
  • a text area eg, a first text area
  • a text area eg, a second text area
  • the first text area may mean a text area having a first size among texts located within a predetermined range within the screenshot image
  • the second text area may be located within a predetermined range within the screenshot image. It may mean a text area having a second size among texts. More specifically, the first text area may mean an area of adjacent texts within a predetermined range among texts having the first size in the screenshot image. Similarly, the second text area may mean an area of adjacent texts within a predetermined range among texts having the second size in the screenshot image.
  • the controller 100 may obtain a text area (eg, a first text area) having a first font from among texts included in the screenshot image by using the first neural network model NN1 , and the second A text area having a font (eg, a second text area) may be acquired.
  • a text area eg, a first text area
  • the second A text area having a font eg, a second text area
  • the first text area may mean a text area having a first font among texts located within a predetermined range in the screenshot image
  • the second text area may be located within a predetermined range within the screenshot image. It may mean a text area having the second font among texts. More specifically, the first text area may mean an area of adjacent texts within a predetermined range among texts having the first font in the screenshot image. Similarly, the second text area may mean an area of adjacent texts within a predetermined range among texts having the second font in the screenshot image.
  • a text area having a first size or a text area having a first font may be plural.
  • there may be a plurality of text areas having a first size in the screenshot image and the plurality of text areas may exist at different positions in the screenshot image.
  • there may be a plurality of text areas having the first font in the screenshot image and the plurality of font areas may exist in different positions within the screenshot image.
  • the controller 100 uses a first neural network model NN1 to have a first text area TA1 having a first size among texts included in a screenshot image SI and a second size.
  • a second text area TA2 and a third text area TA3 having a third size may be obtained.
  • the text area having the first size may be a plurality (eg, the third text area TA3 , the fifth text area TA5 , and the sixth text area TA6 ), and the text area having the second size may be plural.
  • There may be a plurality of text areas eg, a seventh text area TA7 and an eighth text area TA8).
  • the controller 100 uses the first neural network model NN1 to use a first text area TA1 having a first font among texts included in the screenshot image SI, and a fourth text area having a second font. (TA4), a third text area TA3 having a third font may be obtained. In this case, there may be a plurality of text areas having the third font (eg, the third text area TA3 , the fifth text area TA5 , and the sixth text area TA6 ).
  • the text area may mean an area in which text is included in the acquired screenshot image.
  • the text area may mean a text-related image area in a screenshot image.
  • an area including text is defined as a text area, but it may also be defined as a text-related image area.
  • Text information included in the text area may be extracted through the step of extracting the text ( S1400 ), and the step of extracting the text ( S1400 ) may be performed on the user terminal 1000 . That is, the controller 100 of the user terminal 1000 may obtain a plurality of text areas from the screenshot image and extract text information from the plurality of text areas.
  • FIG. 7 and 8 are diagrams for explaining a method of extracting text from a text area.
  • the controller 100 may extract text information by analyzing the screenshot image.
  • the controller 100 may extract text information from the screenshot image through the character reading module TM.
  • the controller 100 may output text information as output data by using the entire screenshot image as input data of the character reading module TM.
  • the controller 100 may extract text information by analyzing a partial region of the screenshot image.
  • the controller 100 may extract text information by analyzing at least some of the plurality of text areas TA included in the screenshot image through the character reading module TM.
  • the controller 100 may output text information as output data by using at least some of the plurality of text areas TA included in the screenshot image as input data of the character reading module TM.
  • the data processing speed of the character reading module TM may be improved.
  • the character reading module TM may include a character recognition algorithm (program or function), and obtain a character image based on image data using the character recognition algorithm (program or function), and convert the acquired character image into pixels Characters can be recognized by analyzing units.
  • the character reading module TM may include an optical character reading algorithm (program or function), and using the optical character reading algorithm (program or function) to obtain data as a result of reading an image for a scanned character can do.
  • the character reading module (TM) may be a lightweight algorithm as a character reading algorithm.
  • the controller 100 may extract the first text information from the first text area TA1 through the character reading module TM.
  • the controller 100 may extract the second text information from the second text area TA2 through the character reading module TM. That is, each of the plurality of text areas may be individually input to the text reading module TM, and the text reading module TM may extract respective text information from each text area.
  • the controller 100 may extract text information from a text area having a first size among texts included in the screenshot image through the character reading module TM.
  • the controller 100 may output text information from a text area having a first font among texts included in the screenshot image through the character reading module TM.
  • 9 to 11 are diagrams for exemplarily explaining text information extracted from a text area.
  • text information may be extracted from a screenshot image captured by a user.
  • the user may capture the SNS screen displayed through the output unit 500 of the user terminal 1000 as shown in FIG. By acquiring an image, text information as shown in FIG. 9(b) may be extracted.
  • the controller 100 controls a plurality of text areas (eg, a first text area TA1 , a second text area TA2 , and a third text area from the obtained screenshot image). area TA3) may be obtained. Thereafter, the controller 100 may extract text information as shown in FIG. 9B from at least one of the plurality of acquired text areas by using the character reading module TM.
  • a plurality of text areas eg, a first text area TA1 , a second text area TA2 , and a third text area from the obtained screenshot image.
  • area TA3 may be obtained.
  • the controller 100 may extract text information as shown in FIG. 9B from at least one of the plurality of acquired text areas by using the character reading module TM.
  • the user may capture the messenger screen displayed through the output unit 500 of the user terminal 1000 as shown in FIG. It is possible to extract text information as shown in (b) of FIG.
  • the controller 100 may obtain a plurality of text areas (eg, the first text areas TA1 to TA6) from the obtained screenshot image. can Thereafter, the controller 100 may extract text information as shown in FIG. 10B from at least one of the plurality of acquired text areas by using the character reading module TM.
  • a plurality of text areas eg, the first text areas TA1 to TA6
  • the controller 100 may extract text information as shown in FIG. 10B from at least one of the plurality of acquired text areas by using the character reading module TM.
  • the user may capture a news screen displayed through the output unit 500 of the user terminal 1000 , and the controller 100 may control the captured image It is possible to extract text information as shown in (b) of FIG. 11 by obtaining .
  • the controller 100 may obtain a plurality of text areas (eg, first text areas TA1 to ninth text areas TA9) from the obtained screenshot image. can Thereafter, the controller 100 may extract text information as shown in FIG. 11B from at least one of the plurality of acquired text areas by using the character reading module TM.
  • a plurality of text areas eg, first text areas TA1 to ninth text areas TA9
  • the controller 100 may extract text information as shown in FIG. 11B from at least one of the plurality of acquired text areas by using the character reading module TM.
  • FIG. 12 is a diagram for explaining a method for extracting image information IN from a non-text area NTA.
  • the controller 100 may extract image information IN from a plurality of non-text areas NTA obtained from a screenshot image through the image reading module IM.
  • the step of extracting object information from the non-text area among the plurality of areas may be performed.
  • the step of extracting object information from the non-text area may be performed through a pre-trained neural network model.
  • the pre-trained neural network model may be a model trained to obtain object information based on an image.
  • the object information may be stored by being mapped to a screenshot image, and a user may classify and/or search for the screenshot image based on the object information mapped to and stored on the screenshot image.
  • the user may search for a desired screenshot image by using the stored object information and/or keywords mapped to the screenshot image.
  • the user may search for a desired screenshot image by inputting the content related to the stored object information mapped to the screenshot image as a search term.
  • the step of extracting the object information from the non-text area may be performed in the user terminal 1000, but is not limited thereto.
  • the step of extracting object information from the non-text area may be performed in the server S.
  • control unit 100 of the user terminal 1000 may transmit a non-text area (NTA) obtained from the screenshot image to the server (S), and the server (S) receives the received non-text area (NTA).
  • NTA non-text area
  • Object information (TI) may be obtained from an image reading module (IM) or a neural network model.
  • control unit 100 of the user terminal 1000 may transmit a screenshot image to the server S, and the server S may use an image reading module IM from the received screenshot image or an object through a neural network model.
  • Information TI can be obtained.
  • the control unit 100 may map the extracted text information with the screenshot image and then store the extracted text information with the screenshot image ( S1500 ), and may map the extracted text information with the screenshot image, and then store it in the storage unit 300 .
  • the controller 100 may map the extracted text information to the screenshot image.
  • the mapping of the extracted text information to the screenshot image may mean, for example, to display identification so that the control unit 100 or the server S can recognize that the text information is extracted from the screenshot image. have.
  • the controller 100 may transmit the extracted text to the server through the step of transmitting the extracted text to the server ( S1600 ).
  • the controller 100 may transmit the text information to the server in order to extract a keyword from the extracted text information.
  • the method for extracting a keyword from text information in the server includes the steps of receiving text information from a user terminal (S2100), analyzing text information to obtain a keyword (S2200), and taking a screenshot of the obtained keyword It may include a step of storing the image after mapping (S2300), and a step of transmitting the obtained keyword to the user terminal (S2400).
  • the server S may receive text information from the user terminal 1000 through the step S2100 of receiving the text information from the user terminal.
  • the text information may be information about text extracted from the screenshot image by the controller 100 of the user terminal 1000 as described above.
  • the server S may analyze the text information obtained through the user terminal 1000 through a step S2200 of analyzing the text information to obtain a keyword, and obtain a keyword based thereon.
  • FIG. 14 is a diagram for explaining a method of acquiring a keyword from text information through a second neural network model according to an embodiment.
  • the server S may obtain a keyword from text information through the second neural network model NN2.
  • the second neural network model NN2 may be trained to obtain keywords from text information.
  • the second neural network model NN2 may be trained to obtain keywords based on text information.
  • the server S may acquire important keywords from text information using a keyword extraction algorithm.
  • the keyword extraction algorithm may include various known algorithms defined to extract key words or sentences from a plurality of words or sentence structures.
  • the keyword may mean a key word, number, sentence, etc. from various texts (eg, words, numbers, sentences, etc.) included in text information.
  • the keyword may mean at least one or more key words, numbers, sentences, etc. that can represent a screenshot image from various texts (eg, words, numbers, sentences, etc.) included in text information.
  • the server S may obtain two keywords from the text information, and the same importance may be reflected in both obtained keywords.
  • the server S may obtain two keywords from the text information, and one of the obtained keywords may reflect a higher importance than the other.
  • a keyword reflecting a relatively high importance may be related to the screenshot image with a higher probability.
  • the keyword to which the relatively high importance is reflected may be a word or sentence more representative of the screenshot image.
  • the keyword may be any one of various texts included in the text information, but is not limited thereto.
  • the keyword may be a word, number, or sentence not included in the text information.
  • the keyword may be a new word, number, sentence, or a combination thereof extracted based on various texts included in text information.
  • the server S may map the acquired keyword to the screenshot image and store the screenshot image to which the keyword is mapped through the step (S2300) of mapping the acquired keyword to the screenshot image and then storing it.
  • the mapping of the obtained keyword to the screenshot image may mean, for example, to display identification so that the controller 100 or the server S can recognize that the keyword represents the screenshot image.
  • the server S may transmit the obtained keyword to the user terminal 1000 through the step S2400 of transmitting the obtained keyword to the user terminal.
  • the server S may transmit the acquired keyword to the user terminal 1000 so that a screen shot image search and classification can be performed on the user terminal 1000 based on the acquired keyword.
  • 15 is a diagram for exemplarily explaining keywords and target information extracted from a screenshot image.
  • various types of text or non-text may be included in a screenshot image SI captured and stored by a user's motion.
  • at least one or more keywords and/or object information may be acquired from the screenshot image through the user terminal 1000 and/or the server S.
  • the first keyword may be obtained from the first text area TA1 included in the screenshot image SI, and the second keyword may be obtained from the second text area TA2.
  • a third keyword may be obtained from the third text area TA3
  • a fourth keyword may be obtained from the fourth text area TA4
  • a fifth keyword may be obtained from the fifth text area TA2 .
  • a sixth keyword may be obtained from the sixth text area TA6
  • a seventh keyword may be obtained from the seventh text area TA7
  • an eighth keyword may be obtained from the eighth text area TA8 . can be obtained.
  • importance may be reflected in the plurality of keywords (eg, first to eighth keywords) obtained as described above.
  • keywords eg, first to eighth keywords
  • the highest importance may be reflected in the first keyword.
  • first object information may be obtained from the first non-text area NTA1 and second object information may be obtained from the second non-text area NTA2 .
  • At least one or more keywords and/or object information obtained from the screenshot image SI may be defined as tag information, and the tag information may be mapped to the screenshot image SI and stored in the user terminal 1000 . .
  • the user can search for a desired screenshot image by using the tag information. A method of retrieving a screenshot image stored in the user terminal 1000 using the acquired at least one or more tag information will be described later.
  • the controller 100 may acquire any one of a plurality of screenshot images stored in the user terminal 1000 based on at least one piece of tag information obtained from the screenshot image SI.
  • the control unit 100 may select at least one of a plurality of screenshot images stored in the storage unit 500 based on a user input obtained through the search unit SE and provide it to the user.
  • the controller 100 may display at least one or more tag information obtained from the screenshot image SI through the recommendation tag display unit REC of the output unit 500 and provide it to the user. Specifically, the controller 100 may display tag information that meets a predetermined criterion through the output unit 500 and recommend it to the user. In this case, the tag information recommended to the user may be a recommendation tag.
  • the user terminal 1000 may transmit the first text information to the server, and the server may transmit the first text information After extracting the first tag information based on , it may be transmitted to the user terminal 1000 .
  • the user terminal 1000 may transmit the second text information to the server, and the server may send the second text information to the server based on the second text information. 2 After the tag information is extracted, it may be transmitted to the user terminal 1000 .
  • the user terminal 1000 may receive the first tag information from the server at a first time and receive the second tag information at a second time. In this case, the user terminal 1000 may determine the first tag information as the recommendation tag, and the first time point may be earlier than the second time point.
  • the user terminal 1000 may receive a plurality of tag information extracted from a plurality of screenshot images from the server. In this case, the user terminal 1000 may determine the recommendation tag based on the frequency at which a plurality of pieces of tag information are received from the server. For example, the user terminal 1000 may determine, as a recommendation tag, tag information with the highest frequency received from the server among a plurality of tag information and provide it to the user.
  • the server may extract tag information based on the obtained keyword.
  • the keyword may be at least one word, number, sentence, or a combination thereof representing a screenshot image among texts included in the text information.
  • the user terminal 1000 may extract tag information from the keyword, and may display the recommendation tag through the recommendation tag display unit REC by using it.
  • a recommendation tag composed of characters may be displayed on the recommendation tag display unit REC as shown in (a) of FIG. 16 . That is, the recommendation tag extracted by the above-described method may be provided in the form of text.
  • a recommendation tag composed of text and a representative image may be displayed on the recommendation tag display unit REC.
  • a recommendation tag composed of a representative image may be displayed on the recommendation tag display unit REC.
  • the representative image displayed on the recommendation tag display unit REC may be an image determined based on the keyword.
  • the representative image is determined based on the keyword, and may be an image that can represent the type or information of the screenshot image well.
  • the controller 100 may acquire a screenshot image captured and stored by the user through the step of acquiring the screenshot image ( S3100 ).
  • the control unit 100 captures by the user's action through the step (S3100) of obtaining a screenshot image and stores at least one screenshot image or two or more screenshot images among a plurality of screenshot images stored in the storage unit 300 image can be obtained.
  • the control unit 100 may obtain a screenshot image from the outside of the user terminal 1000, and in this regard, the description is duplicated because it has been described above through the step (S1100) of obtaining the screenshot image of FIG. 5 . is omitted.
  • the controller 100 may analyze the screenshot image obtained through the step of analyzing the screenshot image ( S3200 ).
  • the controller 100 may perform a step of pre-processing the image so that the obtained screenshot image is suitable for analysis.
  • the step of analyzing the screenshot image ( S3200 ) is the same as or corresponding to the step of analyzing the screenshot image of FIG. 5 ( S1200 ), so a redundant description will be omitted.
  • the controller 100 may extract text information by analyzing the screenshot image obtained through the step of extracting text information from the screenshot image ( S3300 ). More specifically, the controller 100 may divide the screenshot image into a text area and/or a non-text area, and extract text information from the text area through a character reading module. In this regard, since the steps of dividing the screenshot image of FIG. 5 into a plurality of areas ( S1300 ) and extracting text from the text area among the plurality of areas ( S1400 ) have been described above, the overlapping description will be omitted.
  • the controller 100 may transmit the obtained text information to the server S.
  • the server S may obtain a plurality of keywords from the received text information.
  • the server S may acquire a plurality of keywords from text information through a pre-trained neural network model or a known algorithm. .
  • the controller 100 may select a recommendation tag to be provided to the user from among the acquired keywords through the step of selecting a recommendation tag from among the plurality of keywords ( S3500 ).
  • the recommendation tag may be selected based on a predetermined criterion among tag information stored in the storage unit 400 .
  • the control unit 100 may classify the tag information stored in the storage unit 400 according to time order and provide it to the user as a recommended tag. For example, the control unit 100 may provide the tag information stored in the storage unit 400 to the user in the order of recently acquired and stored tags. That is, the recommendation tag provided to the user by the controller 100 may be tag information extracted from the most recently captured and stored screenshot image.
  • the control unit 100 may classify the tag information stored in the storage unit 400 according to the frequency and provide it to the user as a recommended tag. More specifically, since a plurality of tag information is stored in the storage unit 400 , in this case, there may be repeatedly stored tag information. In this case, the control unit 100 may provide the user with the order of the most stored tags among the tag information stored in the storage unit 400 . That is, the recommendation tag provided to the user by the controller 100 may be a tag stored repeatedly the most.
  • the controller 100 may classify the tag information stored in the storage 400 according to the user's activity log and provide it to the user as a recommended tag. After obtaining information about the user's activity log in the user terminal 1000 , the controller 100 may provide suitable tag information to the user based thereon.
  • the control unit 100 may classify the tag information stored in the storage unit 400 by type and provide it to the user as a recommended tag. More specifically, the plurality of screenshot images stored in the storage 400 may be classified by type based on tag information. For example, a screenshot image to which tag information 'webpage' is mapped among a plurality of screenshot images stored in the storage unit 400 may be classified as a first type, and a screen to which tag information 'gifticon' is mapped. The shot image may be classified into a second type, a screenshot image to which tag information of 'shopping mall' is mapped may be classified as a third type, and a screenshot image to which tag information of 'SNS' is mapped may be classified as a fourth type can be classified as
  • the screenshot image may be classified into a plurality of types based on the mapped tag information, and the controller 100 provides tag information based on classifying the screenshot image into a plurality of types as a recommended tag to the user. can do. That is, the recommendation tag provided by the controller 100 to the user may be a 'web page', a 'gifticon', a 'shopping mall', or the like.
  • the control unit 100 may display the recommendation tag selected through the step of providing the recommendation tag to the user ( S3600 ) on the output unit 500 and provide it to the user.
  • the controller 100 provides the recommendation tag to the user through the output unit 500 .
  • the controller 100 may provide only text related to the recommendation tag to the user through the output unit 500 .
  • the controller 100 may provide, through the output unit 500 , at least a partial region of the screenshot image to which the recommendation tag is mapped together with the text related to the recommendation tag to the user.
  • the recommendation tag is provided to the user in the form of an image together with text, the user may more intuitively understand information about the recommendation tag.
  • the shape of the image provided to the user together with the recommendation tag may relate to at least a partial area of the screenshot image.
  • the shape of the image may relate to any one of a text area and/or a non-text area among the screenshot images.
  • the form of the image provided to the user together with the recommendation tag may be an image processed based on the recommendation tag.
  • the user terminal 1000 includes a storage unit 300 for storing a plurality of screenshot images, an output unit 500 for displaying at least one screenshot image among the plurality of screenshot images, and an input unit for receiving a user input ( 400), a communication unit 700 for communicating with an external server, and a control unit 100 for determining at least one screenshot image to be displayed on the output unit among the plurality of screenshot images based on the user input can
  • the output unit 500 includes a search unit SE that allows a user to input a search word for a screen shot image search, and a related search unit SE that displays keywords related to a user input input to the search unit. '), and a search result unit RES for displaying a screen shot image searched for in response to a user input.
  • a plurality of screenshot images may be stored in the storage unit 300 , and at least one piece of tag information may be mapped to each of the plurality of screenshot images.
  • the tag information is extracted from the screenshot image, and since this has been described above, a redundant description will be omitted.
  • the controller 100 may determine the same tag information as the user input from among the tag information stored in the storage 300 or tag information showing a degree of similarity greater than or equal to a certain standard as the matching tag information.
  • the user input obtained through the user input unit 400 may include a first user input and a second user input.
  • the first user input is input at a first time point
  • the second user input is input at a second time point, but the first time point and the second time point may be different.
  • the controller 100 may determine matching tag information based on at least one of a first user input and a second user input. For example, the controller 100 may determine tag information corresponding to at least one of the first user input and the second user input as the matching tag information.
  • control unit 100 may determine the matching tag information based on a first user input and a second user input, and determine the matching tag information by giving importance to the first user input. More specifically, the controller 100 may determine, as the matching tag information, tag information having a higher probability of corresponding to the first user input from among the plurality of tag information corresponding to the first user input and the second user input. .
  • captured time information may be input to the screenshot image stored in the storage unit 300 .
  • the controller 100 may control the plurality of screenshot images to which the matching tag information is mapped to be output to the output unit 500 after arranging them according to the chronological order stored in the storage unit 300 .
  • control unit 100 may control the output to the output unit 500 after sorting in the order most recently stored in the storage unit 300 among a plurality of screenshot images to which the matching tag information is mapped. .
  • the matching tag information is determined based on a user input, and more specifically, tag information corresponding to a user input among a plurality of tag information is determined as the matching tag information.
  • the matching tag information may be determined based on a probability value corresponding to a user input among a plurality of tag information. For example, tag information in which a probability value corresponding to a user input is greater than or equal to a predetermined criterion among a plurality of tag information may be determined as matching tag information.

Landscapes

  • Engineering & Computer Science (AREA)
  • Business, Economics & Management (AREA)
  • Theoretical Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Strategic Management (AREA)
  • Economics (AREA)
  • General Business, Economics & Management (AREA)
  • Marketing (AREA)
  • Tourism & Hospitality (AREA)
  • Development Economics (AREA)
  • Human Resources & Organizations (AREA)
  • General Health & Medical Sciences (AREA)
  • General Engineering & Computer Science (AREA)
  • Health & Medical Sciences (AREA)
  • Entrepreneurship & Innovation (AREA)
  • Finance (AREA)
  • Accounting & Taxation (AREA)
  • Library & Information Science (AREA)
  • Artificial Intelligence (AREA)
  • Game Theory and Decision Science (AREA)
  • Multimedia (AREA)
  • Data Mining & Analysis (AREA)
  • Computational Linguistics (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Databases & Information Systems (AREA)
  • Signal Processing (AREA)
  • Primary Health Care (AREA)
  • Audiology, Speech & Language Pathology (AREA)
  • Quality & Reliability (AREA)
  • Operations Research (AREA)
  • Educational Administration (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Biomedical Technology (AREA)
  • Biophysics (AREA)
  • Evolutionary Computation (AREA)
  • Molecular Biology (AREA)
  • Computing Systems (AREA)
  • Mathematical Physics (AREA)
  • Software Systems (AREA)

Abstract

Selon un mode de réalisation décrit dans la présente invention, un système d'extraction d'informations d'étiquette à partir d'une image de capture d'écran comprend un terminal d'utilisateur dans lequel au moins une image de capture d'écran est stockée, et un serveur qui transmet et reçoit des informations concernant l'image de capture d'écran au terminal d'utilisateur et en provenance de ce dernier. Le terminal d'utilisateur peut acquérir, dans l'image de capture d'écran, une zone de texte dans laquelle un texte est présent et une zone sans texte dans laquelle un texte n'est pas présent, extraire des informations de texte de la zone de texte au moyen d'un module de lecture de lettre, puis transmettre les informations de texte au serveur; et le serveur peut extraire des mots-clés à partir des informations de texte et transmettre, au terminal d'utilisateur, des informations d'étiquette dérivées sur la base des mots-clés.
PCT/KR2021/015463 2020-12-04 2021-10-29 Procédé et système d'extraction d'informations d'étiquette à partir d'une image de capture d'écran Ceased WO2022119136A1 (fr)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
KR20200168173 2020-12-04
KR10-2020-0168173 2020-12-04

Publications (1)

Publication Number Publication Date
WO2022119136A1 true WO2022119136A1 (fr) 2022-06-09

Family

ID=81853205

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/KR2021/015463 Ceased WO2022119136A1 (fr) 2020-12-04 2021-10-29 Procédé et système d'extraction d'informations d'étiquette à partir d'une image de capture d'écran

Country Status (2)

Country Link
KR (3) KR20220079431A (fr)
WO (1) WO2022119136A1 (fr)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2024015457A1 (fr) * 2022-07-15 2024-01-18 Zebra Technologies Corporation Systèmes et procédés d'extraction automatisée de chaînes de textes cibles

Families Citing this family (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
KR102815945B1 (ko) * 2024-07-18 2025-06-09 주식회사 위캐리컴퍼니 컨텐츠 식별 기반 ai를 이용한 태그 자동 생성 시스템

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JPH08287086A (ja) * 1995-04-13 1996-11-01 Nippon Telegr & Teleph Corp <Ntt> 適合度順画像強調表示方法及び装置
US6397213B1 (en) * 1999-05-12 2002-05-28 Ricoh Company Ltd. Search and retrieval using document decomposition
JP2009290798A (ja) * 2008-05-30 2009-12-10 Hitachi Ltd 映像検索装置及び映像検索方法
KR101831783B1 (ko) * 2016-10-27 2018-02-27 주식회사 처음마음 출력 인쇄물에 포함된 이미지 및 텍스트 인식 장치 및 그 방법
KR20200100027A (ko) * 2018-08-31 2020-08-25 엔에이치엔 주식회사 텍스트 기반 이미지 검색 방법 및 장치

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JPH08287086A (ja) * 1995-04-13 1996-11-01 Nippon Telegr & Teleph Corp <Ntt> 適合度順画像強調表示方法及び装置
US6397213B1 (en) * 1999-05-12 2002-05-28 Ricoh Company Ltd. Search and retrieval using document decomposition
JP2009290798A (ja) * 2008-05-30 2009-12-10 Hitachi Ltd 映像検索装置及び映像検索方法
KR101831783B1 (ko) * 2016-10-27 2018-02-27 주식회사 처음마음 출력 인쇄물에 포함된 이미지 및 텍스트 인식 장치 및 그 방법
KR20200100027A (ko) * 2018-08-31 2020-08-25 엔에이치엔 주식회사 텍스트 기반 이미지 검색 방법 및 장치

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2024015457A1 (fr) * 2022-07-15 2024-01-18 Zebra Technologies Corporation Systèmes et procédés d'extraction automatisée de chaînes de textes cibles

Also Published As

Publication number Publication date
KR20220079432A (ko) 2022-06-13
KR20220079433A (ko) 2022-06-13
KR20220079431A (ko) 2022-06-13

Similar Documents

Publication Publication Date Title
WO2019107981A1 (fr) Dispositif électronique reconnaissant du texte dans une image
WO2016209020A1 (fr) Appareil de traitement d&#39;image et procédé de traitement d&#39;image
WO2016085173A1 (fr) Dispositif et procédé pour fournir un contenu écrit à la main dans celui-ci
WO2016017987A1 (fr) Procédé et dispositif permettant d&#39;obtenir une image
WO2016048108A1 (fr) Appareil de traitement d&#39;image et procédé de traitement d&#39;image
WO2016117836A1 (fr) Appareil et procédé de correction de contenu
WO2015016619A1 (fr) Appareil électronique et son procédé de commande, et appareil et procédé de reproduction d&#39;image
WO2015178716A1 (fr) Procédé et dispositif de recherche
WO2017030262A1 (fr) Appareil photo, et procédé de commande associé
WO2022119136A1 (fr) Procédé et système d&#39;extraction d&#39;informations d&#39;étiquette à partir d&#39;une image de capture d&#39;écran
WO2020116983A1 (fr) Appareil électronique, procédé de commande d&#39;appareil électronique et support lisible par ordinateur
WO2015137666A1 (fr) Appareil de reconnaissance d&#39;objet et son procédé de commande
WO2018074768A1 (fr) Procédé d&#39;affichage d&#39;image et dispositif électronique associé
WO2018124842A1 (fr) Procédé et dispositif de fourniture d&#39;informations sur un contenu
WO2016126083A1 (fr) Procédé, dispositif électronique et support d&#39;enregistrement pour notifier des informations de situation environnante
EP3198557A1 (fr) Appareil de traitement d&#39;image et procédé de traitement d&#39;image
WO2020085696A1 (fr) Dispositif électronique et son procédé de commande
WO2021230485A1 (fr) Procédé et appareil de fourniture d&#39;image
WO2014178578A1 (fr) Appareil et procédé de génération de données d&#39;image dans un terminal portable
WO2019098732A1 (fr) Procédé et système de gestion et d&#39;opération sur une image dans un système informatique
WO2021125795A1 (fr) Procédé de fourniture d&#39;étiquettes et dispositif électronique le prenant en charge
WO2021221385A1 (fr) Dispositif électronique et son procédé de fonctionnement
WO2022065561A1 (fr) Procédé de classification d&#39;intention de chaîne de caractères et programme informatique
WO2021235641A1 (fr) Dispositif et procédé d&#39;estimation d&#39;âge
WO2019124775A1 (fr) Dispositif électronique et procédé de fourniture des informations de service associées à un contenu de diffusion dans un dispositif électronique

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: 21900824

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: 21900824

Country of ref document: EP

Kind code of ref document: A1