KR100721409B1 - Method for searching situation of moving picture and system using the same - Google Patents

Abstract

The present invention relates to a video search method and a video search system, and more particularly, to a video scene search method and a video scene search system for searching a partial scene in a video. In accordance with an aspect of the present invention, a video scene search method comprises: receiving a search word including a scene search keyword from a user, recognizing the search term as a video scene search by identifying the scene search keyword, and the search word Searching for a scene search database in which a script representing a content characteristic of a scene video is stored using the query word generated using the scene video (the scene video corresponds to a part of the entire video divided into content-based units), and the query word And providing a search result for a scene video corresponding to a script matched with. Therefore, a desired scene can be effectively searched for in the entire video by using keywords related to the video content.

Video Search, Scene Search, Content Based Search

Description

Video scene search method and scene search system using the same {METHOD FOR SEARCHING SITUATION OF MOVING PICTURE AND SYSTEM USING THE SAME}

1 is a flowchart illustrating a video scene search method according to an exemplary embodiment of the present invention.

2 and 3 are conceptual diagrams showing a relationship between an entire video and a scene video.

4 is a diagram illustrating an example of data stored in a scene search database of the present invention.

5 is an operation flowchart illustrating an example of a search result providing step illustrated in FIG. 1.

6 is a diagram illustrating an example of data stored in a scene search database when two or more scripts are allocated to one scene video.

7 is a diagram illustrating an example of a video scene search result page.

8 is a diagram illustrating a screen for confirming whether to play back the entire video including the played scene video after the playback of the scene video ends.

9 is a diagram illustrating a guide screen regarding a video scene search restriction.

10 is a block diagram illustrating a video scene search system according to an embodiment of the present invention.

<Explanation of symbols for the main parts of the drawings>

1010: search term input unit

1020: query generator

1030: scene search database

1040: scene video search

1050: search result provider

The present invention relates to a video search method and a video search system, and more particularly, to a video scene search method and a video scene search system for searching a partial scene in a video.

As the download speed through the Internet has become faster, users can easily transmit or receive large amounts of data through the Internet. Therefore, it is becoming common to download high-quality video data, which has not been achieved before, through the Internet and to play in real time.

In fact, most TV programs can be viewed again through the Internet, and various videos such as over-the-air / satellite TV, movies, music videos, and sports broadcasts are provided through the Internet.

Unlike a TV broadcasted according to a unilaterally arranged program, a video provided through the Internet has an advantage of playing a video of a desired content at a desired time. For example, a user who wants to watch the 9 o'clock news may search for and watch the 9 o'clock news through the Internet even the next morning.

However, if a user wants to watch a video through the Internet, the user has to search for a desired video among numerous videos provided through the Internet. That is, a user who wants to watch a TV program called "Imaginary Plus" should find a desired program video by searching a keyword using a program title of "Imaginary Plus" or the name of a program coordinator "Tak Jae-hoon".

Therefore, various attempts have been made to provide a video search service that can effectively search a user's desired video.

On the other hand, a user is often interested in some contents of a specific program. In other words, the user is often interested in the contents of a part of the program rather than the entire program.

For example, a user who wants to watch a soccer relay may want to watch only a scene in which a specific player scores a goal, rather than watching the entire soccer relay program. In addition, there may be a case where the entertainment program viewer wants to watch only a scene in which Baeulgi performs a retro dance. However, since a general video search method uses the entire soccer relay or the entire entertainment program as the search target, it is impossible to search for some scenes of the video desired by the user.

In the past, there have been attempts to search a video by using dialogue information (or subtitle information) included in the video in order to enable searching for a part of the video. In other words, the user wanted to search for a part of a specific video that has a dialogue or subtitle related to a search word input by the user, thereby searching for a part of the desired video.

However, in many cases, dialogue or subtitles alone cannot adequately reflect the content of the video. For example, if you want to find a kiss scene of a drama, because the kiss scene often has no dialogue, it is very difficult to find a scene in which drama heroes kiss only by the dialogue or subtitles. In addition, in the case of a video without dialogue or subtitles, video search using this method is not possible at all.

Accordingly, there is an urgent need for a moving picture scene retrieval method and moving picture scene retrieval system capable of appropriately retrieving a desired scene in the moving picture using the contents of the moving picture.

The present invention has been made to solve the problems of the prior art as described above, and an object of the present invention is to enable a user to effectively search for a desired scene in an entire video by using keywords related to video content.

In addition, the present invention is to divide the entire video according to the content, and search by using a script indicating the characteristics of the content for the divided section, so that the user can input a search term related to the video content to search for a scene related to the content For the purpose of

In addition, an object of the present invention is to identify a scene search keyword and automatically recognize a user's intention to perform a video scene search, thereby providing a search result corresponding to the user's intention.

In addition, an object of the present invention is to improve a user's search satisfaction by assigning two or more scripts to a scene video and using a script similar to the user's search word as a title of the scene video.

In order to achieve the above object and solve the problems of the prior art, a video scene search method according to an embodiment of the present invention, receiving a search word including a scene search keyword from a user, by identifying the scene search keyword Recognizing the search term as a video scene search; searching a scene search database in which a script indicating a feature of the content of a scene video is stored using a query word generated using the search term (the scene video is a content-based unit) And a search result for a scene video corresponding to a script matched with the query word.

In this case, the content-based unit means a unit in which the content of the corresponding part of the entire video is meaningful. In other words, the content-based unit is a unit including a meaningful part so that users want to appreciate only that part.

In addition, the video scene search method according to another embodiment of the present invention, the step of receiving a search word from the user, using the query generated using the search word, the scene search stored the narrative script indicating the characteristics of the content of the scene video Searching a database (the scene video corresponds to a part of the entire video divided into content-based units), and a search result corresponding to the scene video corresponding to the narrative script matching the query word matching the query word It characterized in that it comprises a step of providing.

In this case, two or more scripts (or narrative scripts) are allocated to one scene video, and the providing of the search result may include one or more scripts (or narrative scripts) for the scene video selected as the search result. A script (or narrative script) having the highest similarity with the search word may be set as a title of the scene video. For example, two or more scripts (or narrative scripts) allocated to one scene video may be represented by different subjects of actions for one scene video.

In addition, the video scene search system according to another embodiment of the present invention, a search term input unit for receiving a search word from a user, a query word generation unit for generating a query using the search word, a part of the entire video divided by content-based unit Scene search database for storing a script indicating the characteristics of the content of the corresponding scene video, Scene video search unit for searching the scene search database using the query, Search results for the scene video corresponding to the script matching the query Characterized in that it comprises a search result providing unit for providing a.

In the present invention, a moving picture is a concept encompassing video data related to a TV program, a movie, a sports relay, and the like, and multimedia data in which video data and audio data are combined.

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

1 is a flowchart illustrating a video scene search method according to an exemplary embodiment of the present invention.

Referring to FIG. 1, in a video scene search method according to an embodiment of the present invention, a search word including a scene search keyword is input from a user (S110).

The scene search keyword is a keyword that lets the user know that the search word input by the user is a video scene search. For example, the scene search keyword may be "scene", "situation" or "corner" or the like.

For example, a user who wants to search for a scene in which a performer named "Baesulgi" performs a retro dance in a TV program called "actual Saturday" may input "scene where Baesulgi performs retro dance" as a search word. In this case, the keyword "scene" may be a scene search keyword.

In addition, the video scene search method determines whether the scene search keyword is included in the search word input from the user (S120).

In this case, the video scene search method determines that the user inputs the keyword for searching the video scene if the scene search keyword is included in the search word input from the user. Otherwise, the user searches for the general search such as integrated search. It may be determined that a keyword is input.

At this time, the video scene search method stores scene search keywords such as "scene" or "situation" stored in the database, and if a search word is input from the user, whether the search term includes one or more of the keywords stored in the database. Can be determined.

For example, if a user enters "scenes where Baeulgi dances retro dance" as a search term, the video scene search method includes "scene" in the search term entered by the user, so the user enters a search term to search for video scenes. You can judge that. For example, if a user enters "when Lim's plunger enters Hong Jin-ho's headquarters" as a search term, the video scene search method includes a "situation" in the search term entered by the user. It may be determined that a search term has been entered.

As a result of the determination in step S120, when it is determined that the scene search keyword is not included in the search word input by the user, the video scene search method performs a general keyword search such as integrated search (S130).

If it is determined in step S120 that the scene search keyword is included in the search word input from the user, the video scene search method generates a query using the input search word (S140).

For example, if the search word input from the user is "the scene of retro dancing", the search word may be a query word as it is, or the query query "the retro dancing of the traditional dance" is subtracted from the search term only. In addition, only "Baesulgi" and "retro dance" may be extracted as a query.

After the query word is generated, the moving image scene search method searches for a scene retrieval database in which a script indicating a feature of the content of the scene moving image is stored using the generated query word (S150). In this case, the scene video corresponds to a part of the entire video divided into content-based units.

The content-based unit refers to a unit in which the content of the corresponding part of the entire video has meaning. In other words, the content-based unit is a unit including a meaningful part so that users want to appreciate only that part.

For example, if a player named Yoo-hwan sacked a unit called a game to the opponent's base and annihilated his or her worker's unit, and the content is meaningful enough for users to search only that part of the video, The part that used to board the opponent's base and annihilate the workers becomes a content-based unit.

2 and 3 are conceptual diagrams showing a relationship between an entire video and a scene video.

Referring to FIG. 2, it can be seen that the 60-minute full video 210 includes seven scene videos 221 to 227. At this time, each of the scene moving pictures 221 to 227 correspond to a predetermined content feature. That is, since each scene video (221 ~ 227) has a feature of content that can be divided in the entire video 210 is set to be divided in the full video (210).

Each of the scene videos 221 to 227 is divided into content-based units, and corresponds to a predetermined content feature. Therefore, each scene video image 221 to 227 is assigned a script representing the feature feature. In this case, the script may be a narrative script that describes a situation occurring in each scene video (221 ~ 227).

For example, the entire video 210 may be a soccer relay program. In this case, all of the seven scene videos 221 to 227 may correspond to main scenes (highlights) of the corresponding soccer game. In this case, the script assigned to the scene videos 221 ˜ 227 may be a “Park Ji-Sung hitting a crossbar”, “A scene where Lee Un-Jae runs out”, or “A scene where Lee Chun-Soo hit Maldini”.

  For example, the entire video 210 may be a drama program. In this case, the seven scene videos 221 ˜ 227 may correspond to main scenes (highlights) of the drama. In this case, the scripts assigned to the scene videos 221 to 227 may be "scenes of Daniel Henney kissing the Consort," "scene of hitting Daniel Henney," or "scene of Sam Soon drinking and singing."

As such, by dividing a main part having a characteristic of content that can be divided in the entire video into a scene video, the video scene search method according to an embodiment of the present invention provides a search target for a scene of interest to users. It can be set effectively and can reduce unnecessary scene video segmentation. Furthermore, by assigning a script indicating the characteristics of the content to the scene video, even if there is no dialogue, it is possible to appropriately search the video scene through the search word regarding the content of the video.

Referring to FIG. 3, it can be seen that the 60-minute entire video 310 is divided into six scene videos 321 to 326. In this case, the case shown in FIG. 3 may be a TV program divided into respective corners. As described above, in the TV program consisting of each corner, all the corners become one content-based unit, and when the scene videos 321 to 326 are summed together, the entire video 310 is obtained. For example, since a program such as "gag concert" is composed of corners such as "Bombang Institute" or "family of love", the entire video 310 is divided into scene videos 321 to 326 as shown in FIG. do.

In this case, since each of the scene videos 321 to 326 corresponds to each corner in the entire video, a script indicating a corner name may be assigned to each scene video 321 to 326.

For example, in the case where the entire video 310 is a TV gag program called "look up", the six scene videos 321 to 326 all correspond to the corners of the program, and the respective scene videos 321 to The script assigned to 326 may be a corner name of a corresponding program such as "image height".

As such, by dividing the entire video consisting of corners by corners and setting the scene video, the video scene search method according to an embodiment of the present invention enables users to search for video in the corner unit of a program, Use names to help you find the corner you want.

In FIG. 2 and FIG. 3, each scene video may correspond to a pointer indicating a part of which the entire video is divided, or may be newly generated by copying a part of the entire video divided.

As shown in Fig. 2 and 3, by dividing the entire video into a scene video based on the content characteristics, it is possible to effectively set the target of the keyword search for the content of the entire video, the entire video is provided for a fee, Scene videos can be provided free of charge, and various service strategies can be set up as needed. For example, if the full video is paid, but some scene video is provided free of charge, users who watched the scene video for free may be interested in the entire video and feel the desire to watch the full video for a fee. Can be.

Referring back to FIG. 1, the video scene search method according to an embodiment of the present invention provides a search result corresponding to a search word input by a user (S160).

In this case, step S160 may provide a search result for a scene video corresponding to a script matching a query when performing a video scene search. In this case, the match means that the similarity between the query and the script to be searched is equal to or higher than a predetermined level.

4 is a diagram illustrating an example of data stored in a scene search database of the present invention.

Referring to FIG. 4, the scene search database stores data 410 corresponding to a first scene video and a data 420 corresponding to a second scene video generated by dividing an entire video corresponding to a predetermined entertainment program. have.

In this case, as shown in FIG. 2, the first scene video and the second scene video may be divided into main parts having characteristics of contents that may be divided in the entire video and set as the scene video.

Data 410 corresponding to the first scene video includes a first scene script 411, a first scene start / completion time 412, and a first scene still cut 413.

The first scene script 411 is a narrative script that describes the situation of the first scene video in a narrative formula, and it can be seen that "Jung Woo-hyuk is jumping and drinking and dancing."

The first scene start / end time 412 corresponds to a playback time corresponding to the first scene video in the entire video.

The first scene still cut 413 may be a capture screen that may well represent the characteristics of the content of the first scene video. That is, since the first scene video is “the scene where Jang Woo-hyuk jumps up and dances”, the first scene still cut 413 may be an image file 13_1.jpg that captures the scene where Jang Woo-hyuk is dancing.

Similarly, the data 420 corresponding to the second scene motion picture includes a second scene script 421, a second scene start / completion time 422, and a second scene still cut 423.

The second scene script 421 is a narrative script that describes the situation of the second scene video by the narrative, and is stored as "a scene where Kang Ho Dong strikes Yoo Jae Suk."

The second scene start / end time 422 corresponds to a playback time corresponding to the second scene video in the entire video.

The second scene still cut 423 may be an image file (hit.jpg) that captures a screen on which Kang Ho Dong is hitting Yoo Jae-seok as a capture screen that may well represent characteristics of the second scene video.

According to an exemplary embodiment, the data 410 and 420 corresponding to the first and second scene moving images may also include dialogue information, image quality information, capacity information, and source information corresponding to the corresponding scene moving image, respectively.

FIG. 5 is a flowchart illustrating an example of providing a search result (S160) illustrated in FIG. 1.

Referring to FIG. 5, the search result providing step selects one of two or more scripts corresponding to the found scene video (S510).

That is, two or more scripts described with reference to FIG. 1 may be allocated to one scene video.

In this case, two or more scripts expressed for each subject of an action may be allocated to one scene video. For example, a scene video corresponding to a scene in which Rooney receives a reverse goal in the Manchester United football game may be assigned "Park Ji Sung's season 3 assist scene" and "Rooney's reverse goal scene".

Step S510 selects a script having a higher similarity to the search word among the two or more assigned scripts. In this case, the similarity may be a concept considering not only the similarity between the search word and the words included in the script but also the word order. For example, if a user performs a video scene search using "Park Ji Sung Assist Scene" as a search term, "Park Ji Sung's Season 3 Assist Scene" is selected instead of "Rooney's Reverse Hot Pot Scene", and the user selects "Rooney's Goal Scene." If a video search is performed using "," "Lunie's reverse hotpot scene" is selected instead of "Park Ji Sung's season 3 assist scene."

In operation S520, the search result providing step sets the selected script as the title of the scene video (S520).

That is, even though the video selected as the search result is the same, "Park Ji Sung's Season 3 Assist Scene" or "Rooney's Reverse Hot Pot Scene" is optionally set as the title of the scene video according to the user's search term. .

Therefore, the user can obtain a search result having a title similar to the search word input by the user, thereby improving the user's search satisfaction. In other words, when a user performs a video scene search with "Park Ji Sung assist scene", even if a video of the same scene is provided as a search result, rather than providing a scene video titled "Rooney's reverse hot pot scene", "Park Ji Sung season 3" The provision of a scene video titled "Hose Assist Scene" is more in line with the intention of the searcher.

In operation S530, the search result providing step may be provided to the user.

6 is a diagram illustrating an example of data stored in a scene search database when two or more scripts are allocated to one scene video.

Referring to FIG. 6, three scene scripts 610, a scene start / completion time 620, and a scene still cut 630 are stored in the scene search database as data corresponding to one scene video.

At this time, the three scene scripts 610 are "scenes of the kiss of rain in the love of this killer", "scenes of the rain kissing Shin Min Ah", and "scenes of Shin Min Ah a surprise kiss from the rain". The script most similar to the search term entered can be used as the title of the scene video.

7 is a diagram illustrating an example of a video scene search result page.

Referring to FIG. 7, a video scene search result page includes a player field 710 for playing a scene video, a scene video title field 720, a scene video dialogue information field 730, and an additional information field 740.

In the player field 710, a player for playing the scene video selected as the search result is displayed. At this time, the scene video may be set to be played only when the user clicks the play button of the player.

The scene video title field 720 displays the title of the scene video selected as the search result. According to the search word entered by the user, in addition to the "scene of rain kissing in the love of this killing man" shown in FIG. Can be as described above.

The scene video dialogue information field 730 displays dialogue information when there is an ambassador of the scene video selected as a search result. At this time, the displayed metabolic information can be limited to about two lines.

The additional information field 740 displays image quality information, file type information, capacity information, playback time information, screen size information, source information, and the like.

According to an exemplary embodiment, the video scene search method may search not only a scene video selected as a result of a scene video search but also a video search database in which a script for all videos is stored using a query input by a user when providing a search result. The user may be provided with a search result including the entire video selected as a match. In this case, the script for the entire video may include a title, episode, source, file URL, resolution, screen size, person, category, and still cut.

That is, the video scene search method may display both the entire video search result and the scene video search result corresponding to the search word input by the user. In this case, the video scene search method may not determine whether the scene search keyword is included in the search word input by the user.

For example, when a user searches for a "scene of kisses of love to kill" with a search term, the video scene search method is a video search database in which a script for all videos is stored as a query generated using the input search term. It may be provided together with the entire video search results searched for and the scene video search results searched for the scene search database.

At this time, the video scene search method of the present invention, when a predetermined scene video is provided to the user as a search result and playback of the corresponding scene video is completed, the user plays the entire video including the played scene video. The method may further include determining whether it is desired.

8 is a diagram illustrating a screen for confirming whether to play back the entire video including the played scene video after the playback of the scene video ends.

Referring to FIG. 8, when playback of a scene video is terminated or when the user stops playback of the scene video, the video scene search method displays a screen asking the user whether to play the entire video including the scene video.

In response to this, the user may click the "OK" button to play the entire video including the scene video, and click the "Cancel" button to not play the entire video.

As such, by providing the scene video search result and the entire video search result in association with the user, the user's interest in the entire video may be increased. In particular, if the full video is available for a fee, providing a scene video for the entire video for free makes it more likely that users who watch the scene video for free will be able to watch the full video including the scene video. By displaying a screen inquiring whether to watch the entire video including the scene video, it is possible to increase the user's selection rate for the entire video.

According to an embodiment, the video scene search method may limit the number of scene videos that can be played before the user logs in. This is because if the user plays the scene video indefinitely, there is no need to watch the entire video as a result. In particular, when the entire video is provided for a fee and the scene video is provided for free, it is very important to appropriately limit the number of times the scene video is played for the entire video. At this time, the number of times the scene video playback of the user who is not logged in may be performed by using an IP and a cookie.

In addition, when the user logs in, the scene video search method may allow playback of the scene video within a predetermined ratio of the play time of the entire video within a predetermined time. For example, the scene video search method may prevent a logged-in user from playing more than 10% of the scene video each month.

9 is a diagram illustrating a guide screen regarding a video scene search restriction.

Referring to FIG. 9, when a user who is not logged in plays a scene video more than a predetermined number of times, or a logged in user wants to play a scene video that is more than a predetermined ratio from the playing time of the entire video within a predetermined time, Notice that it displays a message.

In this way, by appropriately restricting the scene search results, it is possible to prevent scene search from being used indefinitely to prevent the paid sale of the video content or to infringe the copyright of the content provider.

The moving picture scene retrieval method according to the present invention can be implemented in the form of program instructions that can be executed by various computer means and recorded in a computer readable medium. The computer readable medium may include program instructions, data files, data structures, etc. alone or in combination. Program instructions recorded on the media may be those specially designed and constructed for the purposes of the present invention, or they may be of the kind well-known and available to those having skill in the computer software arts. Examples of computer-readable recording media include magnetic media such as hard disks, floppy disks, and magnetic tape, optical media such as CD-ROMs, DVDs, and magnetic disks, such as floppy disks. Magneto-optical media, and hardware devices specifically configured to store and execute program instructions, such as ROM, RAM, flash memory, and the like. The medium may be a transmission medium such as an optical or metal wire, a waveguide, or the like including a carrier wave for transmitting a signal specifying a program command, a data structure, or the like. Examples of program instructions include not only machine code generated by a compiler, but also high-level language code that can be executed by a computer using an interpreter or the like. The hardware device described above may be configured to operate as one or more software modules to perform the operations of the present invention, and vice versa.

10 is a block diagram illustrating a video scene search system according to an embodiment of the present invention.

Referring to FIG. 10, a scene search system according to an exemplary embodiment of the present invention may include a search term input unit 1010, a query generation unit 1020, a scene search database 1030, a scene video search unit 1040, and a search result provider. 1050.

The search term input unit 1010 receives a search term from a user. In this case, the search word input unit may receive a search word including a scene search keyword from the user. In this case, although not shown in FIG. 10, the scene search system may further include a scene search determining unit to identify a scene search keyword included in a search word input from the user to determine whether the user wants to perform a scene search. .

The query generator 1020 generates a query using a search word. In this case, the search word may be a query word as it is, or a meaningless part included in the search word may be deleted and only the remaining part may be a query word.

The scene retrieval database 1030 stores a script representing the characteristics of the contents of the scene moving picture corresponding to a part of the entire moving picture, which is divided into content-based units.

In this case, the content-based unit refers to a unit in which the content of the corresponding part of the entire video has meaning. In other words, the content-based unit is a unit including a meaningful part so that users want to appreciate only that part.

In this case, two or more scripts are assigned to one scene video, and among the two or more scripts of the scene video selected as a search result, a script having the highest similarity with the search word may be set as a title of the scene video. In this case, two or more scripts differently expressed for each subject of an action may be allocated to one scene video.

In this case, the script may include a narrative script that describes a situation occurring in the scene video.

The scene video search unit 1040 searches for the scene search database 1030 using the query generated by the query generation unit 1020.

The search result providing unit 1050 provides a search result for a scene video corresponding to a script matching a query.

Content not described in connection with the system described with reference to FIG. 10 may be omitted as it may be applied directly to embodiments related to the video scene search method.

As described above, although the present invention has been described with reference to limited embodiments and drawings, the present invention is not limited to the above embodiments, and those skilled in the art to which the present invention pertains various modifications and variations from such descriptions. This is possible.

Therefore, the scope of the present invention should not be limited to the described embodiments, but should be determined not only by the claims below but also by the equivalents of the claims.

The moving picture scene retrieval method and moving picture scene retrieval system of the present invention can enable a user to effectively search for a desired scene in the entire moving picture by using keywords related to moving picture contents.

In addition, the present invention is to divide the entire video according to the content, and search by using a script indicating the characteristics of the content for the divided section, so that the user can enter a search term related to the video content to search for scenes related to the content. Can be.

In addition, the present invention can identify the scene search keyword to automatically recognize the user's intention to perform the video scene search, thereby providing a search result corresponding to the user's intention.

In addition, the present invention can improve the user's search satisfaction by assigning two or more scripts to one scene video and using a script similar to the user's search term as the title of the scene video.

Claims (16)

delete Receiving a search word from a user; Searching a scene search database in which a script representing a feature of a scene video is stored, using a query generated using the search word, wherein the scene video corresponds to a part of the entire video divided by content-based units; And Providing a search result for a scene video corresponding to a script matching the query word Including, More than one script may be assigned to a scene video. The providing of the search result may include setting a script having the highest similarity to the search word as a title of the scene video among two or more scripts for the scene video selected as the search result. The method of claim 2, And at least two scripts expressed for each subject of an action are allocated to the one scene video. The method of claim 2, And the script comprises a narrative script describing a situation occurring in the scene video. The method of claim 2, And the scene video corresponds to a start time and a completion time set for the entire video. The method of claim 2, The scene video is a moving picture scene search method, characterized in that the newly generated video by copying a partial section of the entire video. The method of claim 2, The script scene search method comprising a corner name of a specific corner of the entire video. Receiving a search word from a user; Searching a scene retrieval database in which a narrative script representing a content characteristic of a scene video is stored by using a query generated using the search term, wherein the scene video corresponds to a part of the entire video divided into content-based units; ; And Providing a search result corresponding to a scene video corresponding to the narrative script matching the query word that matches the query word Including, More than one narrative script is assigned to a scene video, The providing of the search result may include setting a narrative script having the highest similarity with the search word among the two or more narrative scripts for the scene video selected as the search result as a title of the scene video. Way. The method according to claim 2 or 8, The search term includes a scene search keyword, The video scene search method And recognizing the search term as a video scene search by identifying the scene search keyword. The method according to claim 2 or 8, The searching of the scene search database may include searching for a video search database in which a script for the entire video is stored, using the query. Providing the search results And a search result including the entire video matched with the query word in the search result. The method according to claim 2 or 8, The video scene search method, Determining whether the user wants to play the entire video including the played scene video when the playback of the scene video provided as the search result is completed; The video scene search method further comprising. The method according to claim 2 or 8, The video scene search method And a number of scene videos that can be played before the user logs in. The method according to claim 2 or 8, The video scene search method And when the user is logged in, only playing the scene video within a predetermined ratio with respect to the playing time of the entire video within a predetermined time. A computer-readable recording medium in which a program for executing the method of any one of claims 2 to 8 is recorded. A search term input unit for receiving a search term from a user; A query generator for generating a query using the search word; A scene retrieval database for storing a script representing a feature of content of a scene movie corresponding to a part of the entire movie, divided into content-based units; A scene video search unit for searching the scene search database using the query word; And Search result providing unit for providing a search result for the scene video corresponding to the script matching the query Including, More than one script may be assigned to a scene video. And the search result providing unit sets a script having the highest similarity to the search word as a title of the scene video among two or more scripts for the scene video selected as the search result. The method of claim 15, The search term includes a scene search keyword, The video scene search system And a scene search determination unit identifying the scene search keyword and recognizing the search term as a video scene search.

Images