WO2021118050A1 - 하이라이트 영상 자동 편집을 위한 컴퓨터 프로그램 - Google Patents
하이라이트 영상 자동 편집을 위한 컴퓨터 프로그램 Download PDFInfo
- Publication number
- WO2021118050A1 WO2021118050A1 PCT/KR2020/014636 KR2020014636W WO2021118050A1 WO 2021118050 A1 WO2021118050 A1 WO 2021118050A1 KR 2020014636 W KR2020014636 W KR 2020014636W WO 2021118050 A1 WO2021118050 A1 WO 2021118050A1
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- real
- user
- time content
- data
- content
- Prior art date
Links
Images
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N21/00—Selective content distribution, e.g. interactive television or video on demand [VOD]
- H04N21/80—Generation or processing of content or additional data by content creator independently of the distribution process; Content per se
- H04N21/85—Assembly of content; Generation of multimedia applications
- H04N21/854—Content authoring
- H04N21/8549—Creating video summaries, e.g. movie trailer
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR 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; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N5/00—Computing arrangements using knowledge-based models
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N21/00—Selective content distribution, e.g. interactive television or video on demand [VOD]
- H04N21/40—Client devices specifically adapted for the reception of or interaction with content, e.g. set-top-box [STB]; Operations thereof
- H04N21/45—Management operations performed by the client for facilitating the reception of or the interaction with the content or administrating data related to the end-user or to the client device itself, e.g. learning user preferences for recommending movies, resolving scheduling conflicts
- H04N21/466—Learning process for intelligent management, e.g. learning user preferences for recommending movies
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N21/00—Selective content distribution, e.g. interactive television or video on demand [VOD]
- H04N21/40—Client devices specifically adapted for the reception of or interaction with content, e.g. set-top-box [STB]; Operations thereof
- H04N21/45—Management operations performed by the client for facilitating the reception of or the interaction with the content or administrating data related to the end-user or to the client device itself, e.g. learning user preferences for recommending movies, resolving scheduling conflicts
- H04N21/466—Learning process for intelligent management, e.g. learning user preferences for recommending movies
- H04N21/4662—Learning process for intelligent management, e.g. learning user preferences for recommending movies characterized by learning algorithms
- H04N21/4666—Learning process for intelligent management, e.g. learning user preferences for recommending movies characterized by learning algorithms using neural networks, e.g. processing the feedback provided by the user
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N21/00—Selective content distribution, e.g. interactive television or video on demand [VOD]
- H04N21/80—Generation or processing of content or additional data by content creator independently of the distribution process; Content per se
- H04N21/83—Generation or processing of protective or descriptive data associated with content; Content structuring
- H04N21/845—Structuring of content, e.g. decomposing content into time segments
Definitions
- the present disclosure relates to automatic highlight image editing, and more particularly, to a computer program for automatically editing highlight images using an artificial neural network.
- a single creator creates and provides a highlight video by collecting, editing, and summarizing only the interesting parts (highlight parts) from the original video in order to attract the attention of the viewers.
- interesting parts highlight parts
- a separate professional editor is employed to produce a highlight video, manpower costs may occur.
- the criteria for selecting the highlight section from the original content may lack objectivity because the editor's subjective criteria are applied, and the quality of the highlight video generated may vary depending on the editor's competence.
- the present disclosure has been made in response to the above-described background art, and is intended to provide a computer program for automatically editing a highlight image.
- a computer program stored in a computer-readable storage medium is disclosed in an embodiment of the present disclosure for solving the above-described problems.
- the one or more processors When the computer program is executed on one or more processors, the one or more processors cause the one or more processors to perform the following operations for automatically editing a highlight image, the operations comprising: the real-time content from one or more user terminals viewing the real-time content Receiving one or more user reaction data for , identifying one or more highlight sections in the real-time content based on an analysis result of the user reaction data, and generating highlight content based on the one or more highlight sections may include.
- the user reaction data may be based on at least one of a natural language input or an action input of a user viewing the real-time content, and each user reaction data may be associated with at least a portion of the real-time content. .
- an operation of receiving external reaction data for at least one of the real-time content or the relevant information of the real-time content, and user reaction data for the real-time content and the external reaction data to the relevant information of the real-time content may further include generating a knowledge base for the related parties of the real-time content by storing in association with the .
- the external reaction data may be reaction data for at least one of the real-time content or a person concerned with the real-time content, and may be data received through a different channel from the user reaction data for the real-time content.
- the operation of generating a knowledge base for the related party of the real-time content by storing the user reaction data and the external reaction data for the real-time content in association with the related party information of the real-time content includes the external reaction
- the knowledge base is created by extracting external narrative information related to the real-time content or related party information of the real-time content from data, and extracting internal narrative information related to the real-time content or related party information of the real-time content from the user response data. It may include an action to create.
- the knowledge base is generated for each of the parties of the real-time content, and is used to identify a user reaction unique to each of the parties, and a semantic analysis of the externally narrated information or the internally narrated information. or based on at least one of quantitative analysis.
- the identifying one or more highlight sections in the real-time content based on the analysis result of the user reaction data includes the real-time based on at least one of semantic analysis or quantitative analysis of the user reaction data. It may include the operation of identifying one or more highlight sections in the content.
- the quantitative analysis may be performed based on a threshold determined based at least in part on the number of one or more user terminals viewing the real-time content.
- the method may further include the operation of pre-processing the user response data based on a knowledge base on the related information of the real-time content.
- the user reaction data further includes user reaction user identification information related to the user terminal that has transmitted the user reaction data, and based on a result of analysis of the user reaction data, one or more items in the real-time content
- the operation of identifying the highlight section includes an operation of assigning a weight to each user reaction data based on the user reaction user identification information and an analysis result of the user reaction data based on a weight assigned to each user reaction data. and identifying one or more highlight sections in the real-time content based on the analysis result.
- the operation of assigning a weight to each user reaction data based on the user reaction user identification information includes a highlight section prediction contribution rate for real-time content of the user who generated the user reaction data, the user's and assigning the weight to each of the user response data based on at least one of a viewing contribution rate for the content matched with the related information of the real-time content and a matching rate for the knowledge base of the user response data.
- the highlight section prediction contribution rate for the real-time content is information on whether user reaction data is included in the pre-produced highlight content in relation to the related party of the real-time content, and is included in the pre-produced highlight content. It may be calculated based on at least one of quantitative analysis and semantic analysis of the included user response data.
- the viewing contribution rate for the content matched with the relevant information of the real-time content is information about the participation history of each of the one or more user terminals for the content matched with the relevant information of the real-time content, wherein the matching It may be calculated based on the number of contents and the participation history.
- the matching rate of the user reaction data to the knowledge base may be calculated based on determination of a similarity between the user reaction data and at least one of external and internal narration information included in the knowledge base. .
- a method of automatically editing a highlight image performed by one or more processors of a computing device includes receiving one or more user response data for the real-time content from one or more user terminals viewing the real-time content, and identifying one or more highlight sections in the real-time content based on an analysis result of the user reaction data. and generating highlight content based on the one or more highlight sections.
- a server for automatically editing a highlight image includes a processor including one or more cores, a memory including program codes executable in the processor, and a network unit for transmitting and receiving data to and from a user terminal and an external server, and the processor is one for viewing real-time content Receive one or more user reaction data for the real-time content from the user terminal, identify one or more highlight sections in the real-time content based on an analysis result of the user reaction data, and based on the one or more highlight sections You can create highlight content.
- the present disclosure may provide a computer program for automatically editing a highlight image.
- FIG. 1 is a block diagram of a computing device for automatically editing a highlight image according to an embodiment of the present disclosure.
- FIG. 2 is an exemplary diagram for explaining a method of identifying a highlight section based on user reaction data related to real-time content related to an embodiment of the present disclosure.
- FIG. 3 is an exemplary diagram for explaining a method of identifying a highlight section based on user reaction data related to real-time content related to an embodiment of the present disclosure related to another embodiment of the present disclosure.
- FIG. 4 is a diagram exemplarily illustrating a method of allocating different weights to each of user response data related to an embodiment of the present disclosure.
- FIG. 5 is a schematic diagram illustrating a network function according to an embodiment of the present disclosure.
- FIG. 6 is a flowchart for generating highlight content related to an embodiment of the present disclosure.
- FIG. 7 is a diagram illustrating logic for implementing a method for generating highlight content related to an embodiment of the present disclosure.
- FIG. 8 depicts a simplified, general schematic diagram of an exemplary computing environment in which embodiments of the present disclosure may be implemented.
- a component can be, but is not limited to being, a process running on a processor, a processor, an object, a thread of execution, a program, and/or a computer.
- an application running on a computing device and the computing device may be a component.
- One or more components may reside within a processor and/or thread of execution.
- a component may be localized within one computer.
- a component may be distributed between two or more computers.
- these components can execute from various computer readable media having various data structures stored therein.
- Components may communicate via a network such as the Internet with another system, for example via a signal having one or more data packets (eg, data and/or signals from one component interacting with another component in a local system, distributed system, etc.) may communicate via local and/or remote processes depending on the data being transmitted).
- a network such as the Internet with another system
- one or more data packets eg, data and/or signals from one component interacting with another component in a local system, distributed system, etc.
- the network function of the present disclosure and an artificial neural network and a neural network may be used interchangeably.
- the streaming server, one or more user terminals and one or more streamer terminals and computing device 100 may transmit and receive data through an interconnection through a wireless and/or wired connection. have.
- the user terminal may be a terminal related to a plurality of users viewing real-time content transmitted from the streamer terminal.
- the user terminal may access the streaming server to view real-time content transmitted from the streamer terminal, and may transmit user response data related to the real-time content to the streaming server.
- the streamer terminal may be a terminal related to a streamer for transmitting real-time content to a plurality of user terminals.
- the streamer terminal may transmit various real-time content such as video content and audio content through a streaming server.
- the user terminal and the streamer terminal may include a personal computer (PC), a notebook (note book), a mobile terminal, a smart phone, a tablet PC, etc., It may include all types of terminals capable of accessing a wired/wireless network.
- PC personal computer
- notebook note book
- mobile terminal a smart phone
- tablet PC etc.
- It may include all types of terminals capable of accessing a wired/wireless network.
- the detailed description of the user terminal and the streamer terminal described above is only an example, and the present disclosure is not limited thereto.
- the streaming server may be a server that provides a service for transmitting and viewing real-time content through the Internet that interconnects real-time content related to each of one or more streamer terminals to each of one or more user terminals related to content viewing.
- the streaming server may receive user response data related to the corresponding real-time content from each user terminal in relation to the real-time content viewed by each of one or more user terminals and transmit it to the computing device 100 .
- the streaming server and the computing device 100 may include a streaming server in the computing device 100 according to an embodiment of the present disclosure to perform content transmission and highlight editing image providing functions in one integrated device.
- the computing device 100 may receive user response data for real-time content from one or more user terminals.
- the computing device 100 may generate a highlight edited image of the streamer's real-time content.
- the computing device 100 may identify highlight sections of interest to viewers in the streamer's real-time content, and generate a highlight edited image based on the identified highlight sections.
- the computing device 100 of the present disclosure may include all kinds of computing devices capable of calculating data in electronic form, for example, general computing devices such as personal computers, server computers, and mobile terminals (smartphones (smartphones) It may include a computing device having limited computing power, such as a smartphone), a tablet (tablet), and the like.
- a detailed description of a method in which the computing device 100 generates a highlight edited image from real-time content will be described later with reference to the following drawings.
- the computing device 100 may include a network unit 150 , a processor 110 , and a memory 130 .
- the above-described components are exemplary, and the scope of the present disclosure is not limited to the above-described components. That is, additional components may be included or some of the above-described components may be omitted depending on implementation aspects for the embodiments of the present disclosure.
- the computing device 100 may include a network unit 150 for transmitting and receiving data.
- the network unit 150 may transmit and receive information on real-time content and highlight content according to an embodiment of the present disclosure with at least one of one or more user terminals, a plurality of streamer terminals, and a streaming server.
- the computing device 100 may receive real-time content and user reaction data related to reactions of a plurality of user terminals viewing real-time content from a streaming server or a streamer terminal through the network unit 150 .
- the computing device 100 may transmit the highlight content generated in response to the real-time content to the streamer terminal through the network unit 150 .
- the detailed description of the information transmitted and received by the above-described network unit is only an example, and the present disclosure is not limited thereto.
- the memory 130 of the computing device 100 may store any type of information generated or determined by the processor 110 and any type of information received by the network unit 150 . Also, the memory 130 may store data supporting various functions of the computing device 100 . The memory 130 may store a plurality of application programs and commands driven by the computing device 100 . At least one of these applications may exist in the computing device 100 from the time of release. For example, the memory 130 may store real-time content and application programs for identifying a highlight section based on user reaction data related to the real-time content. As another example, the memory 130 may store a pre-trained neural network for generating a personalized knowledge base for each streamer of the computing device 100 . The description of the information stored in the above-described memory is only an example, and the present disclosure is not limited thereto.
- the present disclosure may consist of one or more cores, and a central processing unit (CPU) of a computing device, a general purpose graphics processing unit (GPGPU), a tensor processing unit ( It may include a processor for data analysis and deep learning, such as a tensor processing unit (TPU).
- the processor 120 may read a computer program stored in the memory 130 and perform data processing for machine learning according to an embodiment of the present disclosure.
- the processor 120 may perform an operation for learning the neural network.
- the processor 120 is for learning of a neural network such as processing input data for learning in deep learning (DL), extracting features from input data, calculating an error, and updating the weight of the neural network using backpropagation. calculations can be performed.
- DL deep learning
- At least one of a CPU, a GPGPU, and a TPU of the processor 120 may process learning of a network function.
- the CPU and the GPGPU can process learning of a network function and data classification using the network function together.
- learning of a network function and data classification using the network function may be processed by using the processors of a plurality of computing devices together.
- the computer program executed in the computing device according to an embodiment of the present disclosure may be a CPU, GPGPU, or TPU executable program.
- the computing device 100 may distribute and process a network function using at least one of a CPU, a GPGPU, and a TPU. Also, in an embodiment of the present disclosure, the computing device 100 may distribute and process a network function together with other computing devices. Descriptions of specific details regarding network function distributed processing of the computing device 100 are provided in US patent applications US15/161080 (filed on May 20, 2016) and US15/217475 (filed on July 22, 2016), which are incorporated herein by reference in their entireties. specifically discussed.
- the processor 120 may typically process the overall operation of the computing device 100 .
- the processor 120 processes signals, data, information, etc. input or output through the above-described components, or by driving an application program stored in the memory 130, to provide or process appropriate information or functions to the user terminal. have.
- the processor 110 may receive one or more user reaction data for the real-time content from one or more user terminals viewing the real-time content.
- the user response data may be based on at least one of a natural language input of a user of a user terminal viewing real-time content or an action input.
- the user reaction data may include user reaction user identification information related to the user terminal that has transmitted the user reaction data.
- the user response user identification information may be identification information for distinguishing each of one user terminal viewing real-time content.
- the user response data may be based on at least one of a natural language input input from a user of the user terminal in association with a part of real-time content, or an action input.
- the natural language input of the user response data may include a chatting input input from a user of one or more user terminals through a chat window related to real-time content.
- the action input of the user response data may mean an emoticon input, a donation input (eg, super chat, star balloon, etc.) input from a user of one or more user terminals, and a voice input.
- the user response data may be a chat input received from the first user terminal through a chat window in association with a first time point in the game content transmitted in real time by the first streamer terminal.
- the user response data may be a donation input received from the second user terminal in association with the second time point in the communication content transmitted in real time by the second streamer terminal.
- the processor 110 may receive user response data associated with a portion of the real-time content from each of one or more user terminals viewing the real-time content. In other words, the processor 110 identifies a point in time at which the user reaction data is received in the real-time content through the user reaction data having time correlation with the real-time content, and whether the user reaction data is received from any user terminal. can be identified.
- the processor 110 may receive external reaction data.
- the processor 110 may receive external response data related to a post posted in relation to real-time content by one or more user terminals from an external server.
- the external server is a server that posts information posted through one or more user terminals, and may be, for example, an SNS, a blog, an Internet broadcasting bulletin board, or the like.
- the detailed description of the above-described external server is only an example, and the present disclosure is not limited thereto.
- the processor 110 may receive external response data for at least one of real-time content and related information of real-time content from an external server. That is, the external reaction data is reaction data for at least one of real-time content or related information of real-time content, and a user's reaction input through a different channel (eg, a chat window of real-time content) from user reaction data for real-time content. data received over a different channel than the data).
- a different channel eg, a chat window of real-time content
- the related party information of real-time content is information for identifying each of a plurality of real-time content exposed to one or more user terminals, and may include, for example, at least one of broadcast content information and streamer identification information.
- the broadcast content information is information about a category of a broadcast transmitted by a corresponding streamer, and may include, for example, game broadcast content, eating broadcast content, outdoor broadcast content, communication broadcast content, and the like.
- the streamer identification information is information for identifying the streamer who transmits the content and the performer appearing in the content, information about the name, age, and gender of the streamer and the performer, ID, nickname, nickname of the streamer and the performer may include information about
- the detailed description of the broadcast content information and the streamer identification information included in the related information of the real-time content described above is only an example, and the present disclosure is not limited thereto.
- the related information of real-time content includes information that the streamer related to real-time content is 'A streamer', category information related to real-time content is 'invited singer live broadcast', and performer information related to real-time content is ' Singer Hong Gil-dong' may include information.
- the detailed description of the relationship information of the real-time content described above is only an example, and the present disclosure is not limited thereto.
- the processor 110 may use real-time content or relevant information of the real-time content as an input to the natural language processing model to determine the meaning and form of a sentence corresponding to the real-time content and the relevant information of the real-time content.
- the natural language processing model may be composed of one or more network functions learned through a training data set, and may be a model that performs semantic analysis and shape analysis on the input sentence by taking the input sentence as an input.
- the processor 110 may crawl the real-time content and related information of the real-time content as external response data from an external server based on the meaning and form of a sentence output in response to the real-time content and related information of the real-time content.
- the processor 110 when the related information of real-time content includes information that the streamer related to the real-time content is 'A streamer' and the performer information related to the real-time content is 'Singer Gil-dong Hong', the processor 110 provides the information These are input to the natural language processing model, and the meaning and form of the sentence corresponding to the corresponding information can be grasped. In this case, the processor 110 may receive, from an external server, a post written in relation to 'A streamer' or 'Hong Gil-dong' corresponding to the meaning and form of the sentence corresponding to the related information of the real-time content.
- the detailed description of the related information of the real-time content described above is only an example, and the present disclosure is not limited thereto.
- the processor 110 may obtain external response data corresponding to specific real-time content by selecting and crawling only real-time content from an external server through a natural language processing model or external user reaction data corresponding to relevant information of real-time content. have.
- the processor 110 may determine the user's reaction of the user terminal to the real-time content based on the external reaction data. Specifically, the processor 110 acquires external response data for at least one of real-time content or related information of real-time content, and grasps the meaning of the external response data through a natural language processing model, so as to obtain one or more real-time content viewers. It may be determined whether the user terminals have a positive or negative response to the real-time content.
- the processor 110 processes the sentence as an input of the natural language processing model, so that the sentence is It can be determined that the evaluation of Hong Gil-dong contains a negative meaning.
- Specific description of the above-described external reaction data is only an example, and the present disclosure is not limited thereto.
- the processor 110 may provide feedback information to a terminal concerned with real-time content corresponding to the external reaction data based on the external reaction data. Specifically, the processor 110 determines whether the user's response to the real-time content of the user terminal is positive or negative through the external reaction data, and provides feedback information to the terminal concerned with the real-time content based on the determined reaction. You can decide to send it.
- the processor 110 when receiving external response data 'These days, BJ true content is not interesting..' in relation to the second real-time content on SNS, the processor 110 processes the sentence as an input of the natural language processing model, It can be determined that the viewer's reaction to the recent content of BJ True (ie, the streamer) is negative. The processor 110 generates feedback information indicating that there is a need for improvement because the viewer's reaction to the latest content is negative based on the determined reaction, and the terminal concerned with the real-time content (eg, a terminal of BJ True that transmits real-time content) ) can be determined to be transmitted.
- the terminal concerned with the real-time content eg, a terminal of BJ True that transmits real-time content
- the processor 110 receives, from an external server, external reaction data related to a post posted in relation to the real-time content by users of one or more user terminals who have viewed the real-time content, thereby reacting the user terminals who have viewed the real-time content. can detect In addition, the processor 110 transmits feedback information to the relevant terminal of the corresponding real-time content based on the external reaction data of the user terminals who have viewed the real-time content, thereby allowing the person concerned with the real-time content to see the viewers' reaction to their content. By making it easy to recognize, it is possible to suggest the direction of improvement for the content. That is, the processor 110 may provide a streamer management effect for managing the streamer through feedback information based on external reaction data.
- the processor 110 may generate a knowledge base for a person concerned with real-time content. Specifically, the processor 110 may generate a knowledge base for the related party of the real-time content by storing the user reaction data and the external reaction data for the real-time content in association with the related information of the real-time content.
- a knowledge base is created corresponding to each stakeholder of the real-time content, and the relationship can be used to identify a user response unique to each.
- the knowledge base may be generated based on at least one of semantic analysis or quantitative analysis of external narration information or internal narration information. That is, the knowledge base may be a combination of keywords reflecting the individual characteristics of each streamer terminal that transmits a real-time video.
- the processor 110 extracts internal narrative information related to real-time content or relevant information of real-time content from the user reaction data based on at least one of semantic analysis or quantitative analysis of the user reaction data, and the extracted internal A knowledge base can be created by matching the narration information with the relevant information of the real-time content.
- the processor 110 is a natural language
- ' ⁇ ' may be a neologism with a negative connotation for the streamer wile (eg, a list of consonants in the sentence 'wield away').
- the processor 110 extracts internal narration information from the user response data through semantic analysis of the user response data, and matches the extracted internal narration information with the related information of the first real-time content, thereby providing information to the related party of the first real-time content.
- the detailed description of the above-described user reaction data and internal narrative information extracted from the user reaction data is an example, and the present disclosure is not limited thereto.
- user reaction data including information on the subject information of the first real-time content as “streamer wilde” and information on the word ‘godweil’ among a plurality of user reaction data received from each of one or more user terminals.
- the processor 110 determines that a plurality of user reaction data related to real-time content is more exposed through the chat window, and 'God Weil' corresponding to the user reaction data is internally descriptive information can be extracted as
- the processor 110 may generate a knowledge base on the related party of the first real-time content by matching the extracted internal narration information with the related party information of the first real-time content and storing the same.
- the processor 110 extracts text semantically suitable to the person concerned with the real-time content as internal narrative information through semantic analysis of a plurality of user response data received from one or more user terminals through a real-time chatting window related to the real-time content. to create a knowledge base.
- the processor 110 extracts the text frequently exposed in the real-time content as internal narrative information through quantitative analysis of a plurality of user response data received from one or more user terminals through a real-time chat window related to the real-time content as internal narrative information. You can create a base.
- the processor 110 extracts external narration information related to real-time content or related information of real-time content from the external reaction data based on at least one of semantic analysis or quantitative analysis of the external response data, and extracts the extracted external narration information.
- a knowledge base can be created by matching with the relevant information of real-time content.
- the processor 110 may extract external description information of 'Dr ⁇ '.
- 't ⁇ ' may be a buzzword with a positive meaning for BJ true (eg, 'true up'). That is, the processor 110 extracts external narration information from the external response data through semantic analysis of the external response data, and matches the extracted external narration information with the related information of the second real-time content, thereby providing information to the related party of the second real-time content.
- the processor 110 extracts a text semantically suitable for a person concerned with the real-time content as external narration information through semantic analysis of a plurality of external response data posted through one or more user terminals in relation to the real-time content, thereby extracting the knowledge base can create
- the processor 110 displays text frequently exposed in the external server in relation to the real-time content through quantitative analysis of a plurality of external reaction data posted through one or more user terminals to the external server related to the real-time content as external narrative information. can be extracted to create a knowledge base.
- the processor 110 receives user reaction data and external reaction data from each of the postings of the real-time chat window and the external server in relation to specific real-time content, and performs semantic analysis and quantitative analysis for each of the user reaction data and the external reaction data.
- the processor 110 extracts internal narration information and external narration information from user response data and external response data, respectively, it is possible to generate a knowledge base corresponding to a person concerned with a specific real-time content.
- the processor 110 may identify one or more highlight sections in real-time content based on the analysis result of the user reaction data. Specifically, the processor 110 may identify one or more highlight sections in the real-time content based on at least one of semantic analysis or quantitative analysis of the user reaction data.
- the processor 110 may generate one or more real-time sub-contents by dividing the real-time content into predetermined time intervals. Also, the processor 110 may identify one or more highlight sections in the real-time content based on the semantic analysis of the user response data received in response to each real-time sub content. The processor 110 performs semantic analysis on one or more user response data included in each of the one or more real-time sub-contents, and as a result of the semantic analysis, corresponds to the sub-content with the maximum number of user response data including a positive meaning in each sub-content. The point of view can be identified as a highlight section.
- the processor 110 divides the first real-time content into five periods and divides the first real-time content into five real-time content.
- Sub-contents ie, first to fifth real-time sub-contents may be generated.
- the processor 110 may identify a time point (ie, 1 to 2 minutes) corresponding to the first real-time sub-content as a highlight section. have.
- a time point ie, 1 to 2 minutes
- the detailed description of the above-described real-time content, the predetermined time period, and the number of user reaction data is merely an example, and the present disclosure is not limited thereto.
- the processor 110 may identify one or more highlight sections in real-time content based on quantitative analysis of user reaction data. Specifically, the processor 110 may perform quantitative analysis on the number of user response data received in association with the real-time content, and identify one or more highlight sections in the real-time content based on the quantitative analysis result. The quantitative analysis may be performed based on a threshold determined based at least in part on the number of one or more user terminals viewing the real-time content.
- the processor 110 may generate one or more real-time sub-contents by dividing the real-time content into predetermined time intervals. Also, the processor 110 may identify one or more real-time sub-contents in which the number of user response data received corresponding to each real-time sub-content exceeds a threshold value among the one or more real-time sub-contents as a highlight section.
- a threshold value serving as a criterion for identifying a highlight section through quantitative analysis based on the number of user reaction data may be determined based at least in part on the number of one or more user terminals viewing real-time content.
- the threshold value may have a positive correlation with the number of user terminals viewing the real-time content at a time point corresponding to each real-time sub-content.
- the threshold which is a criterion for identifying a highlight section through quantitative analysis based on the number of user response data, can be variably adjusted based on the number of user terminals viewing real-time content at a time point corresponding to each sub-content.
- the threshold may be preset to 50% of the number of user terminals viewing real-time content at a time point corresponding to each real-time sub-content.
- the threshold value may be 50, and the second real-time content is a time period of 2-3 minutes.
- the threshold value may be 25. That is, the threshold value was 50 as there were 100 viewers in the 1 to 2 minute interval of the real-time content, but the threshold value can be adjusted to 25 as the number of viewers decreases to 50 in the 2-3 minute interval, which is a later point in time. .
- the above-described real-time content, sub-content, and specific values of the threshold value and the number of user terminals viewed at a time corresponding to the sub-content are merely examples, and the present disclosure is not limited thereto.
- the processor 110 converts the first real-time content to the real-time sub content of 14 sections. (that is, the first real-time sub-contents to the fourteenth real-time sub-contents).
- the number of one or more user terminals viewing the real-time content at a time point corresponding to each of the first to fourth real-time sub-contents may be flexibly changed.
- a threshold value for each real-time sub-content may be variably adjusted according to a fluid change in the number of user terminals viewing the real-time content at a time point corresponding to each real-time sub-content.
- the number of a plurality of user response data received from one or more user terminals during the time period corresponding to the A section 210 is 90, and the corresponding section
- the processor 110 determines that the number of user response data received in the corresponding section exceeds the threshold value
- Section A 210 may be identified as a highlight section.
- the processor 110 controls the user received in the corresponding section. It is determined that the number of response data is less than or equal to the threshold value, so that section B 220 may not be identified as a highlight section.
- the number of a plurality of user response data received from one or more user terminals during the time period corresponding to the C section 230 is 32
- the processor 110 determines that the number of user response data received in the section exceeds the threshold value. It is determined that the section C 230 can be identified as a highlight section.
- the B section 220 may not be included in the highlight section despite receiving relatively more user response data compared to the A section 210 and the C section 230 .
- the processor 110 does not simply identify a section with a large number of user response data received from one or more user terminals as a highlight section, but identifies a section with a high number of user response data compared to the viewer as a highlight section in each section. can That is, even though relatively little user response data is received in section C 230 , the processor 110 may determine that a large amount of user response data has been received in relation to the number of viewers at that time, and may identify it as a highlight section.
- the processor 110 identifies a change amount of user reaction data for each section of real-time content, and identifies a highlight section based on a point in time when the identified change amount of user reaction data exceeds a threshold change amount.
- the threshold change amount may be determined based at least in part on the change amount of the number of one or more user terminals viewing the real-time content.
- the processor 110 may generate one or more real-time sub-contents by dividing the real-time content into predetermined time intervals.
- the processor 110 may identify one or more real-time sub-contents in which a change amount of user response data received corresponding to each real-time sub-content among one or more real-time sub content exceeds a threshold change amount as one or more highlight sections.
- the processor 110 converts the second real-time content into real-time sub-contents of 14 sections. (that is, the first real-time sub-contents to the fourteenth real-time sub-contents).
- the processor 110 may calculate a change amount of the user reaction data based on the user reaction data received from each of the real-time sub-contents.
- the threshold change amount in each section may be variably adjusted.
- the processor 110 determines the user corresponding to the section It is determined that the amount of change in the response data exceeds the threshold change amount, so that the A section 310 and the B section 320 can be identified as the highlight section. In an additional embodiment, the processor 110 may identify only the section B 320 as the highlight section.
- the amount of change in the plurality of user response data received from one or more user terminals during the time period corresponding to the C section 330 and the D section 340 is 2
- the processor 110 determines that the change amount of the user response data at the time point corresponding to the section is less than the threshold change amount C section 330 and D section 340 ) may not be identified as a highlight section.
- the highlight It may not be included in the section.
- the processor 110 does not identify a highlight section based on the number of user response data in each section, but identifies a section in which the amount of change in the user response data is large in each section as a highlight section in relation to the amount of change of the viewer in each section.
- a different number of viewers may be retained depending on the awareness of the streamer. For example, when user A, a popular streamer, transmits real-time content, an average of 3,000 user terminals can view the content, while user B, an unpopular streamer, receives an average of 50 when sending real-time content. A corresponding user terminal may view the corresponding content. Accordingly, when identifying a highlight section based on the number of user response data received from a user terminal that simply watches real-time content, the highlight section is identified from the real-time content of each streamer as each streamer has a different number of viewers. There is a possibility that the standards for doing so may differ.
- the processor 110 identifies the highlight section through the amount of change in user response data at each point in time of the real-time content, so that the same criterion for identifying the highlight section from each of the real-time contents of a plurality of streamers having different numbers of viewers is used. can present Accordingly, the processor 110 may identify the highlight section with the same reference for each of the real-time contents of various streamers such as popular streamers and unpopular streamers.
- the processor 110 may perform pre-processing on user response data received in response to real-time content.
- the preprocessing of the user reaction data may mean assigning different weights to the user reaction data received from each of one or more user terminals through a chat window related to real-time content.
- the processor 110 may assign a weight to each user reaction data based on the user reaction user identification information.
- the processor 110 determines at least one of a highlight section prediction contribution rate for the real-time content of the user who generated the user reaction data, a viewing contribution rate for the content matched with the related information of the user's real-time content, and a matching rate for the knowledge base of the viewer reaction data.
- a weight may be assigned to each user response data based on one.
- the highlight section prediction contribution rate for real-time content is information on whether user reaction data is included in the pre-produced highlight content in relation to the person involved in the real-time content, and quantitative analysis of the user reaction data included in the pre-produced highlight content and semantic analysis.
- the processor 110 controls the first user terminal.
- a higher weight may be given to the first user response data of the first user terminal to the first real-time content based on the identification information.
- the process of generating a highlight image by identifying a highlight section from the real-time content pre-produced at the previous time point of the streamer A, including the user reaction data of the first user terminal means that the user reaction data of the first user terminal is the previous real-time It may contribute to identifying a highlight section in the content.
- the processor 110 determines that the user of the first user terminal is a user who contributes a lot to detecting an interesting part in the real-time content, so that the user reaction data related to the first user terminal is more than the user reaction data related to other user terminals.
- a relatively high weight can be assigned.
- the degree of weight given to the user response data related to the first user terminal may be based on at least one of quantitative analysis and semantic analysis.
- the processor 110 may assign different weights to each of the user reaction data through semantic analysis of each user reaction data.
- the processor 110 may perform semantic analysis on the first response data of the first user terminal, and as a result of the semantic analysis, the content included in the first response data (ie, 'aaaaaaaa') is real-time content.
- a weight of '0' may be assigned to the first response data 410 by identifying it as a meaningless sentence not related to . That is, the processor 110 determines that the meaning of the first reaction data 410 is not related to the real-time content, and assigns a weight of '0' to the first reaction data 410 to remove it as noise. By doing so, it is possible to prevent the first reaction data from being counted in the quantitative analysis process.
- the processor 110 may perform semantic analysis on the second reaction data 420 of the second user terminal, and as a result of the semantic analysis, the content included in the second reaction data 420 (that is, 'this is a highlight') 'Wow') means that the highlight section is more appropriately predicted, and a relatively high weight of '3' may be assigned to the second response data 420 .
- the processor 110 may perform semantic analysis on the third reaction data 430 of the third user terminal, and as a result of the semantic analysis, the content included in the third reaction data 430 (ie, 'Awesome! !') may be identified as contributing to the highlight section identification, and a weight of '2.5' may be assigned to the third response data 430 .
- the processor 110 may perform semantic analysis on the fourth reaction data 440 of the fourth user terminal, and as a result of the semantic analysis, the content included in the fourth reaction data 440 (ie, 'What is it? ') may be identified as having a meaning opposite to the highlight section prediction, and a weight of '-1' may be assigned to the fourth response data 440 .
- the processor 110 may perform semantic analysis on the fifth response data 450 of the fifth user terminal, and as a result of the semantic analysis, the contents included in the fifth response data 450 (that is, 'that is what anyone is doing') ') as having the opposite meaning to the highlight section prediction, so that a weight of '-1' may be assigned to the fifth response data 450 .
- the processor 110 may remove meaningless user reaction data as noise based on the result of semantic analysis of each of a plurality of user reaction data corresponding to each user terminal included in the pre-made highlight image.
- a relatively high weight may be given to a user terminal that has predicted more clearly, and a negative weight may be given to user response data having a meaning opposite to that of the highlight section.
- the detailed description of the above-described user response data of each user terminal and a weight given corresponding thereto is only an example, and the present disclosure is not limited thereto.
- the processor 110 may Through quantitative analysis of the reaction data, a higher weight may be given to the user reaction data of the fourth user terminal, to which the user reaction data of the third user terminal is inputted more times in the highlight section than the user reaction data of the third user terminal. That is, user reaction data of a user terminal that has inputted user reaction data a greater number of times at a time point corresponding to the highlight section through quantitative analysis of a plurality of user reaction data corresponding to each user terminal included in the pre-produced highlight image. can be given a relatively high weight.
- the detailed description of the above-described user response data of each user terminal and a weight given corresponding thereto is only an example, and the present disclosure is not limited thereto.
- the viewing contribution rate for the content matched with the related information of the real-time content is information about the participation history of each of one or more user terminals for the content matched with the relationship between the real-time content information, and is based on the number of matched content and the participation history can be calculated by
- a user terminal that frequently views content related to a specific streamer may be more likely to recognize an interesting part in a real-time content image of the streamer. Accordingly, the processor 110 determines that each of the plurality of user terminals views the real-time content through the viewing contribution rate for the content matched with the relevant information (ie, information for identifying a particular streamer) of the particular real-time content by each user terminal. Different weights can be given to the user response data of .
- the processor 110 may give a relatively high weight to the first user terminal. That is, a different weight may be assigned to the user response data of each user terminal based on the degree of the participation history of each user terminal for the content of the previous point in time related to the streamer A.
- the participation history of the user terminal for the contents matched with the related information of the specific real-time content may be based on whether the user reaction data received from each user terminal is included in the content that was matched with the A streamer and transmitted at the previous point in time. have.
- the user terminal that has performed chatting input or action input corresponding to user response data in the previous content transmission process related to streamer A has a history of participating in content matched with streamer A by the processor 110 .
- the detailed description of the content matched to the above-described specific streamer and the participation history of the specific user terminal is merely an example, and the present disclosure is not limited thereto.
- the matching rate of the viewer reaction data to the knowledge base may be calculated based on the determination of a similarity between the user reaction data and at least one of external narration information and internal narration information included in the knowledge base.
- the processor 110 may generate a personalized knowledge base for the person concerned with the real-time content, and view the real-time content with at least one of external narration information and internal narration information included in the knowledge base for the person concerned with the real-time content.
- a different weight may be assigned to each of the plurality of user response data received from each of the one or more user terminals viewing real-time content based on the similarity determination of each of the plurality of user response data received from the one or more user terminals.
- the knowledge base of the person concerned with the first real-time content includes internal narration information of “Godweil” and external narration information of “T ⁇ ”
- a plurality of pieces received from one or more user terminals viewing the first real-time content A high weight may be given to user terminals that have inputted user response data including information such as “Godweil” or “Duh” among the user response data of .
- the knowledge base personalized for each streamer may be composed of buzzwords or new words corresponding to the streamer, the user terminal that has input chatting input and action input related to internal narrative information and external narrative information included in the knowledge base
- the user may be a user with a high degree of understanding of the content of the corresponding streamer.
- the processor 110 may give a high weight to the user reaction data of the user terminals that input the user reaction data included in the knowledge base of the person concerned with the real-time content in the process of watching the real-time content.
- a personalized knowledge base is used for each of the stakeholders of a plurality of real-time contents
- the specific user reaction data is not valid as user reaction data of the first real-time content, but as user reaction data of the second real-time content can be determined to be valid.
- each streamer may use a personalized knowledge base to give different weights to each of a plurality of user reaction data.
- the detailed description of the knowledge base of each person concerned is merely an example, and the present disclosure is not limited thereto.
- the processor 110 performs pre-processing of adjusting the weights for the user reaction data received in relation to the real-time content and assigns different weights to each of the plurality of user reaction data for identifying the highlight section, thereby making it more sophisticated than the highlight section. can be clearly identified.
- the processor 110 generates an analysis result for the user reaction data based on a weight assigned to each user reaction data, and selects one or more highlight sections in real-time content based on the analysis result. can be identified. Specifically, the processor 110 assigns different weights to each of the user reaction data through pre-processing of the user reaction data, and generates an analysis result for each of the user reaction data to which the different weights have been given, so as to highlight one or more highlights in real-time content. section can be identified.
- the first user reaction data and the second user reaction data are received from each of the first user terminal and the second user terminal in relation to the first real-time content, and a weight assigned to each received user reaction data is '
- the processor 110 may perform quantitative analysis based on user reaction data based on a weight assigned to each user reaction data. That is, a plurality of user reaction data received in the real-time content viewing environment may be recognized as different weights according to weights assigned to each user reaction data in the quantitative analysis process. For example, in the quantitative analysis process, as the weight given to the second user response data is '2', the second user response data may be recognized as having a weight twice that of the first user response data with a weight of '1'. have.
- each user reaction data is different in the quantitative analysis process according to the weight assigned to each user reaction data. can be recognized as a proportion.
- the processor 110 may generate highlight content based on one or more highlight sections. Specifically, the processor 110 may generate highlight content including one or more highlight sections identified in the real-time content.
- the first real-time content corresponds to 5 minutes, and the predetermined time interval is 1 minute
- the first real-time content includes five real-time sub-contents (that is, the first real-time sub-content to the fifth real-time sub content).
- the processor 110 corresponds to the first real-time sub content and the fourth real-time sub content. It is possible to create highlight content including a time period. That is, the highlight content generated by the processor 110 may be composed of content at points in time corresponding to 0 to 1 minute and 3 to 4 minutes in the real-time content.
- the description of specific times of the above-described real-time content and real-time sub content is merely an example, and the present disclosure is not limited thereto.
- the processor 110 may generate highlight content including at least partially the identified one or more highlight sections. Specifically, the processor 110 may generate highlight content by removing at least a portion of one or more highlight sections identified based on the reproduction time setting information or adding one or more supplementary highlight sections.
- the reproduction time setting information is information for limiting the reproduction time of the highlight content generated from the real-time content, and may be predetermined based on a user's input requesting the production of the highlight content for the real-time content. For example, when the reproduction time setting information is predetermined by the user as 5 minutes, the processor 110 may generate highlight content having a reproduction time of 5 minutes from the real-time content. That is, the reproduction time setting information may be predetermined in response to a user's request to produce highlight content having a specific reproduction time in relation to the real-time content.
- the processor 110 removes at least some sections from the identified one or more highlight sections based on the comparison of the play time information of the identified one or more highlight sections and the predetermined play time setting information, or adds a supplementary highlight section You can create highlight content.
- the processor Step 110 may generate highlight content by removing at least some of the identified three highlight sections.
- the highlighted section partially removed from among the identified three highlight sections may be based on at least one of semantic analysis and quantitative analysis of user response data corresponding to each section.
- the processor 110 may generate highlight content in which at least some of the three highlight sections are removed based on the results of semantic analysis and quantitative analysis on a plurality of user response data corresponding to each of the identified three highlight sections. .
- the processor 110 controls the user from among the identified three highlight sections.
- the highlight content may be generated by removing the section C (ie, the section corresponding to the eighth real-time sub-content) that is the section in which the reception of response data is least.
- the processor 110 may identify a supplementary highlight section in sections other than the four identified highlight sections among the second real-time contents.
- a section identified as a supplementary highlight section in a section other than the identified highlight section may be based on at least one of semantic analysis and quantitative analysis of user response data corresponding to each section. That is, the processor 110 may identify the supplementary highlight section based on the results of semantic analysis and quantitative analysis on a plurality of user response data corresponding to each of the identified four highlight sections. For example, as a result of quantitative analysis of user response data corresponding to each of the identified sections other than the four highlighted sections, the first real-time sub content is the section in which the number of user response data is the largest in sections other than the identified four highlight sections.
- the processor 110 When it is determined as section A corresponding to , the processor 110 identifies section A as a supplementary highlight section, and adds the supplementary highlight section to the four highlight sections to generate highlight content.
- the supplemental highlight section may be added to the four highlight sections based on the viewpoint in the real-time content. That is, the processor 110 generates a section A (ie, a section corresponding to the first real-time sub-content), a section D (a section corresponding to the second real-time sub-content), and section E (a section corresponding to the fourth real-time sub-content).
- section F section corresponding to the fifth real-time sub content
- section H section corresponding to the ninth real-time sub content
- the description of specific numerical values for the above-described one or more highlight sections, the predetermined reproduction time setting information, and the number of user response data is merely an example, and the present disclosure is not limited thereto.
- the processor 110 may generate highlight content by removing or adding a specific section from the identified highlight section, it may be possible to produce highlight content having a specific reproduction time that meets the user's needs. That is, since it is possible to reproduce the highlight content corresponding to the playback time length desired by the user in the real-time content, the convenience of the user who wants to produce a highlight image may be increased.
- FIG. 5 is a schematic diagram illustrating a network function related to an embodiment of the present disclosure.
- a neural network may be composed of a set of interconnected computational units, which may generally be referred to as nodes. These nodes may also be referred to as neurons.
- a neural network is configured by including at least one or more nodes. Nodes (or neurons) constituting the neural networks may be interconnected by one or more links.
- one or more nodes connected through a link may relatively form a relationship between an input node and an output node.
- the concepts of an input node and an output node are relative, and any node in an output node relationship with respect to one node may be in an input node relationship in a relationship with another node, and vice versa.
- an input node-to-output node relationship may be created around a link.
- One or more output nodes may be connected to one input node through a link, and vice versa.
- the value of the output node may be determined based on data input to the input node.
- a node interconnecting the input node and the output node may have a weight.
- the weight may be variable, and may be changed by the user or algorithm in order for the neural network to perform a desired function. For example, when one or more input nodes are interconnected to one output node by respective links, the output node sets values input to input nodes connected to the output node and links corresponding to the respective input nodes. An output node value may be determined based on the weight.
- one or more nodes are interconnected through one or more links to form an input node and an output node relationship within the neural network.
- the characteristics of the neural network may be determined according to the number of nodes and links in the neural network, the correlation between the nodes and the links, and the value of a weight assigned to each of the links. For example, when the same number of nodes and links exist and there are two neural networks having different weight values between the links, the two neural networks may be recognized as different from each other.
- a neural network may include one or more nodes. Some of the nodes constituting the neural network may configure one layer based on distances from the initial input node. For example, a set of nodes having a distance n from the initial input node may constitute n layers. The distance from the initial input node may be defined by the minimum number of links that must be traversed to reach the corresponding node from the initial input node. However, the definition of such a layer is arbitrary for description, and the order of the layer in the neural network may be defined in a different way from the above. For example, a layer of nodes may be defined by a distance from the final output node.
- the initial input node may refer to one or more nodes to which data is directly input without going through a link in a relationship with other nodes among nodes in the neural network.
- it may mean nodes that other input nodes connected by a link do not have.
- the final output node may refer to one or more nodes that do not have an output node in relation to other nodes among nodes in the neural network.
- the hidden node may mean nodes constituting the neural network other than the first input node and the last output node.
- the neural network according to an embodiment of the present disclosure may be a neural network in which the number of nodes in the input layer may be the same as the number of nodes in the output layer, and the number of nodes decreases and then increases again as progresses from the input layer to the hidden layer.
- the neural network according to another embodiment of the present disclosure may be a neural network in which the number of nodes in the input layer may be less than the number of nodes in the output layer, and the number of nodes decreases as the number of nodes progresses from the input layer to the hidden layer. have.
- the neural network according to another embodiment of the present disclosure may be a neural network in which the number of nodes in the input layer may be greater than the number of nodes in the output layer, and the number of nodes increases as the input layer progresses to the hidden layer.
- the neural network according to another embodiment of the present disclosure may be a neural network in a combined form of the aforementioned neural networks.
- a deep neural network may refer to a neural network including a plurality of hidden layers in addition to an input layer and an output layer.
- Deep neural networks can be used to identify the latent structures of data. In other words, it can identify the potential structure of photos, texts, videos, voices, and music (for example, what objects are in the photos, what the text and emotions are, what the texts and emotions are, etc.) .
- Deep neural networks include convolutional neural networks (CNNs), recurrent neural networks (RNNs), auto encoders, generative adversarial networks (GANs), and restricted Boltzmann machines (RBMs). boltzmann machine), a deep belief network (DBN), a Q network, a U network, a Siamese network, and the like.
- CNNs convolutional neural networks
- RNNs recurrent neural networks
- GANs generative adversarial networks
- RBMs restricted Boltzmann machines
- the network function may include an auto-encoder.
- the auto-encoder may be a kind of artificial neural network for outputting output data similar to input data.
- the auto encoder may include at least one hidden layer, and an odd number of hidden layers may be disposed between the input/output layers.
- the number of nodes in each layer may be reduced from the number of nodes of the input layer to an intermediate layer called the bottleneck layer (encoding), and then expanded symmetrically with reduction from the bottleneck layer to the output layer (symmetrically with the input layer).
- the nodes of the dimension reduction layer and the dimension reconstruction layer may or may not be symmetrical.
- the auto-encoder can perform non-linear dimensionality reduction.
- the number of input layers and output layers may correspond to the number of sensors remaining after pre-processing of input data.
- the number of nodes of the hidden layer included in the encoder may have a structure that decreases as the distance from the input layer increases. If the number of nodes in the bottleneck layer (the layer with the fewest nodes between the encoder and the decoder) is too small, a sufficient amount of information may not be conveyed, so a certain number or more (e.g., more than half of the input layer, etc.) ) may be maintained.
- the neural network may be trained by at least one of supervised learning, unsupervised learning, and semi-supervised learning.
- the training of the neural network is to minimize the error in the output.
- iteratively inputs the training data to the neural network, calculates the output of the neural network and the target error for the training data, and calculates the error of the neural network from the output layer of the neural network to the input layer in the direction to reduce the error. It is a process of updating the weight of each node in the neural network by backpropagation in the direction.
- teacher learning learning data in which the correct answer is labeled in each learning data is used (ie, labeled learning data), and in the case of comparative learning, the correct answer may not be labeled in each learning data.
- learning data in the case of teacher learning related to data classification may be data in which categories are labeled in each of the learning data.
- Labeled training data is input to the neural network, and an error can be calculated by comparing the output (category) of the neural network with the label of the training data.
- an error may be calculated by comparing the input training data with the neural network output. The calculated error is back propagated in the reverse direction (ie, from the output layer to the input layer) in the neural network, and the connection weight of each node of each layer of the neural network may be updated according to the back propagation.
- a change amount of the connection weight of each node to be updated may be determined according to a learning rate.
- the computation of the neural network on the input data and the backpropagation of errors can constitute a learning cycle (epoch).
- the learning rate may be applied differently according to the number of repetitions of the learning cycle of the neural network. For example, a high learning rate can be used at the early stage of training of a neural network to increase efficiency by allowing the neural network to quickly acquire a certain level of performance, and a low learning rate can be used to increase accuracy at a later stage of learning.
- the training data may be a subset of real data (that is, data to be processed using the trained neural network), and thus the error on the training data is reduced, but the error on the real data is reduced.
- Overfitting is a phenomenon in which errors on actual data increase by over-learning on training data as described above. For example, a phenomenon in which a neural network that has learned a cat by showing a yellow cat does not recognize that it is a cat when it sees a cat other than yellow may be a type of overfitting. Overfitting can act as a cause of increasing errors in machine learning algorithms. In order to prevent such overfitting, various optimization methods can be used. In order to prevent overfitting, methods such as increasing training data, regularization, and dropout in which a part of nodes in the network are omitted in the process of learning may be applied.
- a computer-readable medium storing a data structure is disclosed according to an embodiment of the present disclosure.
- the data structure may refer to the organization, management, and storage of data that enables efficient access and modification of data.
- a data structure may refer to an organization of data to solve a specific problem (eg, data retrieval, data storage, and data modification in the shortest time).
- a data structure may be defined as a physical or logical relationship between data elements designed to support a particular data processing function.
- the logical relationship between data elements may include a connection relationship between data elements that a user thinks.
- Physical relationships between data elements may include actual relationships between data elements physically stored on a computer-readable storage medium (eg, a hard disk).
- a data structure may specifically include a set of data, relationships between data, and functions or instructions applicable to data.
- a computing device can perform an operation while using the resources of the computing device to a minimum.
- the computing device may increase the efficiency of operations, reads, insertions, deletions, comparisons, exchanges, and searches through effectively designed data structures.
- a data structure may be classified into a linear data structure and a non-linear data structure according to the type of the data structure.
- the linear data structure may be a structure in which only one piece of data is connected after one piece of data.
- the linear data structure may include a list, a stack, a queue, and a deck.
- a list may mean a set of data in which an order exists internally.
- the list may include a linked list.
- the linked list may be a data structure in which data is linked in such a way that each data is linked in a line with a pointer. In a linked list, a pointer may contain information about a link with the next or previous data.
- a linked list may be expressed as a single linked list, a doubly linked list, or a circularly linked list according to a shape.
- a stack can be a data enumeration structure with limited access to data.
- a stack can be a linear data structure in which data can be processed (eg, inserted or deleted) at only one end of the data structure.
- the data stored in the stack may be a data structure LIFO-Last in First Out.
- a queue is a data listing structure that allows limited access to data. Unlike a stack, the queue may be a data structure that comes out later (FIFO-First in First Out) as data stored later.
- a deck can be a data structure that can process data at either end of the data structure.
- the nonlinear data structure may be a structure in which a plurality of data is connected after one data.
- the nonlinear data structure may include a graph data structure.
- a graph data structure may be defined as a vertex and an edge, and the edge may include a line connecting two different vertices.
- a graph data structure may include a tree data structure.
- the tree data structure may be a data structure in which one path connects two different vertices among a plurality of vertices included in the tree. That is, it may be a data structure that does not form a loop in the graph data structure.
- the data structure may include a neural network.
- the data structure including the neural network may be stored in a computer-readable medium.
- Data structures, including neural networks may also include data input to the neural network, weights of the neural network, hyperparameters of the neural network, data obtained from the neural network, activation functions associated with each node or layer of the neural network, and loss functions for learning the neural network.
- a data structure comprising a neural network may include any of the components disclosed above.
- the data structure including the neural network includes all or all of the data input to the neural network, the weights of the neural network, hyperparameters of the neural network, data obtained from the neural network, the activation function associated with each node or layer of the neural network, and the loss function for training the neural network may be configured including any combination of
- a data structure including a neural network may include any other information that determines a characteristic of a neural network.
- the data structure may include all types of data used or generated in the operation process of the neural network, and is not limited to the above.
- Computer-readable media may include computer-readable recording media and/or computer-readable transmission media.
- a neural network may be composed of a set of interconnected computational units, which may generally be referred to as nodes. These nodes may also be referred to as neurons.
- a neural network is configured by including at least one or more nodes.
- the data structure may include data input to the neural network.
- a data structure including data input to the neural network may be stored in a computer-readable medium.
- the data input to the neural network may include learning data input in a neural network learning process and/or input data input to the neural network in which learning is completed.
- Data input to the neural network may include pre-processing data and/or pre-processing target data.
- the preprocessing may include a data processing process for inputting data into the neural network.
- the data structure may include data to be pre-processed and data generated by pre-processing.
- the above-described data structure is merely an example, and the present disclosure is not limited thereto.
- the data structure may include the weights of the neural network.
- a weight and a parameter may be used interchangeably.
- a data structure including a weight of a neural network may be stored in a computer-readable medium.
- the neural network may include a plurality of weights.
- the weight may be variable, and may be changed by the user or algorithm in order for the neural network to perform a desired function. For example, when one or more input nodes are interconnected to one output node by respective links, the output node sets values input to input nodes connected to the output node and links corresponding to the respective input nodes. An output node value may be determined based on the parameter.
- the above-described data structure is merely an example, and the present disclosure is not limited thereto.
- the weight may include a weight variable in a neural network learning process and/or a weight in which neural network learning is completed.
- the variable weight in the neural network learning process may include a weight at the start of the learning cycle and/or a variable weight during the learning cycle.
- the weight for which neural network learning is completed may include a weight for which a learning cycle is completed.
- the data structure including the weight of the neural network may include a data structure including the weight variable in the neural network learning process and/or the weight in which the neural network learning is completed. Therefore, it is assumed that the above-described weights and/or combinations of weights are included in the data structure including the weights of the neural network.
- the above-described data structure is merely an example, and the present disclosure is not limited thereto.
- the data structure including the weights of the neural network may be stored in a computer-readable storage medium (eg, memory, hard disk) after being serialized.
- Serialization can be the process of converting a data structure into a form that can be reconstructed and used later by storing it on the same or a different computing device.
- the computing device may serialize the data structure to send and receive data over the network.
- a data structure including weights of the serialized neural network may be reconstructed in the same computing device or in another computing device through deserialization.
- the data structure including the weights of the neural network is not limited to serialization.
- the data structure including the weights of the neural network is a data structure to increase the efficiency of computation while using the resources of the computing device to a minimum (e.g., B-Tree, Trie, m-way search tree, AVL tree, Red-Black Tree).
- a minimum e.g., B-Tree, Trie, m-way search tree, AVL tree, Red-Black Tree.
- the data structure may include hyper-parameters of the neural network.
- the data structure including the hyperparameters of the neural network may be stored in a computer-readable medium.
- the hyperparameter may be a variable variable by a user. Hyperparameters are, for example, learning rate, cost function, number of iterations of the learning cycle, weight initialization (e.g., setting the range of weight values subject to weight initialization), Hidden Unit The number (eg, the number of hidden layers, the number of nodes of the hidden layer) may be included.
- the above-described data structure is merely an example, and the present disclosure is not limited thereto.
- FIG. 6 is a flowchart for generating highlight content related to an embodiment of the present disclosure.
- the computing device 100 may receive one or more user response data to the real-time content from one or more user terminals viewing the real-time content ( 510 ).
- the computing device 100 may identify one or more highlight sections in real-time content based on the analysis result of the user reaction data ( 520 ).
- the computing device 100 may generate highlight content based on one or more highlight sections ( 530 ).
- FIG. 7 is a diagram illustrating logic for implementing a method for generating highlight content related to an embodiment of the present disclosure.
- the computing device 100 may be implemented by the following logic.
- the computing device 100 includes a logic 610 for receiving one or more user reaction data for real-time content from one or more user terminals viewing real-time content, and an analysis result of the user reaction data and logic 620 for identifying one or more highlight sections in the real-time content based on , and logic 630 for generating highlight content based on one or more highlight sections.
- the user reaction data may be based on at least one of a natural language input or an action input of a user viewing the real-time content, and each user reaction data may be associated with at least a portion of the real-time content. .
- logic for receiving external reaction data for at least one of the real-time content or related party information of the real-time content, and user reaction data and the external reaction data for the real-time content are related to the real-time content person By storing in association with the information, it may further include logic for generating a knowledge base for the related parties of the real-time content.
- the external reaction data may be reaction data for at least one of the real-time content or a person concerned with the real-time content, and may be data received through a different channel from the user reaction data for the real-time content.
- the logic for generating the knowledge base for the relevant person of the real-time content may include: The knowledge base by extracting external narrative information related to the real-time content or related party information of the real-time content from reaction data, and extracting internal narrative information related to the real-time content or related party information of the real-time content from the user reaction data It may include logic for creating
- the knowledge base is generated for each of the parties of the real-time content, and is used to identify a user reaction unique to each of the parties, and a semantic analysis of the externally narrated information or the internally narrated information. or based on at least one of quantitative analysis.
- the logic for identifying one or more highlight sections in the real-time content based on an analysis result of the user reaction data may include: based on at least one of a semantic analysis or a quantitative analysis of the user reaction data. It may include logic for identifying one or more highlight sections in real-time content.
- the quantitative analysis may be performed based on a threshold determined based at least in part on the number of one or more user terminals viewing the real-time content.
- a logic for pre-processing the user reaction data based on a knowledge base of related information of the real-time content may be further included.
- the user reaction data further includes user reaction user identification information related to the user terminal that has transmitted the user reaction data, and based on a result of analysis of the user reaction data, one or more items in the real-time content
- the logic for identifying the highlight section includes logic for assigning a weight to each user reaction data based on the user reaction user identification information and an analysis of the user reaction data based on a weight assigned to each of the user reaction data and logic for generating a result and identifying one or more highlight sections in the real-time content based on the analysis result.
- the logic for assigning a weight to each user reaction data based on the user reaction user identification information may include a highlight section prediction contribution rate for real-time content of the user who generated the user reaction data, the user and logic for assigning the weight to each of the user reaction data based on at least one of a viewing contribution rate for content matched with the related information of the real-time content and a matching rate for the knowledge base of the user reaction data. can do.
- the highlight section prediction contribution rate for the real-time content is information on whether user reaction data is included in the pre-produced highlight content in relation to the related party of the real-time content, and is included in the pre-produced highlight content. It may be calculated based on at least one of quantitative analysis and semantic analysis of the included user response data.
- the viewing contribution rate for the content matched with the relevant information of the real-time content is information about the participation history of each of the one or more user terminals for the content matched with the relevant information of the real-time content, wherein the matching It may be calculated based on the number of contents and the participation history.
- the matching rate of the user reaction data to the knowledge base may be calculated based on determination of a similarity between the user reaction data and at least one of external and internal narration information included in the knowledge base. .
- logic for implementing the computing device may be implemented by a module, circuit, or means for implementing the computing device.
- FIG. 8 shows a simplified, general schematic diagram of an exemplary computing environment in which one embodiment of the present disclosure may be implemented, in conjunction with one embodiment of the present disclosure.
- program modules include routines, procedures, programs, components, data structures, etc. that perform particular tasks or implement particular abstract data types.
- methods of the present disclosure are suitable for single-processor or multiprocessor computer systems, minicomputers, mainframe computers as well as personal computers, handheld computing devices, microprocessor-based or programmable consumer electronics, etc. (each of which is It will be appreciated that other computer system configurations may be implemented, including those that may operate in connection with one or more associated devices.
- the described embodiments of the present disclosure may also be practiced in distributed computing environments where certain tasks are performed by remote processing devices that are linked through a communications network.
- program modules may be located in both local and remote memory storage devices.
- Computers typically include a variety of computer-readable media. Any medium accessible by a computer may be a computer-readable medium, and the computer-readable medium may include a computer-readable storage medium and a computer-readable transmission medium. Such computer-readable storage media includes volatile and nonvolatile media, removable and non-removable media. Computer readable storage media includes volatile and nonvolatile media, removable and non-removable media implemented in any method or technology for storage of information such as computer readable instructions, data structures, program modules or other data.
- a computer-readable storage medium may be RAM, ROM, EEPROM, flash memory or other memory technology, CD-ROM, digital video disk (DVD) or other optical disk storage device, magnetic cassette, magnetic tape, magnetic disk storage device, or other magnetic storage device. device, or any other medium that can be accessed by a computer and used to store the desired information.
- Computer-readable transmission media typically embody computer-readable instructions, data structures, program modules, or other data in a modulated data signal, such as a carrier wave or other transport mechanism. information delivery media.
- modulated data signal means a signal in which one or more of the characteristics of the signal is set or changed so as to encode information in the signal.
- computer-readable transmission media includes wired media such as a wired network or direct-wired connection, and wireless media such as acoustic, RF, infrared, and other wireless media. Combinations of any of the above are also intended to be included within the scope of computer-readable transmission media.
- An example environment 1100 implementing various aspects of the disclosure is shown including a computer 1102 , the computer 1102 including a processing unit 1104 , a system memory 1106 , and a system bus 1108 . do.
- the system bus 1108 couples system components, including but not limited to system memory 1106 , to the processing device 1104 .
- the processing device 1104 may be any of a variety of commercially available processors. Dual processor and other multiprocessor architectures may also be used as processing unit 1104 .
- the system bus 1108 may be any of several types of bus structures that may further interconnect a memory bus, a peripheral bus, and a local bus using any of a variety of commercial bus architectures.
- System memory 1106 includes read only memory (ROM) 1110 and random access memory (RAM) 1112 .
- ROM read only memory
- RAM random access memory
- a basic input/output system (BIOS) is stored in non-volatile memory 1110, such as ROM, EPROM, EEPROM, etc., which is the basic input/output system (BIOS) that helps transfer information between components within computer 1102, such as during startup. contains routines.
- BIOS basic input/output system
- RAM 1112 may also include high-speed RAM, such as static RAM, for caching data.
- the computer 1102 may also be configured with an internal hard disk drive (HDD) 1114 (eg, EIDE, SATA) - this internal hard disk drive 1114 may also be configured for external use within a suitable chassis (not shown).
- HDD hard disk drive
- FDD magnetic floppy disk drive
- optical disk drive 1120 eg, a CD-ROM
- the hard disk drive 1114 , the magnetic disk drive 1116 , and the optical disk drive 1120 are connected to the system bus 1108 by the hard disk drive interface 1124 , the magnetic disk drive interface 1126 , and the optical drive interface 1128 , respectively.
- the interface 1124 for implementing an external drive includes at least one or both of Universal Serial Bus (USB) and IEEE 1394 interface technologies.
- drives and their associated computer-readable media provide non-volatile storage of data, data structures, computer-executable instructions, and the like.
- drives and media correspond to storing any data in a suitable digital format.
- computer readable media refers to HDDs, removable magnetic disks, and removable optical media such as CDs or DVDs, those skilled in the art will use zip drives, magnetic cassettes, flash memory cards, cartridges, etc. It will be appreciated that other tangible computer-readable media such as etc. may also be used in the exemplary operating environment and any such media may include computer-executable instructions for performing the methods of the present disclosure.
- a number of program modules may be stored in the drive and RAM 1112 , including an operating system 1130 , one or more application programs 1132 , other program modules 1134 , and program data 1136 . All or portions of the operating system, applications, modules, and/or data may also be cached in RAM 1112 . It will be appreciated that the present disclosure may be implemented in various commercially available operating systems or combinations of operating systems.
- a user may enter commands and information into the computer 1102 via one or more wired/wireless input devices, for example, a pointing device such as a keyboard 1138 and a mouse 1140 .
- Other input devices may include a microphone, IR remote control, joystick, game pad, stylus pen, touch screen, and the like.
- these and other input devices are often connected to the processing unit 1104 through an input device interface 1142 that is connected to the system bus 1108, parallel ports, IEEE 1394 serial ports, game ports, USB ports, IR interfaces, It may be connected by other interfaces, etc.
- a monitor 1144 or other type of display device is also coupled to the system bus 1108 via an interface, such as a video adapter 1146 .
- the computer typically includes other peripheral output devices (not shown), such as speakers, printers, and the like.
- Computer 1102 may operate in a networked environment using logical connections to one or more remote computers, such as remote computer(s) 1148 via wired and/or wireless communications.
- Remote computer(s) 1148 may be workstations, server computers, routers, personal computers, portable computers, microprocessor-based entertainment devices, peer devices, or other common network nodes, and are generally Although including many or all of the components described, only memory storage device 1150 is shown for simplicity.
- the logical connections shown include wired/wireless connections to a local area network (LAN) 1152 and/or a larger network, eg, a wide area network (WAN) 1154 .
- LAN and WAN networking environments are common in offices and companies, and facilitate enterprise-wide computer networks, such as intranets, all of which can be connected to a worldwide computer network, for example, the Internet.
- the computer 1102 When used in a LAN networking environment, the computer 1102 is connected to the local network 1152 through a wired and/or wireless communication network interface or adapter 1156 .
- Adapter 1156 may facilitate wired or wireless communication to LAN 1152 , which also includes a wireless access point installed therein for communicating with wireless adapter 1156 .
- the computer 1102 may include a modem 1158 , connected to a communication server on the WAN 1154 , or otherwise establishing communications over the WAN 1154 , such as over the Internet. have the means
- a modem 1158 which may be internal or external and a wired or wireless device, is coupled to the system bus 1108 via a serial port interface 1142 .
- program modules described for computer 1102 may be stored in remote memory/storage device 1150 . It will be appreciated that the network connections shown are exemplary and other means of establishing a communication link between the computers may be used.
- Computer 1102 may be associated with any wireless device or object that is deployed and operates in wireless communication, for example, printers, scanners, desktop and/or portable computers, portable data assistants (PDAs), communications satellites, wireless detectable tags. It operates to communicate with any device or place, and phone. This includes at least Wi-Fi and Bluetooth wireless technologies. Accordingly, the communication may be a predefined structure as in a conventional network or may simply be an ad hoc communication between at least two devices.
- PDAs portable data assistants
- communications satellites for example, printers, scanners, desktop and/or portable computers, portable data assistants (PDAs), communications satellites, wireless detectable tags. It operates to communicate with any device or place, and phone. This includes at least Wi-Fi and Bluetooth wireless technologies. Accordingly, the communication may be a predefined structure as in a conventional network or may simply be an ad hoc communication between at least two devices.
- Wi-Fi Wireless Fidelity
- Wi-Fi is a wireless technology such as cell phones that allows these devices, eg, computers, to transmit and receive data indoors and outdoors, ie anywhere within range of a base station.
- Wi-Fi networks use a radio technology called IEEE 802.11 (a, b, g, etc.) to provide secure, reliable, and high-speed wireless connections.
- Wi-Fi can be used to connect computers to each other, to the Internet, and to wired networks (using IEEE 802.3 or Ethernet).
- Wi-Fi networks may operate in unlicensed 2.4 and 5 GHz radio bands, for example, at 11 Mbps (802.11a) or 54 Mbps (802.11b) data rates, or in products that include both bands (dual band). have.
- the various embodiments presented herein may be implemented as methods, apparatus, or articles of manufacture using standard programming and/or engineering techniques.
- article of manufacture includes a computer program, carrier, or media accessible from any computer-readable device.
- computer-readable media include magnetic storage devices (eg, hard disks, floppy disks, magnetic strips, etc.), optical disks (eg, CDs, DVDs, etc.), smart cards, and flash memory. devices (eg, EEPROMs, cards, sticks, key drives, etc.).
- various storage media presented herein include one or more devices and/or other machine-readable media for storing information.
- machine-readable medium includes, but is not limited to, wireless channels and various other media capable of storing, holding, and/or carrying instruction(s) and/or data.
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Databases & Information Systems (AREA)
- Multimedia (AREA)
- Signal Processing (AREA)
- Physics & Mathematics (AREA)
- Artificial Intelligence (AREA)
- Evolutionary Computation (AREA)
- Software Systems (AREA)
- General Physics & Mathematics (AREA)
- Mathematical Physics (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- Biophysics (AREA)
- Molecular Biology (AREA)
- General Health & Medical Sciences (AREA)
- Biomedical Technology (AREA)
- Life Sciences & Earth Sciences (AREA)
- Health & Medical Sciences (AREA)
- Computer Security & Cryptography (AREA)
- Information Transfer Between Computers (AREA)
Abstract
Description
Claims (16)
- 컴퓨터 판독가능 저장 매체에 저장된 컴퓨터 프로그램으로서, 상기 컴퓨터 프로그램은 하나 이상의 프로세서에서 실행되는 경우, 상기 하나 이상의 프로세서들로 하여금 하이라이트 영상 자동 편집을 위한 이하의 동작들을 수행하도록 하며, 상기 동작들은:실시간 컨텐츠를 시청하는 하나 이상의 사용자 단말로부터 상기 실시간 컨텐츠에 대한 하나 이상의 사용자 반응 데이터를 수신하는 동작;상기 사용자 반응 데이터에 대한 분석 결과에 기초하여 상기 실시간 컨텐츠에서 하나 이상의 하이라이트 구간을 식별하는 동작; 및상기 하나 이상의 하이라이트 구간에 기초하여 하이라이트 컨텐츠를 생성하는 동작;을 포함하는,컴퓨터 판독가능 저장 매체에 저장된 컴퓨터 프로그램.
- 제 1 항에 있어서,상기 사용자 반응 데이터는,상기 실시간 컨텐츠를 시청하는 사용자의 자연어 입력, 또는 액션 입력 중 적어도 하나에 기초하며, 각각의 사용자 반응 데이터는 상기 실시간 컨텐츠의 적어도 일부와 연관되는,컴퓨터 판독가능 저장 매체에 저장된 컴퓨터 프로그램.
- 제 1 항에 있어서,상기 실시간 컨텐츠, 또는 상기 실시간 컨텐츠의 관계자 정보 중 적어도 하나에 대한 외부 반응 데이터를 수신하는 동작; 및상기 실시간 컨텐츠에 대한 사용자 반응 데이터 및 상기 외부 반응 데이터를 상기 실시간 컨텐츠의 관계자 정보와 연관하여 저장함으로써, 상기 실시간 컨텐츠의 관계자에 대한 지식 베이스를 생성하는 동작;을 더 포함하는,컴퓨터 판독가능 저장 매체에 저장된 컴퓨터 프로그램.
- 제 3 항에 있어서,상기 외부 반응 데이터는,상기 실시간 컨텐츠, 또는 상기 실시간 컨텐츠의 관계자 중 적어도 하나에 대한 반응 데이터로, 상기 실시간 컨텐츠에 대한 사용자 반응 데이터와 상이한 채널을 통해 수신되는 데이터인,컴퓨터 판독가능 저장 매체에 저장된 컴퓨터 프로그램.
- 제 3 항에 있어서,상기 실시간 컨텐츠에 대한 사용자 반응 데이터 및 상기 외부 반응 데이터를 상기 실시간 컨텐츠의 관계자 정보와 연관하여 저장함으로써, 상기 실시간 컨텐츠의 관계자에 대한 지식 베이스를 생성하는 동작은,상기 외부 반응 데이터에서, 상기 실시간 컨텐츠, 또는 상기 실시간 컨텐츠의 관계자 정보와 관련한 외부 서술 정보를 추출하고, 상기 사용자 반응 데이터에서 상기 실시간 컨텐츠 또는 상기 실시간 컨텐츠의 관계자 정보와 관련한 내부 서술 정보를 추출함으로써 상기 지식 베이스를 생성하는 동작;을 포함하는,컴퓨터 판독가능 저장 매체에 저장된 컴퓨터 프로그램.
- 제 5 항에 있어서,상기 지식 베이스는,상기 실시간 컨텐츠의 관계자 각각에 대하여 생성되며, 상기 관계자 각각에 고유한 사용자 반응을 식별하기 위하여 사용되며, 상기 외부 서술 정보 또는 상기 내부 서술 정보에 대한 의미 분석 또는 정량 분석 중 적어도 하나에 기초하여 생성되는,컴퓨터 판독가능 저장 매체에 저장된 컴퓨터 프로그램.
- 제 1 항에 있어서,상기 사용자 반응 데이터에 대한 분석 결과에 기초하여 상기 실시간 컨텐츠에서 하나 이상의 하이라이트 구간을 식별하는 동작은,상기 사용자 반응 데이터에 대한 의미 분석 또는 정량 분석 중 적어도 하나에 기초하여 상기 실시간 컨텐츠에서 하나 이상의 하이라이트 구간을 식별하는 동작;을 포함하는,컴퓨터 판독가능 저장 매체에 저장된 컴퓨터 프로그램.
- 제 7 항에 있어서,상기 정량 분석은,상기 실시간 컨텐츠를 시청하는 하나 이상의 사용자 단말의 수에 적어도 부분적으로 기초하여 결정되는 임계값에 기초하여 수행되는,컴퓨터 판독가능 저장 매체에 저장된 컴퓨터 프로그램.
- 제 1 항에 있어서,상기 사용자 반응 데이터를 상기 실시간 컨텐츠의 관계자 정보에 대한 지식 베이스에 기초하여 전처리하는 동작;을 더 포함하는,컴퓨터 판독가능 저장 매체에 저장된 컴퓨터 프로그램.
- 제 1 항에 있어서,상기 사용자 반응 데이터는, 상기 사용자 반응 데이터를 전송한 사용자 단말에 관련한 사용자 반응 사용자 식별 정보를 더 포함하고,상기 사용자 반응 데이터에 대한 분석 결과에 기초하여 상기 실시간 컨텐츠에서 하나 이상의 하이라이트 구간을 식별하는 동작은,상기 사용자 반응 사용자 식별 정보에 기초하여 각각의 사용자 반응 데이터에 대한 가중치를 부여하는 동작; 및사용자 반응 데이터 각각에 부여된 가중치에 기초하여 상기 사용자 반응 데이터에 대한 분석 결과를 생성하고, 상기 분석 결과에 기초하여 상기 실시간 컨텐츠에서 하나 이상의 하이라이트 구간을 식별하는 동작;을 포함하는,컴퓨터 판독가능 저장 매체에 저장된 컴퓨터 프로그램.
- 제 10 항에 있어서,상기 사용자 반응 사용자 식별 정보에 기초하여 각각의 사용자 반응 데이터에 대한 가중치를 부여하는 동작은,상기 사용자 반응 데이터를 생성한 사용자의 실시간 컨텐츠에 대한 하이라이트 구간 예측 기여율, 상기 사용자의 상기 실시간 컨텐츠의 관계자 정보와 매칭된 컨텐츠에 대한 시청 기여율 및 상기 사용자 반응 데이터의 지식 베이스에 대한 매칭률 중 적어도 하나에 기초하여 상기 각각의 사용자 반응 데이터에 대한 상기 가중치가 부여되는 동작;을 포함하는,컴퓨터 판독가능 저장 매체에 저장된 컴퓨터 프로그램.
- 제 11 항에 있어서,상기 실시간 컨텐츠에 대한 하이라이트 구간 예측 기여율은,상기 실시간 컨텐츠의 관계자에 관련하여 사전 제작된 하이라이트 컨텐츠에서 사용자 반응 데이터가 포함되는지 여부에 관한 정보이며, 상기 사전 제작된 하이라이트 컨텐츠에 포함된 사용자 반응 데이터에 대한 정량 분석 및 의미 분석 중 적어도 하나에 기초하여 산출되는,컴퓨터 판독가능 저장 매체에 저장된 컴퓨터 프로그램.
- 제 11 항에 있어서,상기 실시간 컨텐츠의 관계자 정보와 매칭된 컨텐츠에 대한 시청 기여율은,상기 실시간 컨텐츠의 관계자 정보와 매칭된 컨텐츠에 대한 상기 하나 이상의 사용자 단말 각각의 참여 이력에 관한 정보로, 상기 매칭된 컨텐츠의 개수 및 상기 참여 이력에 기초하여 산출되는,컴퓨터 판독가능 저장 매체에 저장된 컴퓨터 프로그램.
- 제 11 항에 있어서,상기 사용자 반응 데이터의 지식 베이스에 대한 매칭률은,상기 지식 베이스에 포함된 외부 서술 정보 및 내부 서술 정보 중 적어도 하나와 상기 사용자 반응 데이터의 유사도 판별에 기초하여 산출되는,컴퓨터 판독가능 저장 매체에 저장된 컴퓨터 프로그램.
- 컴퓨팅 장치의 하나 이상의 프로세서에서 수행되는 하이라이트 영상을 자동 편집하는 방법으로,실시간 컨텐츠를 시청하는 하나 이상의 사용자 단말로부터 상기 실시간 컨텐츠에 대한 하나 이상의 사용자 반응 데이터를 수신하는 단계;상기 사용자 반응 데이터에 대한 분석 결과에 기초하여 상기 실시간 컨텐츠에서 하나 이상의 하이라이트 구간을 식별하는 단계; 및상기 하나 이상의 하이라이트 구간에 기초하여 하이라이트 컨텐츠를 생성하는 단계;를 포함하는,컴퓨팅 장치의 하나 이상의 프로세서에서 수행되는 하이라이트 영상을 자동 편집하는 방법.
- 하이라이트 영상을 자동 편집하는 서버로서,하나 이상의 코어를 포함하는 프로세서;상기 프로세서에서 실행가능한 프로그램 코드들을 포함하는 메모리; 및사용자 단말 및 외부 서버와 데이터를 송수신하는 네트워크부;를 포함하고, 그리고상기 프로세서는,실시간 컨텐츠를 시청하는 하나 이상의 사용자 단말로부터 상기 실시간 컨텐츠에 대한 하나 이상의 사용자 반응 데이터를 수신하고,상기 사용자 반응 데이터에 대한 분석 결과에 기초하여 상기 실시간 컨텐츠에서 하나 이상의 하이라이트 구간을 식별하고, 그리고상기 하나 이상의 하이라이트 구간에 기초하여 하이라이트 컨텐츠를 생성하는,하이라이트 영상을 자동 편집하는 서버.
Applications Claiming Priority (2)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
KR10-2019-0163539 | 2019-12-10 | ||
KR1020190163539A KR102286638B1 (ko) | 2019-12-10 | 2019-12-10 | 하이라이트 영상 자동 편집을 위한 컴퓨터 프로그램 |
Publications (1)
Publication Number | Publication Date |
---|---|
WO2021118050A1 true WO2021118050A1 (ko) | 2021-06-17 |
Family
ID=76330510
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
PCT/KR2020/014636 WO2021118050A1 (ko) | 2019-12-10 | 2020-10-26 | 하이라이트 영상 자동 편집을 위한 컴퓨터 프로그램 |
Country Status (2)
Country | Link |
---|---|
KR (1) | KR102286638B1 (ko) |
WO (1) | WO2021118050A1 (ko) |
Families Citing this family (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
KR20230126907A (ko) | 2022-02-24 | 2023-08-31 | 명지대학교 산학협력단 | 연합학습을 이용한 비디오 영상의 특정 구간 추출 방법 및 시스템 |
KR102593762B1 (ko) * | 2023-05-23 | 2023-10-26 | 주식회사 스푼라디오 | 보이스 인식 기반의 라이브 인터넷 방송에 대한 하이라이트 콘텐츠를 생성하는 방법 및 이를 위한 서버 |
KR102678303B1 (ko) | 2023-11-27 | 2024-06-25 | 주식회사엠제이에스파트너스 | 시청자 반응 데이터 분석을 기반으로 한 영상 콘텐츠 기획 및 자동 편집 솔루션 제공 방법, 장치 및 시스템 |
Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
KR20150030387A (ko) * | 2013-09-12 | 2015-03-20 | 엠랩 주식회사 | 스트리밍 동영상에 대한 코멘트 태깅 시스템 및 그 제공방법 |
KR20170029933A (ko) * | 2015-09-08 | 2017-03-16 | 네이버 주식회사 | 동영상 컨텐츠의 하이라이트 영상을 추출하여 제공하는 방법과 시스템 및 기록 매체 |
KR20180017821A (ko) * | 2016-08-11 | 2018-02-21 | 한화테크윈 주식회사 | 실시간 시청자 반응을 전달하는 방송 서비스 장치 |
KR101867082B1 (ko) * | 2017-08-23 | 2018-06-12 | 서울과학기술대학교 산학협력단 | 하이라이트 영상 생성 장치 및 방법 |
KR101957807B1 (ko) * | 2018-05-30 | 2019-03-14 | (주) 비투아이 | 다인칭 시점의 sns 라이브 방송을 위한 오디오 중계 방법 및 시스템 |
Family Cites Families (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
KR101628237B1 (ko) | 2009-01-21 | 2016-06-22 | 삼성전자주식회사 | 하이라이트 영상 구성 방법 및 장치 |
-
2019
- 2019-12-10 KR KR1020190163539A patent/KR102286638B1/ko active IP Right Grant
-
2020
- 2020-10-26 WO PCT/KR2020/014636 patent/WO2021118050A1/ko active Application Filing
Patent Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
KR20150030387A (ko) * | 2013-09-12 | 2015-03-20 | 엠랩 주식회사 | 스트리밍 동영상에 대한 코멘트 태깅 시스템 및 그 제공방법 |
KR20170029933A (ko) * | 2015-09-08 | 2017-03-16 | 네이버 주식회사 | 동영상 컨텐츠의 하이라이트 영상을 추출하여 제공하는 방법과 시스템 및 기록 매체 |
KR20180017821A (ko) * | 2016-08-11 | 2018-02-21 | 한화테크윈 주식회사 | 실시간 시청자 반응을 전달하는 방송 서비스 장치 |
KR101867082B1 (ko) * | 2017-08-23 | 2018-06-12 | 서울과학기술대학교 산학협력단 | 하이라이트 영상 생성 장치 및 방법 |
KR101957807B1 (ko) * | 2018-05-30 | 2019-03-14 | (주) 비투아이 | 다인칭 시점의 sns 라이브 방송을 위한 오디오 중계 방법 및 시스템 |
Also Published As
Publication number | Publication date |
---|---|
KR102286638B1 (ko) | 2021-08-05 |
KR20210073077A (ko) | 2021-06-18 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
WO2021118050A1 (ko) | 하이라이트 영상 자동 편집을 위한 컴퓨터 프로그램 | |
US20230015606A1 (en) | Named entity recognition method and apparatus, device, and storage medium | |
US20190124020A1 (en) | Chatbot Skills Systems And Methods | |
US20190103111A1 (en) | Natural Language Processing Systems and Methods | |
WO2019172656A1 (en) | System and method for language model personalization | |
WO2022250439A1 (ko) | 스트리밍 데이터 기반 영상 편집점 추천 방법, 서버 및 컴퓨터프로그램 | |
WO2019124647A1 (ko) | 대화형 ai 에이전트 시스템을 위한 계층적 대화 흐름 관리 모델을 자동으로 구축 또는 갱신하는 방법, 컴퓨터 장치 및 컴퓨터 판독가능 기록 매체 | |
EP3545487A1 (en) | Electronic apparatus, controlling method of thereof and non-transitory computer readable recording medium | |
WO2020182123A1 (zh) | 用于推送语句的方法和装置 | |
WO2018174443A1 (en) | Electronic apparatus, controlling method of thereof and non-transitory computer readable recording medium | |
WO2018174397A1 (ko) | 전자 장치 및 제어 방법 | |
CN112489676B (zh) | 模型训练方法、装置、设备及存储介质 | |
WO2021075658A1 (en) | System and method for prediction and recommendation using collaborative filtering | |
US20240064348A1 (en) | Image analysis system | |
JP7337172B2 (ja) | 音声パケット推薦方法、装置、電子機器およびプログラム | |
CN111753126B (zh) | 用于视频配乐的方法和装置 | |
WO2019156536A1 (ko) | 학습 데이터 중 식별 가능하지만 학습 가능성이 없는 데이터의 레이블화를 통한, 대화형 ai 에이전트 시스템을 위한 지식베이스 모델의 구축 또는 갱신 방법, 컴퓨터 장치, 및 컴퓨터 판독 가능 기록 매체 | |
JP2024012152A (ja) | テキスト情報の中からターゲット単語と対応する単語を識別する方法(method for identify a word corresponding to a target word in text information) | |
WO2023000782A1 (zh) | 获取视频热点的方法、装置、可读介质和电子设备 | |
CN111339770B (zh) | 用于输出信息的方法和装置 | |
CA3217258A1 (en) | Systems and methods to increase viewership of online content | |
WO2020055173A1 (en) | Method and system for audio content-based recommendations | |
Bhatti et al. | LSTM-based Siamese neural network for Urdu news story segmentation | |
WO2024111803A1 (ko) | 실시간 영상 콘텐츠 생성 방법 및 시스템 | |
WO2022086063A1 (ko) | 소셜 네트워크 서비스 계정의 스타일 정보를 산출하는 방법 및 장치 |
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: 20897767 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: 20897767 Country of ref document: EP Kind code of ref document: A1 |
|
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 14/12/2022) |
|
122 | Ep: pct application non-entry in european phase |
Ref document number: 20897767 Country of ref document: EP Kind code of ref document: A1 |