KR101916874B1 - Apparatus, method for auto generating a title of video contents, and computer readable recording medium - Google Patents

Abstract

A server for communicating with a user terminal through a communication network is provided. The server according to the present invention is a server for automatically generating a title of a video content, which communicates with the user terminal through the communication network. The server includes: a communication unit configured to communicate the user terminal with the communication network; a storage unit configured to receive text information inputted from the user terminal and store the text information through the communication unit; and a control unit configured to generate the title of a video by analyzing the stored text information. The control unit is configured to reconfigure the stored text information as one of four templates, wherein the four templates are a bag of two or three words, a phrase, or a complete sentence including a subject, a verb, and an object. The control unit tags each token with a word class after analyzing a morpheme of the text information and selects the most frequently appearing word to generate the title of the video. Accordingly, the present invention can automatically generate a video clip without the involvement of a user.

Description

TECHNICAL FIELD [0001] The present invention relates to a method, apparatus, and computer readable recording medium for automatically generating a title of a video highlight image,

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a method for generating moving picture video clips, and more particularly, to a method and apparatus for automatically generating a video clip including a title without user's involvement.

As the amount of video content explosively increases and video content is provided in various media, consumers are increasingly inclined to select and view their favorite content. Also, in order to select and view a desired image in a short time, a tendency of preferring a short video clip that highlights a desired scene among all the video contents is also increasing.

Generally, in order to create a short video clip, it is common for an editor to directly select a highlight portion of an entire movie to create. On the other hand, Korean Patent Application No. 10-2017-0029933 discloses a method of automatically editing a highlight image through a comment or cumulative information (the number of times of repeated reproduction, the number of times of capture, the number of times of sharing, etc.) However, such a system relates to a method of editing a video image itself, and does not refer to a method of giving an appropriate title to a video image. In addition, such a system has a problem that a user has to input search information by itself as a search word for moving image search, and if not, it is difficult to search so that it is difficult to be identified and found among many videos.

Korean Patent Application No. 10-2017-0029933

Thus, the present disclosure can provide a method and apparatus for automatically generating video clips without user involvement.

The present disclosure can provide a method and apparatus for assigning titles to automatically generated video clips.

According to an aspect of the present invention, there is provided a server for automatically generating a title of moving picture content, which communicates with a user terminal through a communication network. The server comprising: a communication unit configured to communicate with a user terminal through the communication network; A storage unit configured to receive text information input from the user terminal and store the received text information through the communication unit; And a control unit configured to analyze the stored text information to generate a title of a moving image, wherein the control unit is configured to reconstruct the stored text information into one of four templates, the four templates including two or three words A complete sentence containing a bag, phrase, or subject, verb, (object); After the morpheme analysis of the text information, the control unit is configured to tag each token with a word class, and to select the most frequently appearing word to generate the title of the moving picture.

According to another aspect of the present invention, there is provided a method of automatically generating a title of moving picture content to be communicated to a user terminal through a communication network and performed by a server. The method comprising: receiving a text associated with the moving picture content, the text being transmitted from the user terminal; Detecting a highlight event of the moving picture content; Correcting the detected highlight event; And synthesizing a title for the highlight event of the moving picture content, the method including automatically generating a title of the moving picture content.

In one embodiment, the step of detecting a highlight event of the moving picture content may include: dividing the moving picture content into a plurality of moving picture segments having a predetermined time window; Calculating a number of times a chat message corresponding to each of the plurality of moving image segments or a number of times each of the plurality of moving image segments is posted to a web page; And sorting the plurality of video segments in descending order of the number of video segments and selecting a predetermined number of video segments.

In one embodiment, the step of correcting the detected highlight event comprises marking only the first highlight event to remove a case where a subsequent highlight event belongs to an adjacent time window and which belongs to the same scene, And preventing a highlight event.

In one embodiment, the step of correcting the detected highlight event may include correcting a time difference between the playing time of the moving picture content and the text data associated with the moving picture content.

In one embodiment, compositing a title for the highlight event of the moving picture comprises defining four sentence templates, wherein the four sentence templates are two or three word bags, phrases, , Verb, and (object); Subjecting the text to morphological analysis and tagging each token with a word class; Generating a bundle of words by applying a Word2Vec model to the tagged token; And generating a title by combining two or three words of the calculated words having short distances.

According to another aspect of the present invention there is provided a computer readable medium having stored thereon one or more instructions which when executed by a computer causes the computer to perform any one of the methods described above, A recording medium is provided.

According to the present disclosure, a method and apparatus and the like capable of automatically generating a movie highlight video clip can be provided without depending on a user's input.

According to the present disclosure, the generated video clip can be automatically given an appropriate title without user intervention.

1 is a diagram illustrating an example of a network environment according to an embodiment of the present invention.
Fig. 2 is a view schematically showing an exemplary configuration of the server of Fig. 1. Fig.
3 is a schematic operational flow diagram illustrating operations performed on a server for uploading a video highlight along with a title to a bulletin board on a server, in accordance with an embodiment of the present invention.
FIG. 4 is a diagram illustrating an exemplary method of generating a movie highlight along with a title in a mass interaction website (YouTube interaction, Twitter, Naver sports, or a Dice Insight) according to the operation described in FIG.
5 is a diagram illustrating an embodiment of providing a movie highlight event with a title in each genre.
Figure 6 is a graph showing the average score of four title types for each video category.

Hereinafter, embodiments of the present invention will be described in detail with reference to the accompanying drawings. Hereinafter, when it is determined that there is a possibility that the gist of the present invention may be unnecessarily blurred, a detailed description of known functions and configurations will be omitted. In addition, it should be understood that the following description is only an embodiment of the present invention, and the present invention is not limited thereto.

The terminology used herein is for the purpose of describing particular embodiments only and is not intended to be limiting of the present invention. For example, an element expressed in singular < Desc / Clms Page number 5 > terms should be understood as including a plurality of elements unless the context clearly dictates a singular value. In addition, in the specification of the present invention, it is to be understood that terms such as "include" or "have" are intended to specify the presence of stated features, integers, steps, operations, components, It is not intended that the use of the term exclude the presence or addition of one or more other features, numbers, steps, operations, components, parts, or combinations thereof.

As used herein, the term " block " or " part " means a functional part that performs at least one function or operation, and may be implemented in hardware or software or a combination of hardware and software. Also, a plurality of 'blocks' or 'sub-units' may be embodied in at least one processor integrated with at least one software module, except for 'blocks' or 'sub-units' that need to be implemented in specific hardware .

In addition, all terms used herein, including technical or scientific terms, unless otherwise defined, have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs. Commonly used predefined terms are to be interpreted as having a meaning consistent with the contextual meaning of the related art and are not to be construed as being excessively limited or extended unless explicitly defined otherwise in the specification of the present invention You should know.

Hereinafter, embodiments of the present invention will be described in detail with reference to the accompanying drawings.

1 is a diagram schematically illustrating a configuration of a network environment 100 according to an embodiment of the present invention. Network environment 100 includes one or more user terminals 102a-102n, a communication network 104, and a server 106. As shown in FIG. FIG. 1 is a diagram for explaining the invention. The number of user terminals and the number of servers are not limited to those shown in FIG.

According to one embodiment of the present invention, the plurality of user terminals 102a-102n may be various fixed or portable communication terminals used by the user. Each of the plurality of user terminals 102a-102n may be, for example, a desktop, laptop, smart phone, tablet PC, or various other communication terminals. According to one embodiment of the present invention, the user terminals 102a-102n may communicate with the server 106 via the communication network 104 and send and receive necessary information. According to an embodiment of the present invention, each of the plurality of user terminals 102a-102n accesses the server 106 via the communication network 104 and searches each video on the bulletin board operated by the server 106, You can upload new videos to the bulletin board.

According to an embodiment of the present invention, the communication network 104 may be any wired or wireless communication network and may include, for example, a WiFi network, a LAN network, a WAN network, an Internet network, etc., It is not.

In accordance with one embodiment of the present invention, the server 106 communicates with a plurality of user terminals 102a-102n via a communication network 104, in accordance with a given purpose, and provides commands, codes, files, Or a plurality of computer devices.

In one example, the server may provide a file for the installation of the members application to the user terminal 102a-102n connected via the network 140. [ In this case, the user terminal 102a-102n can install the application using the file provided from the server 106. [ The server 106 may be connected to the server 106 under the control of an operating system (OS) and at least one program included in the user terminals 102a-102n to receive services or contents provided by the server 106 . For example, the user terminal 102a-102n transmits a service request message to the server 106 via the network 140 under the control of the application, and the server 106 transmits a code corresponding to the service request message, The user terminals 102a to 102n can transmit the content to the user terminals 102a to 102n, and the user terminals 102a to 102n can provide the content to the user by constituting and displaying a screen according to the code according to the control of the application.

Figure 2 is a schematic diagram illustrating an exemplary configuration of the server of Figure 1 in accordance with one embodiment of the present invention. The server 102 may include a control unit 204, a storage unit 206, and a communication unit 208. As shown in FIG.

According to an embodiment of the present invention, the control unit 204 may be configured to execute various computer programs and control the operation thereof in order to be connected to and control the storage unit 206 and the communication unit 208. [ The control unit 204 may be configured to receive data from one component and to perform data formatting in a format suitable for use by the other component. The control unit 204 may be connected to the storage unit 206 in a communicable manner. In various embodiments, the control unit 204 may be integrally integrated with other components, such as, for example,

The control unit 204 receives the text information through the communication unit 208 and analyzes the received text information to generate a title of the moving image. According to an embodiment of the present invention, the control unit 204 may detect the highlight event of the moving image content using the moving image content stored in the storage unit 206 and the associated text information, and may generate an appropriate title in the moving image video clip have.

According to an embodiment of the present invention, the control unit 204 can classify the text into one of four sentence templates. In one embodiment, the title of a moving picture may be a complete sentence including a bag, a phrase, or a subject, verb, (object) of two or three words. After morpheme analysis of the chat message or comment, the control unit 204 may tag each token with a word class.

In one embodiment, the control unit 204 may select the most frequently occurring word as the title of the scene. Because the most common words can effectively explain the situation of the scene. In one embodiment, the control unit 204 may select the name of the main character as the title of the corresponding scene. In one embodiment, the control unit 204 may remove an inappropriate word such as a profanity word in advance.

In one embodiment, the control unit 204 may generate a bundle of two or three words by generating a compound noun by combining short words between the words. In one embodiment, the control unit 204 may generate a bundle of words using a Word2Vec model.

In one embodiment, when two or three words are given, the control unit 204 may combine words to synthesize a narrative style sentence describing the scene. To synthesize a scene title sentence, the control unit 204 may select an appropriate noun and consider the object and the verb that can enter the sentence. It is possible to extract nouns and other parts that are mainly referred to as subject of a sentence, and to calculate the similarity between the noun and other parts of the sentence. The control unit 104 can complete a sentence by selecting the shortest word vector and a pair of tokens.

The storage unit 206 can store and update moving picture information and the like transmitted from the terminal 102 via the communication network 104 and the communication unit 208 in Fig. According to an embodiment of the present invention, the storage unit 206 also stores a video related chat message transmitted in real time from each of the terminals 102a to 102c via the communication network 102 and the communication unit 208 in Fig. 1, Comments, video clips, and the like.

The storage unit 206 may also store and update the moving image content transmitted from the external content providing server (not shown) through the communication network 102 and the communication unit 208 of FIG.

The storage unit 206 may be a flash memory type, a hard disk type, a MultiMedia Card (MMC), a card type memory (for example, SD (Secure Digital) card or XD (Random Access Memory), SRAM (Static Random Access Memory), ROM (Read Only Memory), EEPROM (Electrically Erasable Programmable Read-Only Memory), PROM (Programmable Read-Only Memory) A magnetic disk, a magnetic disk, and an optical disk, but those skilled in the art will recognize that the embodiment of the storage unit 206 is not limited thereto.

According to one embodiment of the present invention, the communication unit 208 enables the server 106 to communicate with the user terminal 102 via the communication network 104. [ According to an embodiment of the present invention, the communication unit 208 can transmit information processed by the control unit 104 to the user terminal 102 via the communication network 104 according to a predetermined protocol. According to one embodiment of the present invention, the communication unit 208 can appropriately encode or decode information to be transmitted and received according to the protocol of the communication network 104 in Fig. According to the embodiment of the present invention, the communication unit 208 can receive various information from the user terminal 102 and perform appropriate processing.

In the hardware embodiment, the embodiments described above are applicable to application specific integrated circuits (ASICs), digital signal processors (DSPs), digital signal processing devices (DSPDs) At least one of Programmable Logic Devices (PLDs), Field-Programmable Gate Arrays (FPGAs), processors, controllers, micro-controllers and microprocessors Can be implemented using one.

Embodiments involving procedures, steps, or functions may be implemented with firmware / software modules that are executable on a hardware platform that allows performing at least one function or operation. In this case, the firmware / software module may be implemented by a software application written in a suitable program language.

3 is a schematic operational flow diagram illustrating operations performed on a server for uploading a video highlight along with a title to a bulletin board on a server, in accordance with an embodiment of the present invention.

A step of detecting a highlight event of the moving picture is started (S310). In one embodiment, chat messages and comments can be analyzed to find highlight scenes. For example, the moving picture is divided into a plurality of moving picture segments having a predetermined time window (S313), and the number of times the chat message corresponding to each of the plurality of moving picture segments or the number of times each of the plurality of moving picture segments is posted to the web page (S315). Thereafter, each of the plurality of moving image segments is sorted in descending order of the number of times, and a predetermined number of moving image segments are selected (S317).

Thereafter, the highlight event can be corrected (S320). In one embodiment, correcting a highlight event may prevent duplicate highlight events by marking only the first highlight event, since in reality a successive highlight event belongs to a contiguous time window and this often belongs to the same scene S321).

In one embodiment, the time difference between the moving picture content and the text data associated with the moving picture content can be corrected (S323). For example, a user can generate related text data while watching video content on a television, a PC, or a smart phone. Thus, the time at which the video content starts and the time at which the associated text data is generated may not coincide. Generally, the time for the user to watch the video contents may vary depending on the type of media and communication channel. For example, a terrestrial digital television is transmitted through a radio wave by encoding a moving picture into MPEG2-TS, and the moving picture content on the smart phone is transmitted through wireless LAN, CDMA, 3G or LTE. In the case of a smartphone using a wireless channel , The end-to-end delay is larger than in the case of a radio wave or a wired environment because the video content has to be additionally encoded in MPEG4 or H.264 and there is latency in the radio link. There is also a delay due to a time stamp of a text message on a web site, which increases the time difference between the video content and related text data. This is because the user creates a chat or comment message and then generates a timestamp when sending to the Internet. Therefore, a more accurate highlight event can be selected by considering the time difference between the moving picture scene and the corresponding text message. In this regard, the present invention refers to the time table of the TV program or video streaming site and corrects the start time of the video content based on keywords such as 'start' or 'start' of the chat message of the website.

Thereafter, a title is synthesized for the highlight event of the moving picture (S330).

In one embodiment, four sentence templates may be defined to create a movie highlight event (S331). In one embodiment, the title of the video may be a complete or two-or three-word bag, a phrase, or a complete sentence that includes subject, verb, (object). In one embodiment, after stemming a chat message or comment, each token can be tagged with a word class.

In one embodiment, it is possible to select a noun most frequently appearing as the subject of the scene by paying attention to the noun, since the nouns most frequently can describe the situation of the scene effectively (S333). In one embodiment, the name of the protagonist can be selected as the title of the scene. In one embodiment, inappropriate words such as profanity can be removed in advance.

Next, a bundle of two or three words can be generated by generating a compound noun by combining words having a short distance between words (S335). In one embodiment, a Word2Vec model may be used to generate a bundle of words.

Next, when two or three words are given, it is possible to combine words to synthesize a narrative style sentence describing the scene (S337). To synthesize scene title sentences, you can choose appropriate verbs, objects, and nouns that can be included in the sentence. We can consider the nouns that are mainly mentioned as the subject of the sentence, and after extracting nouns and other parts, we can calculate the similarity between nouns and other parts. The sentence can be completed by selecting the shortest word vector and a pair of tokens.

FIG. 4 is a diagram illustrating an exemplary method of generating a movie highlight along with a title in a mass interaction website (YouTube interaction, Twitter, Naver sports, or a Dice Insight) according to the operation described in FIG.

As shown in FIG. 4, it is possible to collect data from a public mutual website, synthesize highlight titles after detecting highlight events, and provide highlight images along with titles.

5 is a diagram illustrating an embodiment of providing a movie highlight event with a title in each genre.

According to FIG. 5, it can be seen that the generated title generally includes the names of the characters of each scene, and the words are generally associated with the main actors of each scene. Also, it can be seen that the synthesized noun describes the behavior of the actors of the scene with high similarity metric between the words associated with the characters' names. The completed sentence also expresses the name and behavior of the character.

In the case of drama, we were able to observe words related to the actor's appearance and words related to the performance of the scene. In the case of entertainment, words such as 'funny', 'interesting', 'amazing' It became popular.

In the case of sports, I could find many terms related to score or player's foul play, or I could find many words to evaluate play.

In the case of online games, there were many different vocabulary and game-specific terms for characters, abbreviations and vias.

To measure the accuracy, 80 titles were selected for 20 video contents, and the accuracy of title was evaluated by dividing into three grades by hand. We surveyed 80 titles of 20 video contents through questionnaires from 12 applicants. As a result of the survey, the accuracy was 67.5% and 54 out of 80 titles received the same grade.

A title that received a high rating was found to contain interesting or coined words or titles with appropriate expressions. On the other hand, a title that received a low rating was a combination of awkward words or inconsistencies between video content and title.

Figure 6 is a graph showing the average score of four title types for each video category.

Overall, the unionized title received a high score across the genre, with shorter words tending to form more meaningful titles in each scene. Bundled titles of two or three words generally received high marks for dramas and entertainments. On the other hand, the sentence title received a lower score. Users have a tendency to give a lower score if there is a combination of awkward words in the sentence or if there is a discrepancy between the title and the video content.

On the other hand, a three-word title is scored lower than a two-word title, which is analyzed because three words at the top of the frequency often contain a single word that is not directly related to the video footage. On average, people seem to prefer titles with a combination of words.

In the embodiments disclosed herein, the arrangement of the components shown may vary depending on the environment or requirements in which the invention is implemented. For example, some components may be omitted or some components may be integrated into one. In addition, the arrangement order and connection of some components may be changed.

Although the present invention and its various functional elements have been described in terms of specific embodiments, it is to be understood that the invention may be implemented in hardware, software, firmware, middleware, or a combination thereof and may be implemented as a system, subsystem, It should be understood that the invention may be utilized in various other embodiments. When implemented in software, the elements of the present invention may be instructions / code segments for performing necessary tasks. The program or code segments may be stored in a machine-readable medium, such as a processor-readable medium, or a computer program product. A machine-readable medium or a processor-readable medium may include any medium that can be read by a machine (e.g., processor, computer, etc.) and capable of storing or transmitting information in an executable form.

While the present invention has been particularly shown and described with reference to exemplary embodiments thereof, it is to be understood that the invention is not limited to the disclosed exemplary embodiments, but, on the contrary, is intended to cover various modifications and equivalent arrangements included within the spirit and scope of the appended claims. It will be understood by those skilled in the art that various changes in form and details may be made therein without departing from the spirit and scope of the invention. Accordingly, the technical scope of the present invention should be determined only by the appended claims.

Claims (7)

A server for automatically generating a title of moving picture content, communicating with a user terminal through a communication network,
A communication unit configured to receive at least one chat message associated with the moving picture content from the user terminal through the communication network;
A storage unit configured to store a chat message received from the user terminal; And
And a controller configured to analyze the text information of the stored chat message to generate a title of the moving picture,
The control unit
Dividing the moving picture content into a plurality of moving picture segments having a predetermined time window,
Analyzing the stored chat message to determine the number of chat messages corresponding to each of the plurality of video segments,
Selecting a predetermined number of video segments in the order of the number of the chat messages among the plurality of video segments, detecting the highlight events,
Wherein the first highlight event is a first highlight event among consecutive highlight events among the highlight events and corrects the consecutive highlight events as one highlight event,
Wherein the header of the highlight event is configured to stem the chat message associated with the highlight event of the stored chat message and to generate a title of the highlight event based on the most frequently occurring noun in the chat message associated with the highlight event.
A server that automatically generates titles for video content. A method of automatically generating a title of a moving picture content to be communicated to a user terminal through a communication network and performed by a server,
Receiving, from the user terminal, at least one chat message associated with the moving picture content;
Dividing the moving picture content into a plurality of moving picture segments having a predetermined time window;
Analyzing at least one chat message received from the user terminal and determining a number of chat messages corresponding to each of the plurality of video segments;
Selecting a predetermined number of video segments in descending order of the number of chat messages among the plurality of video segments, and detecting highlight events;
Marking only the first highlight event among consecutive highlight events among the highlight events and correcting the consecutive highlight events as one highlight event; And
Generating a title of the highlight event based on the most frequently occurring noun in the chat message associated with the highlight event, by stemming a chat message associated with the highlight event among at least one chat message transmitted from the user terminal
And automatically generating a title of the moving picture content. delete delete 3. The method of claim 2,
And correcting a time difference between the playback time of the moving image content and a chat message associated with the moving image content. 3. The method of claim 2,
The step of generating a title for the highlight event
Defining four sentence templates, wherein said four sentence templates are complete sentences comprising two or three word bags, phrases, or subject, verb, (object);
Morphing a chat message associated with the highlight event, tagging each token with a word class, and applying a Word2Vec model to the tagged token to generate a bundle of words; And
Generating a title by combining words of the generated bundle;
The title of the moving picture content including the title of the moving picture is automatically generated. A computer-readable medium having stored thereon one or more instructions for causing the computer to perform the method according to any of claims 2 and 5 to 6 when executed by the computer. Readable recording medium.

Images