KR102208916B1 - System for recognizing broadcast program based on image recognition - Google Patents

Abstract

The present invention captures a broadcast screen broadcast through a current broadcast output device in a user terminal, and extracts object information such as text or performers appearing on the captured broadcast screen so that the currently broadcast program can be automatically recognized. It's about technology.
In the broadcast program recognition system based on image recognition according to the present invention, a broadcast output device that outputs a broadcast program broadcasted by a broadcaster in real time, an application is installed, and a camera is driven according to the execution of the application to be output in real time from the broadcast output device. In addition to acquiring a plurality of broadcast screens from a broadcast image and transmitting the extracted feature information and broadcast program recognition request information including the broadcast screen collection time to the service server by extracting feature information from the plurality of broadcast screens, service A user terminal that provides a service provided by an application using broadcasting information authorized from a server, a broadcasting program according to broadcasting time, and characteristic information for each broadcasting program are registered and stored, and approved from the user terminal in connection with the application executed in the user terminal. Based on the broadcast program recognition request information, the feature information of the broadcast program corresponding to the broadcast time and feature information received from the user terminal are compared to search for broadcast information corresponding to the feature information requested by the user, and the searched broadcast information is retrieved from the user. It is configured to include a service server transmitted to the terminal, and the characteristic information for recognizing the broadcast program is text information including a broadcaster logo or broadcast program name displayed on a broadcast screen, or a performer or object output on the broadcast screen. It is characterized in that it is object information to include.

Description

System for recognizing broadcast program based on image recognition

The present invention captures a broadcast screen broadcast through a current broadcast output device in a user terminal, and extracts object information such as text or performers appearing on the captured broadcast screen so that the currently broadcast program can be automatically recognized. It's about technology.

Augmented reality is a technology that superimposes a virtual object through a computer on the real world that users see through the eyes.With the spread of smartphones, a service that combines the real world shot through a camera and a virtual image or video is commercialized. have.

In addition, an augmented reality system using mobile devices has become possible with the improvement of computing capabilities of mobile devices such as portable terminals and PDAs and the development of wireless network devices.

For example, when a camera is used to illuminate the surroundings with a smartphone, information such as a location of a nearby store and a phone number may be displayed in a 3D image.

In particular, in recent years, various smart phone AR applications such as recognizing TV programs and providing PPL (Product PLacement) suitable for the recognized broadcast TV programs through smart phones, etc. have appeared. Various services are provided to users in partnership with currently broadcast TV programs.

At this time, the smart phone AR application must recognize the current TV broadcast program, and currently, while the TV and the smart phone are connected, the smart phone AR application downloads the current broadcast program from the TV and recognizes the currently broadcast program.

However, in order to use this method, there is a limitation in that the user must connect the smart phone and the TV in advance using Bluetooth, and the TV must support all of these functions.

In addition, in order for users to use AR services while watching TV, all settings must be set in advance.In reality, it is rare for users to watch TV in a state where all of these functions are set after recognizing the use of these functions. It is difficult to activate the use of phone AR applications.

In this regard, a broadcast-linked augmented reality service system and method are also disclosed in Prior Document 1 (Korean Patent No. 10-1850482).

However, Prior Document 1 creates meta information for AR service in a broadcast server in advance, broadcasts it with broadcast information, and recognizes meta information from a broadcast screen shot in an AR application installed on a smartphone, thereby creating an AR corresponding to a broadcast program. It is configured to provide services.

In other words, Prior Document 1 has a limitation in that a broadcaster-linked service is possible only when a broadcaster server generates a broadcast image by inserting meta information for an augmented reality service into each broadcast image in advance.

In addition, broadcasters are inconvenient to modify their original content for a specific service and input meta information, and users have limitations in having to view unwanted meta information on the screen.

1. Korean Patent Registration No. 10-1850482 (Name of invention: Broadcast-linked augmented reality service system and method)

Accordingly, the present invention was created in view of the above circumstances, and the user terminal captures a broadcast screen broadcast through a broadcast output device, and extracts object information such as text or performers appearing on the captured broadcast screen. There is a technical purpose of providing a broadcast program recognition system based on image recognition that enables automatic recognition of a broadcast program being broadcast.

According to an aspect of the present invention for achieving the above object, a broadcast output device that outputs a broadcast program broadcasted by a broadcaster in real time, an application is installed, and a camera is driven according to the execution of the application to be output in real time from the broadcast output device. In addition to acquiring a plurality of broadcast screens from a broadcast image and transmitting the extracted feature information and broadcast program recognition request information including the broadcast screen collection time to the service server by extracting feature information from the plurality of broadcast screens, service A user terminal that provides a service provided by an application using broadcasting information authorized from a server, a broadcasting program according to broadcasting time, and characteristic information for each broadcasting program are registered and stored, and approved from the user terminal in connection with the application executed in the user terminal. Based on the broadcast program recognition request information, the feature information of the broadcast program corresponding to the broadcast time and feature information received from the user terminal are compared to search for broadcast information corresponding to the feature information requested by the user, and the searched broadcast information is retrieved from the user. Consisting of a service server transmitted to a terminal, the characteristic information for recognizing a broadcast program is text information including a broadcaster logo or broadcast program name displayed on a broadcast screen, or a performer or object output on the broadcast screen. Consists of object information to be included, and the user terminal extracts a certain area on the broadcast screen as a region of interest based on both corners of the upper end of the broadcast output device and recognizes the text. If recognized, the text information is transmitted to the service server as feature information, and text recognition is performed in the second region of interest based on the second corner only when the text is not recognized in the first region of interest based on the first corner. Text information is extracted, and when text information is not extracted, the user terminal extracts the entire broadcast screen as a region of interest, and extracts object information having a certain shape from the region of interest. Perform facial recognition to extract the performer output from the broadcast screen as a first object, and if the first object is not extracted, a moving object is extracted as a second object by comparing multiple broadcast screens, and the second object is not extracted. If not, there is provided a broadcast program recognition system based on image recognition, characterized in that a fixed object output from a broadcast screen is extracted as a third object and one of the first to third objects is transmitted to a service server as feature information.

In addition, according to another aspect of the present invention for achieving the above object, a broadcast output device for real-time output of a broadcast program broadcasted by a broadcaster, and an AR application are installed, and a camera is driven to broadcast according to the execution of the AR application. In addition to acquiring multiple broadcast screens from broadcast images output in real time from output devices, acquiring own location information and requesting AR content including feature information extracted from multiple broadcast screens, broadcast screen collection time and location information In addition to transmitting information to the service server, a user terminal that provides an AR service using AR content information authorized from the service server, and broadcast programs by broadcast time, and feature information by broadcast programs are registered and stored, and broadcast. Different AR content information according to the location of each program is registered and stored, and feature information of a broadcast program corresponding to the broadcast time and user terminal based on AR content request information approved from the user terminal in connection with the AR application executed in the user terminal Comprising a service server that compares the feature information received from and searches for a broadcast program corresponding to the feature information requested by the user, and transmits AR content information corresponding to the user location to the user terminal for the searched broadcast program, The characteristic information for recognizing a broadcast program is text information including a broadcaster logo or broadcast program name output on the broadcast screen, or consists of object information including a performer or object output on the broadcast screen, and the user terminal In the broadcast screen, a certain area based on both corners of the broadcast output device is extracted as a region of interest to recognize text, but when text is recognized in the first region of interest based on the first corner, the corresponding text information is used as feature information. Transmits to the server, and extracts text information by performing text recognition in the second ROI based on the second corner only when the text is not recognized in the first ROI based on the first corner, If text information is not extracted, the user terminal extracts the entire broadcast screen as an ROI, extracts object information having a certain shape from the ROI, and performs face recognition in the ROI to provide performers output from the broadcast screen. If the first object is extracted as an object, and if the first object is not extracted, a moving object is extracted as a second object by comparing multiple broadcast screens, and if the second object is not extracted, the fixed object output from the broadcast screen is a third object. An image recognition-based broadcast program recognition system is provided, which extracts as and transmits one of the first to third objects as feature information to a service server.

delete

delete

delete

delete

delete

According to the present invention, the user terminal captures a broadcast screen output from another external device without the inconvenience of connecting with the current broadcast output device or changing broadcast information and the need for the required specifications of the device, and broadcasting currently being broadcasted. By being able to automatically recognize a program, it is possible to activate an application application that recognizes such a broadcast program and provides a service.

Therefore, it is possible to easily recognize the broadcast program through the application installed in the user terminal without performing the editing process of adding or modifying specific information to the broadcast program (content) broadcasted by the broadcaster. The usability of the application can be further improved.

1 is a diagram illustrating a schematic configuration of a broadcast program recognition system based on image recognition according to a first embodiment of the present invention.
FIG. 2 is a block diagram showing the functionally separated user terminal 200 shown in FIG. 1;
3 is a view for explaining a broadcast image collection process in the image collection unit 220 shown in FIG. 2;
FIG. 4 is a diagram illustrating an ROI for extracting text information in the feature information extracting unit 230 shown in FIG. 2.
5 is a view for explaining the operation of the broadcast program recognition system based on image recognition shown in FIG. 1;

Since the embodiments described in the present invention and the configurations shown in the drawings are only preferred embodiments of the present invention, and do not represent all the technical spirit of the present invention, the scope of the present invention is limited to the embodiments and drawings described in the text. Should not be construed as limited by That is, since the embodiments can be variously changed and have various forms, the scope of the present invention should be understood to include equivalents capable of realizing the technical idea. In addition, since the object or effect presented in the present invention does not mean that a specific embodiment should include all of them or only those effects, the scope of the present invention should not be understood as being limited thereto.

All terms used herein have the same meaning as commonly understood by one of ordinary skill in the field to which the present invention belongs, unless otherwise defined. Terms defined in a commonly used dictionary should be construed as having the meaning of the related technology in context, and cannot be interpreted as having an ideal or excessively formal meaning that is not clearly defined in the present invention.

First, the present invention can be applied to various application applications that perform a service using an application application installed in a user terminal, that is, a smart phone, but perform a service using a broadcast program. An AR application that provides AR (Augmented Reality) services will be described.

1 is a diagram illustrating a schematic configuration of a broadcast program recognition system based on image recognition according to a first embodiment of the present invention.

Referring to FIG. 1, a broadcast program recognition system based on image recognition includes a broadcast output device 100, a user terminal 200, and a service server 300.

The broadcast output device 100 outputs a broadcast program currently broadcast from a broadcaster in real time in response to a selected channel. That is, broadcast information is displayed and output through a screen, and a voice corresponding to the broadcast screen is output.

The broadcast output device 100 may be various types of devices that output a program broadcasted by a broadcaster in real time, including a TV.

An AR (augmented reality) application is installed in the user terminal 200, and a broadcast output from the broadcast output device 100 is recognized according to execution of the AR application, and an AR service is provided using the recognized broadcast information.

At this time, the user terminal 200 drives the camera according to the execution of the AR application to collect a plurality of broadcast screens from the broadcast image output from the broadcast output device, extracts feature information from the plurality of broadcast screens, and extracts feature information and In addition to transmitting broadcast program recognition request information including broadcast screen collection time to the service server, AR service is provided using a broadcast program authorized from the service server.

The user terminal 200 may be various wireless communication terminals including a smart phone or tablet PC having a camera and image processing means and a screen display means.

The service server 300 provides the AR application to the user terminal 200 requesting the service, and checks the broadcast information corresponding to the characteristic object information of the broadcast screen applied from the user terminal 200 by interworking with the AR application. Broadcast information is provided to the user terminal 200.

In this case, the servers server 300 may interlock with each broadcaster to receive broadcast program information broadcasted at the current time.

In addition, the service server 300 pre-registers a broadcast program for each time and feature information for each broadcast program, and features information including at least one of a text or an object applied from the user terminal 200, and a screen from which the feature information is extracted. Based on the broadcast screen collection time corresponding to, the previously registered broadcast program is searched and broadcast information corresponding to the feature information is transmitted to the corresponding user terminal 200.

Here, the service server 300 may store only text, only objects, or both text and object information as characteristic information on the broadcast program, and the object information includes performers appearing in the broadcast program, Applicant information broadcasted on the screen in units of broadcast time may be stored.

That is, object information including all performers may be stored for one broadcast program, or object information including performers displayed on a broadcast screen at a corresponding time in a predetermined broadcast time unit for one broadcast program may be stored. This may be applied and stored in the same manner as object information in consideration of a case where various text information (broadcast program name) is output from one broadcast program.

FIG. 2 is a block diagram showing a functionally separated user terminal 200 shown in FIG. 1.

Referring to FIG. 2, the user terminal 200 includes a communication unit 210, an image recognition unit 220, a feature information extraction unit 230, an AR service unit 240, and a data memory 250.

The communication unit 210 transmits and receives information related to a broadcast program with the service server 300. The communication unit 210 transmits feature information to the service server 300 and receives broadcast information corresponding thereto from the service server 300.

The image recognition unit 220 drives a camera to collect a broadcast image currently displayed and output from the broadcast output device 100. In this case, the image recognition unit 220 may capture and collect a plurality of broadcast image screens at a predetermined time unit through the camera.

That is, the image collection unit 220 displays a screen recognized by the camera through the display window while setting the camera to the running state.

In this case, as shown in FIG. 3, the broadcast screen display area 201 is output on the display window of the user terminal 200, and the broadcast screen 101 displayed and output from the broadcast output device 100 is displayed in the broadcast screen display area ( By positioning within 201), it is possible to collect the broadcast screen 101 of the broadcast video that is broadcast for a certain time.

The feature information extracting unit 230 extracts preset feature information for recognizing a corresponding broadcast program from the broadcast screen collected by the image collecting unit 222.

In this case, the feature information extraction unit 230 may be configured to include one or both of the text extraction module 231 and the object extraction module 232.

That is, as feature information extracted from a broadcast screen, only text, only objects, or both text and objects can be extracted, and objects can be extracted only when text is not extracted.

The text extracting module 231 extracts a region of interest at a preset location to extract text from the entire broadcast screen, and extracts text existing in the region of interest.

In general, the screen information of the program currently being broadcast in Korea is displayed at the top of one side, for example, in the right or left corner, in a broadcaster log with a certain size, or the name of the program currently being broadcast, the name of the program scheduled to be broadcast later, or the name of the program to be broadcast Text information that can be recognized is displayed. 3 illustrates a shape in which a broadcaster logo is displayed on the upper right area of the right broadcast screen 101.

In consideration of this, the text extraction module 231 extracts the upper right or upper left corner area from the entire broadcast screen 101 using an edge detection method as shown in FIG. 4, and is adjacent to the corner based on the corner. The regions of interest 102 and 103 having a predetermined size are extracted, and text information corresponding to the program name is obtained from the extracted regions of interest.

In this case, when text information is not extracted from a corresponding broadcast screen, the text extracting module 231 similarly performs text information extraction until text information is extracted from another broadcast screen.

In addition, the text extraction module 231 extracts text from a region of interest of a predetermined region extracted based on both corners of the upper end of the broadcast output device 100, wherein the text is recognized in the first region of interest based on the first corner. In this case, the text information is set as feature information, and only when the text is not recognized in the first region of interest based on the first corner, text recognition is performed in the second region of interest based on the second corner to characterize the text information. It can be set as information.

Meanwhile, the object extraction module 232 extracts an object having a predetermined shape from the entire broadcast screen. The object may include performers or objects appearing in the program.

At this time, the object extraction module 232 extracts the entire broadcast screen as a region of interest, extracts object information having a certain shape from the region of interest, and performs face recognition in the region of interest to determine the performer output from the broadcast screen as a first object And, if the first object is not extracted, a moving object including a vehicle or animal is extracted as a second object by comparing a plurality of broadcast screens, and if the second object is not extracted, from a broadcast screen such as furniture or trees. The output fixed object may be extracted as a third object.

That is, the object extraction module 232 may determine one of a performer, a moving object, or a fixed object as feature information.

In addition, the object extraction module 232 may extract the performers using a face recognition algorithm, and set the face information of all the extracted performers with face recognition as object information, or multiple times among faces recognized in a plurality of collection screens. By sorting in the order of extracted face information, information on performers less than a certain number can be set as object information.

In addition, the object extraction module 232 extracts the performer output from the broadcast screen as a first object by performing face recognition in the region of interest, and extracts a moving object as a second object by comparing a plurality of broadcast screens. The fixed object output from is extracted as a third object, and all of the first to third objects may be set as feature information.

In this case, it is preferable that the service server 300 stores the object information for each broadcast program for each performer, a moving object, and a fixed object.

The AR service unit 240 provides various AR services according to a service program by using broadcast information currently broadcast through the broadcast output device 100.

For example, as an AR service using broadcast information, information such as furniture provided in a drama being watched or clothes or bags worn by performers may be provided.

The data memory 250 stores various types of information for providing an AR service for a broadcast program including an AR application provided from the service server 300.

Next, the operation of the broadcast program recognition system based on image recognition according to the present invention will be described with reference to FIG. 5.

First, broadcast program information provided from each broadcaster is registered in the service server 300 in advance, and feature information including broadcast programs for each time slot and text information or object information for each broadcast program is stored. In this case, the object information may be classified into an object type and stored as a performer (person), a moving object, and a fixed object.

Then, the user terminal 200 downloads and installs an AR application for AR service to the service server 300 (ST100, ST200).

While the AR application installed in the user terminal 200 is running, the user terminal 200 sets the camera to the running state and displays an image recognized through the camera. At this time, the broadcast screen display area 201 is output on the display window of the user terminal 200 as shown in FIG. 3, and the user sets the location of the camera provided in the user terminal 200 so that the broadcast output device 100 is photographed. When the broadcast output device 100 is included in the image display area output on the display window, the broadcast image is collected by performing a photographing process through the camera (ST300).

Here, the collection of broadcast images may be performed for a predetermined period of time, for example, 3 seconds or longer until the broadcast screen display area is cleared.

Then, the user terminal 200 generates a plurality of broadcast screens by capturing the collected broadcast images in a predetermined time unit (ST400). In this case, the camera may acquire a plurality of broadcast screens by performing a photographing operation in a predetermined time unit in step ST300.

Subsequently, the user terminal 200 extracts feature information for recognizing the broadcast from the broadcast screen, and transmits broadcast program request information including feature information and broadcast time to the service server 300 (ST500, ST600).

The user terminal 200 extracts a region of interest from the broadcast screen, and extracts text information or object information, or both, from the extracted region of interest.

In this case, the user terminal 200 may simultaneously extract the first and second ROI from the first and second corners of the top of the broadcast screen, and perform text recognition processing therefrom.

In addition, in most broadcasters, when an area advertising a broadcaster logo or program name is the first corner, the second region of interest of the second corner is only when text is not extracted from the first region of interest of the first corner. It can be implemented to perform a text extraction operation.

Meanwhile, the user terminal 200 extracts the entire broadcast screen as an ROI, and extracts object information having a predetermined shape from the ROI.

At this time, the user terminal 200 may perform face recognition in the region of interest and extract the performer as a first object. When face recognition is not performed, the moving object is extracted as a second object by comparing multiple broadcast screens. And, when there is no moving object, the fixed object may be extracted as a third object.

In addition, when a plurality of first objects are extracted, the user terminal 200 may count the number of times the first object is extracted from a plurality of broadcast screens, and set a certain number of performer information as a feature object in the order of the number of counts. . This applies equally to the second and third objects to set the feature object.

Also, the object information may include all of the first to third objects for one broadcast program request information.

Preferably, object information may be extracted only when text information is not extracted.

The service server 300 searches for a broadcast program corresponding to the broadcast time from the broadcast program request information applied from the user terminal 200, and includes object information and broadcast program request information previously registered for the searched broadcast program. A broadcast program corresponding to the feature information is determined by comparing the generated feature information (ST700).

The service server 300 transmits the broadcast program information determined in step ST700 to the corresponding smartphone 800 (ST800), and the user terminal 800 performs the AR service selected by the user using the received broadcast program information. Do (ST900).

On the other hand, in the present invention, in the step ST800, the service server 300 transmits content information for an AR service corresponding to the broadcast program to the user terminal 200 and corresponds to the AR application installed in the user terminal 200. AR service for content may be performed.

That is, the user terminal 200 additionally acquires its own current location information using a GPS receiver, etc., and provides AR content request information including feature information extracted from a plurality of broadcast screens and broadcast screen collection time and location information. It transmits to the service server 300 and provides an AR service using AR content information that is authorized from the service server 300 for this.

In addition, the service server 300 pre-registers and stores different AR content information according to the user location for each broadcast program, and based on the AR content request information applied from the user terminal, the feature information of the broadcast program corresponding to the broadcast time and the user The user terminal 200 searches for a broadcast program corresponding to the feature information requested by the user by comparing the feature information received from the terminal 200, and calls AR content information corresponding to the user location for the searched broadcast program. Transfer to.

In this case, the service server 300 pre-registers and stores AR service types for each broadcast program and different AR content information according to locations for each AR service type, and the user terminal 200 additionally includes AR service type information. The information is transmitted to the service server 300, and the service server 300 may transmit AR content information corresponding to the user location for the AR service type to the corresponding user terminal 200.

For example, the service server 300 transmits AR information related to Seoul to the user terminal 200 in the Seoul area for drama A (eg, advertisement or discount coupon for the Seoul area), and the user terminal in the Busan area for the same drama A To 200, AR information related to the Busan area is transmitted.

100: broadcast output device, 200: user terminal,
210: communication unit, 220: image collection unit,
230: feature information extraction unit, 231: text extraction module,
232: object extraction module, 240: AR service unit,
250: data memory, 300: service server.

Claims (6)

A broadcast output device that outputs a broadcast program broadcasted by a broadcaster in real time;
When the application is installed, the camera is driven according to the execution of the application to obtain multiple broadcast screens from the broadcast images output in real time from the broadcast output device, and extract feature information and broadcast screens by extracting feature information from multiple broadcast screens A user terminal that transmits broadcast program recognition request information including time to a service server, and provides a service provided by an application by using broadcast information authorized from the service server for this, and
Broadcast programs for each broadcast time and feature information for each broadcast program are registered and stored, and feature information of a broadcast program corresponding to the broadcast time and the user based on the broadcast program recognition request information that is applied from the user terminal in connection with an application executed in the user terminal Comprising a service server for comparing the feature information received from the terminal to search for broadcast information corresponding to the feature information requested by the user, and transmitting the searched broadcast information to the user terminal,
The characteristic information for recognizing a broadcast program is text information including a broadcaster logo or broadcast program name output on a broadcast screen, or consists of object information including a performer or object output on a broadcast screen,
The user terminal recognizes the text by extracting a certain area on the broadcast screen as a region of interest based on both corners of the top of the broadcast output device, but when the text is recognized in the first region of interest based on the first corner, the corresponding text information is Transmits as feature information to the service server, and extracts text information by performing text recognition in the second ROI based on the second corner only when the text is not recognized in the first ROI based on the first corner,
When the text information is not extracted, the user terminal extracts the entire broadcast screen as a region of interest, extracts object information having a certain shape from the region of interest, and performs face recognition in the region of interest to determine the performers output on the broadcast screen. It extracts as a first object, and if the first object is not extracted, a moving object is extracted as a second object by comparing multiple broadcast screens, and if the second object is not extracted, a fixed object output from the broadcast screen is selected as a third object. An image recognition-based broadcast program recognition system, comprising extracting as an object and transmitting one of the first to third objects as feature information to a service server.
A broadcast output device that outputs a broadcast program broadcasted by a broadcaster in real time;
The AR application is installed, and the camera is driven according to the execution of the AR application to acquire multiple broadcast screens from the broadcast images output in real time from the broadcast output device, acquires own location information, and extracts from multiple broadcast screens. A user terminal that transmits AR content request information including the received feature information and broadcast screen collection time and location information to the service server, and provides AR services using AR content information authorized from the service server for this,
In addition to the registration and storage of broadcasting programs by broadcasting time and characteristic information by broadcasting programs, different AR content information according to the location of each broadcasting program is registered and stored, and AR content approved from the user terminal by interlocking with the AR application executed in the user terminal Based on the request information, the feature information of the broadcast program corresponding to the broadcast time and feature information received from the user terminal are compared to search for a broadcast program corresponding to the feature information requested by the user, and the searched broadcast program corresponding to the user location It is configured to include a service server that transmits AR content information to a corresponding user terminal,
The characteristic information for recognizing a broadcast program is text information including a broadcaster logo or broadcast program name output on a broadcast screen, or consists of object information including a performer or object output on a broadcast screen,
The user terminal extracts a certain area based on both corners of the top of the broadcast output device from the broadcast screen as an ROI and recognizes the text, but when the text is recognized in the first ROI based on the first corner, the corresponding text information is Transmits as feature information to the service server, and extracts text information by performing text recognition in the second ROI based on the second corner only when the text is not recognized in the first ROI based on the first corner,
When the text information is not extracted, the user terminal extracts the entire broadcast screen as a region of interest, extracts object information having a certain shape from the region of interest, and performs face recognition in the region of interest to determine the performers output on the broadcast screen. It extracts as a first object, and if the first object is not extracted, a moving object is extracted as a second object by comparing multiple broadcast screens, and if the second object is not extracted, a fixed object output from the broadcast screen is selected as a third object. An image recognition-based broadcast program recognition system, comprising extracting as an object and transmitting one of the first to third objects as feature information to a service server.
delete delete delete delete

Images