WO2017113605A1 - Procédé et dispositif d'identification de canal - Google Patents

Procédé et dispositif d'identification de canal Download PDF

Info

Publication number
WO2017113605A1
WO2017113605A1 PCT/CN2016/084684 CN2016084684W WO2017113605A1 WO 2017113605 A1 WO2017113605 A1 WO 2017113605A1 CN 2016084684 W CN2016084684 W CN 2016084684W WO 2017113605 A1 WO2017113605 A1 WO 2017113605A1
Authority
WO
WIPO (PCT)
Prior art keywords
image
channel
matrix
identification
images
Prior art date
Application number
PCT/CN2016/084684
Other languages
English (en)
Chinese (zh)
Inventor
杨杰
颜业钢
Original Assignee
深圳Tcl数字技术有限公司
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 深圳Tcl数字技术有限公司 filed Critical 深圳Tcl数字技术有限公司
Publication of WO2017113605A1 publication Critical patent/WO2017113605A1/fr

Links

Images

Classifications

    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04NPICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N21/00Selective content distribution, e.g. interactive television or video on demand [VOD]
    • H04N21/40Client devices specifically adapted for the reception of or interaction with content, e.g. set-top-box [STB]; Operations thereof
    • H04N21/43Processing of content or additional data, e.g. demultiplexing additional data from a digital video stream; Elementary client operations, e.g. monitoring of home network or synchronising decoder's clock; Client middleware
    • H04N21/442Monitoring of processes or resources, e.g. detecting the failure of a recording device, monitoring the downstream bandwidth, the number of times a movie has been viewed, the storage space available from the internal hard disk
    • H04N21/44213Monitoring of end-user related data
    • H04N21/44222Analytics of user selections, e.g. selection of programs or purchase activity
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06VIMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00Arrangements for image or video recognition or understanding
    • G06V10/40Extraction of image or video features
    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04NPICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N21/00Selective content distribution, e.g. interactive television or video on demand [VOD]
    • H04N21/20Servers specifically adapted for the distribution of content, e.g. VOD servers; Operations thereof
    • H04N21/23Processing of content or additional data; Elementary server operations; Server middleware
    • H04N21/234Processing of video elementary streams, e.g. splicing of video streams or manipulating encoded video stream scene graphs
    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04NPICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N21/00Selective content distribution, e.g. interactive television or video on demand [VOD]
    • H04N21/20Servers specifically adapted for the distribution of content, e.g. VOD servers; Operations thereof
    • H04N21/23Processing of content or additional data; Elementary server operations; Server middleware
    • H04N21/234Processing of video elementary streams, e.g. splicing of video streams or manipulating encoded video stream scene graphs
    • H04N21/23418Processing of video elementary streams, e.g. splicing of video streams or manipulating encoded video stream scene graphs involving operations for analysing video streams, e.g. detecting features or characteristics
    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04NPICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N21/00Selective content distribution, e.g. interactive television or video on demand [VOD]
    • H04N21/20Servers specifically adapted for the distribution of content, e.g. VOD servers; Operations thereof
    • H04N21/25Management operations performed by the server for facilitating the content distribution or administrating data related to end-users or client devices, e.g. end-user or client device authentication, learning user preferences for recommending movies
    • H04N21/258Client or end-user data management, e.g. managing client capabilities, user preferences or demographics, processing of multiple end-users preferences to derive collaborative data
    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04NPICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N21/00Selective content distribution, e.g. interactive television or video on demand [VOD]
    • H04N21/20Servers specifically adapted for the distribution of content, e.g. VOD servers; Operations thereof
    • H04N21/25Management operations performed by the server for facilitating the content distribution or administrating data related to end-users or client devices, e.g. end-user or client device authentication, learning user preferences for recommending movies
    • H04N21/258Client or end-user data management, e.g. managing client capabilities, user preferences or demographics, processing of multiple end-users preferences to derive collaborative data
    • H04N21/25866Management of end-user data
    • H04N21/25891Management of end-user data being end-user preferences
    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04NPICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N21/00Selective content distribution, e.g. interactive television or video on demand [VOD]
    • H04N21/40Client devices specifically adapted for the reception of or interaction with content, e.g. set-top-box [STB]; Operations thereof
    • H04N21/43Processing of content or additional data, e.g. demultiplexing additional data from a digital video stream; Elementary client operations, e.g. monitoring of home network or synchronising decoder's clock; Client middleware
    • H04N21/44Processing of video elementary streams, e.g. splicing a video clip retrieved from local storage with an incoming video stream or rendering scenes according to encoded video stream scene graphs
    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04NPICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N21/00Selective content distribution, e.g. interactive television or video on demand [VOD]
    • H04N21/40Client devices specifically adapted for the reception of or interaction with content, e.g. set-top-box [STB]; Operations thereof
    • H04N21/43Processing of content or additional data, e.g. demultiplexing additional data from a digital video stream; Elementary client operations, e.g. monitoring of home network or synchronising decoder's clock; Client middleware
    • H04N21/44Processing of video elementary streams, e.g. splicing a video clip retrieved from local storage with an incoming video stream or rendering scenes according to encoded video stream scene graphs
    • H04N21/44008Processing of video elementary streams, e.g. splicing a video clip retrieved from local storage with an incoming video stream or rendering scenes according to encoded video stream scene graphs involving operations for analysing video streams, e.g. detecting features or characteristics in the video stream
    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04NPICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N21/00Selective content distribution, e.g. interactive television or video on demand [VOD]
    • H04N21/40Client devices specifically adapted for the reception of or interaction with content, e.g. set-top-box [STB]; Operations thereof
    • H04N21/43Processing of content or additional data, e.g. demultiplexing additional data from a digital video stream; Elementary client operations, e.g. monitoring of home network or synchronising decoder's clock; Client middleware
    • H04N21/442Monitoring of processes or resources, e.g. detecting the failure of a recording device, monitoring the downstream bandwidth, the number of times a movie has been viewed, the storage space available from the internal hard disk

Definitions

  • the present invention relates to the field of smart televisions, and in particular, to a channel identification method and apparatus.
  • the main object of the present invention is to solve the problem that the existing channel identification process needs to interact with the server at all times, and the operation speed is slow and the recognition accuracy is low.
  • the present invention provides a channel identification method, the channel identification method comprising the following steps:
  • the points in the difference matrix are binarized, and the obtained binarized image is used as a feature image;
  • the similarity between the preset library images and the feature image is separately calculated, and the channel corresponding to the library image with the highest similarity is used as the channel locked by the user in the recognition period.
  • the present invention provides a channel identification method, the channel identification method comprising the following steps:
  • the similarity between the preset library images and the feature image is separately calculated, and the channel corresponding to the library image with the highest similarity is used as the channel locked by the user in the recognition period.
  • the step of performing feature extraction according to all the identification images in the identification period to obtain the feature image comprises:
  • the points in the difference matrix are binarized, and the obtained binarized image is used as the feature image.
  • the present invention further provides a channel identification apparatus, where the channel identification apparatus includes:
  • An acquisition module configured to collect multiple screen images in a preset recognition period
  • An identifier extraction module configured to respectively extract an identifier image containing a channel identifier from each of the screen images
  • a feature extraction module configured to perform feature extraction according to all the identification images in the recognition period to obtain a feature image
  • a determining module configured to separately calculate a similarity between the preset library images and the feature image, and use a channel corresponding to the library image with the highest similarity as a frequency locked by the user in the identification period Road.
  • the invention collects the screen image by timing, and extracts the identification image containing the channel identifier based on the collected screen image, and extracts the feature image based on the smaller identification image, thereby greatly reducing the calculation amount of the extracted feature image and improving the operation efficiency.
  • only the small logo image is processed, which reduces the complexity of the image and improves the accuracy of the recognition.
  • by collecting the screen image and the pre-stored library image for processing there is no need to interact with the server at all times, which further improves the operation efficiency and saves the cost.
  • FIG. 1 is a schematic flow chart of a first embodiment of a channel identification method according to the present invention.
  • FIG. 2 is a schematic flow chart of a second embodiment of a channel identification method according to the present invention.
  • FIG. 3 is a schematic flow chart of a third embodiment of a channel identification method according to the present invention.
  • FIG. 4 is a schematic flow chart of a fourth embodiment of a channel identification method according to the present invention.
  • FIG. 5 is a schematic diagram of functional modules of a first embodiment of a channel identification apparatus according to the present invention.
  • FIG. 6 is a schematic diagram of functional modules of a second embodiment of a channel identification apparatus according to the present invention.
  • the invention provides a channel identification method.
  • FIG. 1 is a schematic flowchart diagram of a first embodiment of a channel identification method according to the present invention.
  • the channel identification method includes:
  • Step S10 collecting a plurality of screen images in a preset recognition period
  • the smart TV collects a plurality of screen images in a preset recognition period to analyze and determine a channel locked by the user during the recognition period.
  • the smart TV can collect the screen image with a preset sampling period; the smart TV summarizes the collected screen images according to the recognition period for analysis.
  • the sampling period can be set to 5 minutes, and the recognition period can be set to 1 hour, for example, smart TV every 5
  • the process of collecting screen images by smart TVs can be achieved by screen capture.
  • the smart TV can obtain the screen image through the screen capture; the smart TV determines whether the captured image resolution conforms to the preset reference sample specification; if the reference sample specification is met, the captured image is saved as the sample image; For the reference sample size, the captured image is interpolated to obtain an image conforming to the reference sample size, and the image obtained by the interpolation process is saved as a sample image.
  • the reference sample specification makes the channel identification method of the present invention applicable to different models.
  • the resolution and aspect ratio of 1920*1080 can be selected as the reference sample specification, and the data can be selected to not only take care of the high-end machine of 4K resolution. Type, it can also be taken care of for low-end models.
  • the screen capture specification is 1920*1080, then the direct screen capture is used as the sample image; for models other than 1920*1080, the image of the screen capture image needs to be interpolated to convert to 1920*1080. Specifications of the image.
  • the smart TV can read the resolution of the screen before collecting the screen image, and determine whether the read screen resolution conforms to the preset reference sample specification, and if the screen resolution conforms to the benchmark sample specification, The screen image is directly captured as a sample image; if the screen resolution does not conform to the reference sample specification, the interpolation process is performed each time the screen image is captured, and an image conforming to the reference sample specification is obtained for storage. It is not necessary to perform a judgment on whether or not the reference sample specification is met every time the screen image is captured, thereby improving the processing efficiency.
  • Step S20 extracting a logo image containing a channel identifier from each screen image
  • the smart TV separately extracts the identification image of each screen image to obtain a corresponding plurality of identification images.
  • the screen image is an image that conforms to a benchmark sample size.
  • the smart television can extract the identification image from the determined area by determining an area containing the channel identification in the screen image.
  • Step S30 performing feature extraction according to all the identification images in the identification period to obtain a feature image
  • the smart TV extracts features according to all the identification images in the recognition period, and extracts feature images containing feature points from the plurality of identification images, and the feature points are points at which the identification images are fixed in the recognition period.
  • the smart television extracts the feature image extracted according to the plurality of identification images into one, that is, the channel locked by the user determined within one identification period is one.
  • Step S40 respectively calculating the similarity between the preset library images and the feature image, and using the channel corresponding to the library image with the highest similarity as the channel locked by the user in the recognition period.
  • the smart TV separately calculates the similarity between the preset library images and the feature images; the smart TV uses the channel corresponding to the library image with the highest similarity as the channel locked by the user in the recognition period. Specifically, the smart TV reads the preset library image, calculates the similarity between each library image and the feature image, determines the library image with the highest similarity, reads the channel information corresponding to the determined library image, and determines the user according to the channel information. The channel that is locked during the recognition period.
  • the smart TV may further load the preset library image and the corresponding channel information to perform channel identification according to the library image and the corresponding channel information.
  • the smart television may also acquire the library image and the corresponding channel information from the server in the channel identification process.
  • the collected screen images are all one channel, and the calculation accuracy is the highest at this time; if the user switches channels a small amount in one recognition period and is collected to the screen. If the sample sampled by one of the channels A reaches a preset ratio or more, the algorithm has the fault tolerance mechanism, and the channel A can still be calculated; if the user frequently changes the station within one recognition period, and the number of sampling times of one channel does not exceed The preset number of times, this calculation loses meaning, the situation is consistent with reality, users frequently change channels quickly, and they do not have statistical characteristics, such cases are not considered.
  • the screen image is collected periodically, and the identifier image containing the channel identifier is extracted based on the collected screen image, and the feature image is extracted based on the smaller logo image, thereby greatly reducing the calculation amount of the extracted feature image and improving the operation efficiency.
  • only the small logo image is processed, which reduces the complexity of the image and improves the accuracy of the recognition.
  • by collecting the screen image and the pre-stored library image for processing there is no need to interact with the server at all times, which further improves the operation efficiency and saves the cost.
  • FIG. 2 is a schematic flowchart diagram of a second embodiment of a channel identification method according to the present invention. Based on the first embodiment of the above channel identification method, step S30 includes:
  • Step S31 calculating an average matrix according to a matrix corresponding to each identifier image
  • the smart television can generate a corresponding matrix C i (matrix corresponding to the i-th identification image) according to each identification image, and calculate an average matrix C mean based on the generated respective matrices C i . For example, if the resolution of the identification image is 480*270, the corresponding matrix C i scale is 480*270.
  • the smart television may generate a corresponding matrix S i (matrix corresponding to the i-th screen image) according to each screen image, and separately segment each matrix S i to obtain a corresponding matrix C i .
  • the obtained matrix S i has a specification of 1920*1080
  • the matrix C i obtained by performing the sixteen-square grid cutting according to the matrix S i is 480*270.
  • the smart television calculates a mean matrix C mean according to each matrix C i , wherein:
  • Step S32 respectively calculating a distance matrix of the matrix corresponding to each identifier image and the mean matrix
  • Smart TV image respectively calculated for each identifier corresponding to the distance matrix and the matrix C i D i C mean mean matrix (matrix C i corresponding to the distance matrix).
  • the smart television calculates the distance matrix D i of each matrix C i and the mean matrix C mean to obtain a distance matrix set D, where:
  • D [C 1 -C mean , C 2 -C mean ,...C i -C mean ,...,C 12 -C mean ],i ⁇ (1,12)
  • Step S33 performing feature decomposition on each distance matrix to obtain a corresponding feature vector matrix
  • the smart television separately decomposes each distance matrix D i to obtain a corresponding feature vector matrix V i (a feature vector matrix corresponding to the distance matrix D i ).
  • the smart television separately decomposes each distance matrix D i to obtain a corresponding feature vector matrix V i and obtains a feature feature vector matrix set V, wherein:
  • V [V 1 , V 2 ,...,V i ,...,V 12 ],i ⁇ (1,12)
  • Step S34 calculating an entropy matrix corresponding to each feature vector matrix
  • Step S35 calculating an average value of all entropy value matrices to obtain a difference matrix
  • the smart TV calculates the average of all entropy matrix E i to obtain the difference matrix E mean .
  • the difference matrix E mean is calculated, where:
  • step S36 the points in the difference matrix are binarized, and the obtained binarized image is used as the feature image.
  • the smart television binarizes the points in the difference matrix E mean and uses the obtained binarized image as the feature image. Specifically, the smart TV can sort the points in the difference matrix E mean by the size, the last 10,000 points are marked as white, and the remaining points are marked as black, and the corresponding binarized image is obtained.
  • the binarization process can effectively distinguish the channel identification from the background.
  • a feature image containing a feature point is obtained by extracting a fixed feature point in the identification area within the recognition period from the identification image containing the channel identifier.
  • the extracted feature image contains more information related to the channel identification, which improves the calculation accuracy.
  • by comparing the feature image and the library image to determine the channel locked by the user it is not necessary to interact with the server at any time, thereby further improving the operation efficiency and saving the cost.
  • FIG. 3 is a schematic flowchart diagram of a third embodiment of a channel identification method according to the present invention. Based on the first embodiment of the above channel identification method, step S20 includes:
  • Step S21 each screen image is separately cut into sixteen squares to obtain a corresponding area image group
  • Step S22 extracting an area image of the first row and the first column of each area image group as a corresponding identification image.
  • the smart TV divides each screen image into a corresponding 16-square grid to obtain a corresponding area image group; extracts an area image of the first row and the first column of each area image group as a corresponding identification image, and obtains a logo image corresponding to each screen image. To extract the feature image based on the obtained logo image.
  • smart television may also generate the corresponding matrix S i, respectively, each of the smart TV matrix S i for sixteen grids obtained by cutting each of the corresponding sub-matrix based on a screen image; a first row and first column of sub smart TV extraction identity matrix as the matrix C i S i of the image corresponding to the matrix.
  • the obtained matrix S i has a specification of 1920*1080
  • the matrix C i obtained by performing the sixteen-square grid cutting according to the matrix S i is 480*270.
  • the screen image is collected periodically, and the identifier image containing the channel identifier is extracted based on the collected screen image, and the feature image is extracted based on the smaller logo image, thereby greatly reducing the calculation amount of the extracted feature image and improving the operation efficiency.
  • only the small logo image is processed, which reduces the complexity of the image and improves the accuracy of the recognition.
  • FIG. 4 is a schematic flowchart diagram of a fourth embodiment of a channel identification method according to the present invention. Based on the first embodiment of the channel identification method, after step S40, the method further includes:
  • Step S51 performing statistics on channels locked by the user in each recognition period within a preset time period
  • step S52 the channel with the preset number of times of the number of locks in the statistical result is used as the preferred channel of the user within the preset time.
  • the smart TV counts the channels locked by the user in each recognition period in the preset time; the channel in which the number of locks is ranked in the preset number of times in the statistical result is used as the preferred channel of the user within the preset time.
  • the preset number of bits can be set according to the actual settings. For example, it can be set to the top three or the top ten. Further, the smart TV can also upload the top three channels in the statistical results to the server. In order to understand the user's usage habits of the smart TV and the user's preference for the television program through the server, the television content of the user's preference is provided in a targeted manner.
  • the preset time can be 24 hours, 48 hours, and the like.
  • the top three channels are used as the preferred channels of the user in the preset time, so that the service provider can understand the user's usage habits of the smart TV and the user preference television programs. Therefore, the television content of the user's preference is provided in a targeted manner.
  • the execution bodies of the channel identification methods of the above-described first to fourth embodiments may each be a smart TV or an intelligent terminal for playing a television program. Further, the channel identification method may be implemented by a client program installed on a smart TV or a smart terminal for playing a television program, wherein the smart terminal may include, but is not limited to, a mobile phone, a smart phone, a notebook computer, a PDA. Terminals (personal digital assistants), PADs (tablets), PMPs (portable multimedia players), desktop computers, and the like.
  • the invention further provides a channel identification device.
  • FIG. 5 is a schematic diagram of functional modules of a first embodiment of a channel identification apparatus according to the present invention.
  • the channel identification device includes: an acquisition module 10, an identifier extraction module 20, a feature extraction module 30, and a determination module 40.
  • the collecting module 10 is configured to collect a plurality of screen images in a preset recognition period
  • the smart TV collects a plurality of screen images in a preset recognition period to analyze and determine a channel locked by the user during the recognition period.
  • the smart TV can collect the screen image with a preset sampling period; the smart TV summarizes the collected screen images according to the recognition period for analysis.
  • the sampling period can be set to 5 minutes, and the recognition period can be set to 1 hour.
  • the 12 screen images are analyzed as samples to determine the channel that the user has locked within 1 hour.
  • the process of collecting screen images by smart TVs can be achieved by screen capture.
  • the smart TV can obtain the screen image through the screen capture; the smart TV determines whether the resolution of the captured image conforms to the preset.
  • the reference sample specification if the reference sample specification is met, the captured image is saved as a sample image; if the reference sample specification is not met, the intercepted image is interpolated to obtain an image conforming to the reference sample specification, and the interpolation is processed.
  • the resulting image is saved as a sample image.
  • the reference sample specification makes the channel identification method of the present invention applicable to different models.
  • the resolution and aspect ratio of 1920*1080 can be selected as the reference sample specification, and the data can be selected to not only take care of the high-end machine of 4K resolution. Type, it can also be taken care of for low-end models.
  • the screen capture specification is 1920*1080, then the direct screen capture is used as the sample image; for models other than 1920*1080, the image of the screen capture image needs to be interpolated to convert to 1920*1080. Specifications of the image.
  • the smart TV can read the resolution of the screen before collecting the screen image, and determine whether the read screen resolution conforms to the preset reference sample specification, and if the screen resolution conforms to the benchmark sample specification, The screen image is directly captured as a sample image; if the screen resolution does not conform to the reference sample specification, the interpolation process is performed each time the screen image is captured, and an image conforming to the reference sample specification is obtained for storage. It is not necessary to perform a judgment on whether or not the reference sample specification is met every time the screen image is captured, thereby improving the processing efficiency.
  • the identifier extraction module 20 is configured to respectively extract the identifier image containing the channel identifier from each screen image;
  • the smart TV separately extracts the identification image of each screen image to obtain a corresponding plurality of identification images.
  • the screen image is an image that conforms to a benchmark sample size.
  • the smart television can extract the identification image from the determined area by determining an area containing the channel identification in the screen image.
  • the feature extraction module 30 is configured to perform feature extraction according to all the identification images in the recognition period to obtain a feature image
  • the smart TV extracts features according to all the identification images in the recognition period, and extracts feature images containing feature points from the plurality of identification images, and the feature points are points at which the identification images are fixed in the recognition period.
  • the smart television extracts the feature image extracted according to the plurality of identification images into one, that is, the channel locked by the user determined within one identification period is one.
  • the determining module 40 is configured to separately calculate the similarity between the preset library images and the feature image, and use the channel corresponding to the library image with the highest similarity as the channel locked by the user in the recognition period.
  • the smart TV separately calculates the similarity between the preset library images and the feature images; the smart TV uses the channel corresponding to the library image with the highest similarity as the channel locked by the user in the recognition period. specific, The smart TV reads the preset library image, calculates the similarity between each library image and the feature image, determines the library image with the highest similarity, reads the channel information corresponding to the determined library image, and determines the user in the recognition cycle according to the channel information. Internal locked channel.
  • the channel identification device may further include an initialization module, and the initialization module is configured to load the preset library image and the corresponding channel information to perform channel identification according to the library image and the corresponding channel information.
  • the smart TV can load the preset library image and the corresponding channel information through the initialization module to perform channel identification according to the library image and the corresponding channel information.
  • the smart television may also acquire the library image and the corresponding channel information from the server in the channel identification process.
  • the collected screen images are all one channel, and the calculation accuracy is the highest at this time; if the user switches channels a small amount in one recognition period and is collected to the screen. If the sample sampled by one of the channels A reaches a preset ratio or more, the algorithm has the fault tolerance mechanism, and the channel A can still be calculated; if the user frequently changes the station within one recognition period, and the number of sampling times of one channel does not exceed The preset number of times, this calculation loses meaning, the situation is consistent with reality, users frequently change channels quickly, and they do not have statistical characteristics, such cases are not considered.
  • the screen image is collected periodically, and the identifier image containing the channel identifier is extracted based on the collected screen image, and the feature image is extracted based on the smaller logo image, thereby greatly reducing the calculation amount of the extracted feature image and improving the operation efficiency.
  • only the small logo image is processed, which reduces the complexity of the image and improves the accuracy of the recognition.
  • by collecting the screen image and the pre-stored library image for processing there is no need to interact with the server at all times, which further improves the operation efficiency and saves the cost.
  • FIG. 6 is a schematic diagram of functional modules of a second embodiment of the apparatus of the present invention.
  • the feature extraction module 30 includes a calculation unit 31 and a binarization unit 32.
  • the calculating unit 31 is configured to calculate an average matrix according to a matrix corresponding to each identifier image
  • the smart television can generate a corresponding matrix C i (matrix corresponding to the i-th identification image) according to each identification image, and calculate an average matrix C mean based on the generated respective matrices C i . For example, if the resolution of the identification image is 480*270, the corresponding matrix C i scale is 480*270.
  • smart television may be the logo image extraction process, generates a corresponding matrix S i (i-th screen image corresponding to a matrix) in accordance with various screen images, respectively, each of the matrix S i dividing process to obtain the corresponding matrix C i.
  • the resolution of the screen image is 1920*1080
  • the obtained matrix S i has a specification of 1920*1080
  • the matrix C i obtained by performing the sixteen-square grid cutting according to the matrix S i is 480*270.
  • the smart television calculates a mean matrix C mean according to each matrix C i , wherein:
  • the calculating unit 31 is further configured to separately calculate a distance matrix of the matrix corresponding to each identifier image and the mean matrix;
  • Smart TV image respectively calculated for each identifier corresponding to the distance matrix and the matrix C i D i C mean mean matrix (matrix C i corresponding to the distance matrix).
  • the smart television calculates the distance matrix D i of each matrix C i and the mean matrix C mean to obtain a distance matrix set D, where:
  • D [C 1 -C mean , C 2 -C mean ,...C i -C mean ,...,C 12 -C mean ],i ⁇ (1,12)
  • the calculating unit 31 is further configured to perform feature decomposition on each distance matrix to obtain a corresponding feature vector matrix
  • the smart television separately decomposes each distance matrix D i to obtain a corresponding feature vector matrix V i (a feature vector matrix corresponding to the distance matrix D i ).
  • the smart television separately decomposes each distance matrix D i to obtain a corresponding feature vector matrix V i and obtains a feature feature vector matrix set V, wherein:
  • V [V 1 , V 2 ,...,V i ,...,V 12 ],i ⁇ (1,12)
  • the calculating unit 31 is further configured to calculate a matrix of entropy values corresponding to each feature vector matrix
  • the calculating unit 31 is further configured to calculate an average value of all entropy value matrices to obtain a difference matrix
  • the smart TV calculates the average of all entropy matrix E i to obtain the difference matrix E mean .
  • the difference matrix E mean is calculated, where:
  • the binarization unit 32 is configured to perform binarization processing on the points in the difference matrix, and use the obtained binarized image as the feature image.
  • the smart television binarizes the points in the difference matrix E mean and uses the obtained binarized image as the feature image. Specifically, the smart TV can sort the points in the difference matrix E mean by the size, the last 10,000 points are marked as white, and the remaining points are marked as black, and the corresponding binarized image is obtained.
  • the binarization process can effectively distinguish the channel identification from the background.
  • a feature image containing a feature point is obtained by extracting a fixed feature point in the identification area within the recognition period from the identification image containing the channel identifier.
  • the extracted feature image contains more information related to the channel identification, which improves the calculation accuracy.
  • by comparing the feature image and the library image to determine the channel locked by the user it is not necessary to interact with the server at any time, thereby further improving the operation efficiency and saving the cost.
  • the identification extraction module 20 includes a cutting unit 21 and an extraction unit 22;
  • a cutting unit 21 configured to perform a sixteen-square grid cut on each screen image to obtain a corresponding area image group
  • the extracting unit 22 is configured to extract an area image of the first row and the first column of each regional image group as a corresponding identification image.
  • the smart TV divides each screen image into a corresponding 16-square grid to obtain a corresponding area image group; extracts an area image of the first row and the first column of each area image group as a corresponding identification image, and obtains a logo image corresponding to each screen image. To extract the feature image based on the obtained logo image.
  • smart television may also generate the corresponding matrix S i, respectively, each of the smart TV matrix S i for sixteen grids obtained by cutting each of the corresponding sub-matrix based on a screen image; a first row and first column of sub smart TV extraction identity matrix as the matrix C i S i of the image corresponding to the matrix.
  • the obtained matrix S i has a specification of 1920*1080
  • the matrix C i obtained by performing the sixteen-square grid cutting according to the matrix S i is 480*270.
  • the screen image is collected periodically, and the identifier image containing the channel identifier is extracted based on the collected screen image, and the feature image is extracted based on the smaller logo image, thereby greatly reducing the calculation amount of the extracted feature image and improving the operation efficiency.
  • only the small logo image is processed, which reduces the complexity of the image and improves the accuracy of the recognition.
  • the channel identification device further includes a statistics module 50;
  • the statistics module 50 is configured to perform statistics on channels locked by users in each identification period within a preset time period
  • the determining module 40 is further configured to use, as the preferred channel of the user, the channel of the preset number of times of the number of locks in the statistical result.
  • the smart TV counts the channels locked by the user in each recognition period in the preset time; the channel in which the number of locks is ranked in the preset number of times in the statistical result is used as the preferred channel of the user within the preset time.
  • the preset number of bits can be set according to the actual settings. For example, it can be set to the top three or the top ten. Further, the smart TV can also upload the top three channels in the statistical results to the server. In order to understand the user's usage habits of the smart TV and the user's preference for the television program through the server, the television content of the user's preference is provided in a targeted manner.
  • the preset time can be 24 hours, 48 hours, and the like.
  • the top three channels are used as the preferred channels of the user in the preset time, so that the service provider can understand the user's usage habits of the smart TV and the user preference television programs. Therefore, the television content of the user's preference is provided in a targeted manner.

Landscapes

  • Engineering & Computer Science (AREA)
  • Multimedia (AREA)
  • Signal Processing (AREA)
  • Databases & Information Systems (AREA)
  • Social Psychology (AREA)
  • General Health & Medical Sciences (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Health & Medical Sciences (AREA)
  • Computer Graphics (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • Image Analysis (AREA)

Abstract

La présente invention concerne un procédé d'identification de canal, le procédé d'identification de canal comprenant les étapes suivantes : capturer une pluralité d'images d'écran dans une période d'identification préréglée ; extraire une image d'identification comprenant un identificateur de canal à partir, respectivement, de chacune des images d'écran ; réaliser une extraction de caractéristique selon toutes les images d'identification dans la période d'identification, de façon à obtenir une image de caractéristique ; et calculer la similarité entre, respectivement, chaque image de bibliothèque préétablie et l'image de caractéristique, et prendre le canal correspondant à l'image de bibliothèque ayant la similarité la plus élevée comme le canal verrouillé par un utilisateur dans la période d'identification. La présente invention concerne en outre un dispositif d'identification de canal. Au moyen de la présente invention, la quantité de calcul pour extraire l'image de caractéristique est réduite, l'efficacité de fonctionnement est améliorée et, également, uniquement de petites images d'identification sont traitées, réduisant la complexité des images et améliorant la précision d'identification. En outre, le traitement est réalisé au moyen des images d'écran capturées et des images de bibliothèque pré-stockées, sans avoir besoin d'une interaction avec un serveur à tous les instants, permettant ainsi d'améliorer l'efficacité de fonctionnement, et d'économiser les coûts.
PCT/CN2016/084684 2015-12-28 2016-06-03 Procédé et dispositif d'identification de canal WO2017113605A1 (fr)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201510999595.0 2015-12-28
CN201510999595.0A CN105635750B (zh) 2015-12-28 2015-12-28 频道识别方法及装置

Publications (1)

Publication Number Publication Date
WO2017113605A1 true WO2017113605A1 (fr) 2017-07-06

Family

ID=56050178

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2016/084684 WO2017113605A1 (fr) 2015-12-28 2016-06-03 Procédé et dispositif d'identification de canal

Country Status (2)

Country Link
CN (1) CN105635750B (fr)
WO (1) WO2017113605A1 (fr)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
KR20210054618A (ko) 2019-11-05 2021-05-14 (주)일리드 실내 인테리어 디자인 생성 방법

Families Citing this family (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN105635750B (zh) * 2015-12-28 2018-07-03 深圳Tcl数字技术有限公司 频道识别方法及装置

Citations (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101324923A (zh) * 2008-08-05 2008-12-17 北京中星微电子有限公司 用于人脸识别的特征提取方法和装置
WO2009073895A1 (fr) * 2007-12-07 2009-06-11 Verimatrix, Inc. Systèmes et procédés pour effectuer une analyse sémantique d'objets de média
CN103634652A (zh) * 2013-11-06 2014-03-12 小米科技有限责任公司 台标识别方法、装置、电视机和系统
CN104023249A (zh) * 2014-06-12 2014-09-03 腾讯科技(深圳)有限公司 电视频道识别方法和装置
CN104598910A (zh) * 2015-01-16 2015-05-06 科大讯飞股份有限公司 基于梯度方向匹配算法的智能电视台标识别方法及系统
CN104639952A (zh) * 2015-01-23 2015-05-20 小米科技有限责任公司 台标识别方法及装置
CN105635750A (zh) * 2015-12-28 2016-06-01 深圳Tcl数字技术有限公司 频道识别方法及装置

Patent Citations (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2009073895A1 (fr) * 2007-12-07 2009-06-11 Verimatrix, Inc. Systèmes et procédés pour effectuer une analyse sémantique d'objets de média
CN101324923A (zh) * 2008-08-05 2008-12-17 北京中星微电子有限公司 用于人脸识别的特征提取方法和装置
CN103634652A (zh) * 2013-11-06 2014-03-12 小米科技有限责任公司 台标识别方法、装置、电视机和系统
CN104023249A (zh) * 2014-06-12 2014-09-03 腾讯科技(深圳)有限公司 电视频道识别方法和装置
CN104598910A (zh) * 2015-01-16 2015-05-06 科大讯飞股份有限公司 基于梯度方向匹配算法的智能电视台标识别方法及系统
CN104639952A (zh) * 2015-01-23 2015-05-20 小米科技有限责任公司 台标识别方法及装置
CN105635750A (zh) * 2015-12-28 2016-06-01 深圳Tcl数字技术有限公司 频道识别方法及装置

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
KR20210054618A (ko) 2019-11-05 2021-05-14 (주)일리드 실내 인테리어 디자인 생성 방법

Also Published As

Publication number Publication date
CN105635750A (zh) 2016-06-01
CN105635750B (zh) 2018-07-03

Similar Documents

Publication Publication Date Title
US11132555B2 (en) Video detection method, server and storage medium
US11341272B2 (en) Methods and apparatus to assign demographic information to panelists
CN105740707B (zh) 恶意文件的识别方法和装置
CN111881913A (zh) 图像识别方法及装置、存储介质和处理器
US9596520B2 (en) Method and system for pushing information to a client
US20160196478A1 (en) Image processing method and device
WO2017012418A1 (fr) Procédé et appareil de traitement d'image
CN110460838B (zh) 一种镜头切换的检测方法、装置及计算机设备
US10402633B2 (en) Human detection in high density crowds
WO2017166597A1 (fr) Procédé et appareil de reconnaissance de vidéo de dessin animé, et dispositif électronique
CN112653928B (zh) 一种基于内容相同的视频过滤方法、系统及设备
CN110765903A (zh) 行人重识别方法、装置及存储介质
WO2017113605A1 (fr) Procédé et dispositif d'identification de canal
US20140140622A1 (en) System and method facilitating designing of classifier while recognizing characters in a video
Ma et al. Multiscale 2-D singular spectrum analysis and principal component analysis for spatial–spectral noise-robust feature extraction and classification of hyperspectral images
CN109711287B (zh) 人脸采集方法及相关产品
US8988219B2 (en) Alert system based on camera identification
CN110765291A (zh) 检索方法、装置及电子设备
WO2015031350A1 (fr) Systèmes et procédés d'utilisation de mémoire pour la détection d'objets
CN106971150B (zh) 基于逻辑回归的排队异常检测方法及装置
CN112364918A (zh) 异常识别方法、终端和计算机可读存储介质
JP4967045B2 (ja) 背景判別装置、方法及びプログラム
CN115474091A (zh) 一种基于分解元图的动作捕捉方法及装置
Li et al. Autocorrelation and dct based quality metrics for fingerprint samples generated by smartphones
CN112199131B (zh) 一种页面检测方法、装置及设备

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: 16880413

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

32PN Ep: public notification in the ep bulletin as address of the adressee cannot be established

Free format text: NOTING OF LOSS OF RIGHTS PURSUANT TO RULE 112(1) EPC (EPO FORM 1205A DATED 09/11/2018)

122 Ep: pct application non-entry in european phase

Ref document number: 16880413

Country of ref document: EP

Kind code of ref document: A1