KR20220101382A - Server, method and computer program for determining start point of each video content included in series content - Google Patents

Abstract

A server for determining the start point of each main part of video content included in the series comprises: a detection unit for detecting a plurality of fade-out sections and a plurality of non-fade-out sections from each video content; a representative frame selection unit for selecting a representative frame of each of the detected fade-out sections and selecting a representative frame of each of the detected non-fade-out sections; a table generation unit for generating a start point detection table including information on the representative frame of the fade-out section for each of the fade-out sections and information on the representative frame of the non-fade-out section corresponding to the fade-out section for each video content; a pattern extraction unit for extracting a similar pattern by comparing the start point detection tables generated for each video content; and a start point determination unit for determining the start point of the video content based on the similar pattern. Accordingly, it is possible to improve work productivity.

Description

A server, method and computer program that determines the start time of each main part of a plurality of video content included in a series {SERVER, METHOD AND COMPUTER PROGRAM FOR DETERMINING START POINT OF EACH VIDEO CONTENT INCLUDED IN SERIES CONTENT}

The present invention relates to a server, a method, and a computer program for extracting the start time of each main part of a plurality of video contents included in a series.

Recently, there is a lot of video content in the form of series. In such video content in the form of series, there is usually an opening section in which age information, theme song, title, preview, past story, and the like are combined in various forms.

Most users want to skip the opening section in series-type video content because they consider it unnecessary.

Currently, the skip function of the opening section is provided manually by a content manager, and the skip function of the corresponding opening section is not provided in some video contents.

Since the manual operation for skipping the opening section is performed by the content manager, it may cause waste of manpower resources and poor content management, and may cause deterioration of media service quality.

Korean Patent Publication No. 10-1517750 (Registered on April 29, 2015)

SUMMARY OF THE INVENTION The present invention is to solve the problems of the prior art, and to provide a server, a method, and a computer program for extracting the start time of the main part for skipping the opening section of each of a plurality of video contents included in a series.

However, the technical problems to be achieved by the present embodiment are not limited to the technical problems described above, and other technical problems may exist.

As a technical means for achieving the above technical problem, a server that determines the start time of each main part of a plurality of video contents included in the series according to the first aspect of the present invention, a plurality of fade-outs ( a detection unit for detecting a FadeOut) section and a plurality of non-fade-out sections; a representative frame selector for selecting a representative frame of each of the plurality of detected fade-out periods and selecting a representative frame of each of the detected non-fading-out periods; A start time detection table including information on a representative frame of a fade-out period and information on a representative frame of a non-fade-out period corresponding to the fade-out period for each of the respective video contents for each of the plurality of fade-out periods; a table generating unit to create; a pattern extraction unit for extracting similar patterns by comparing the start time detection tables generated for each video content; and a start time determiner configured to determine a start time of the plurality of video contents based on the similar pattern.

The method for determining the start time of each main part of a plurality of video content included in a series according to the second aspect of the present invention is to detect a plurality of fade-out sections and a plurality of non-fade-out sections from each video content. to do; selecting a representative frame of each of the plurality of detected fade-out periods and selecting a representative frame of each of the detected non-fading-out periods; A start time detection table including information on a representative frame of a fade-out period and information on a representative frame of a non-fade-out period corresponding to the fade-out period for each of the respective video contents for each of the plurality of fade-out periods; generating; extracting a similar pattern by comparing the start time detection table generated for each video content; and determining a start time of the plurality of video contents based on the similar pattern.

A computer program stored in a computer-readable recording medium including a sequence of instructions for determining a start time of each main part of a plurality of video contents included in the series according to the third aspect of the present invention, when executed by a computing device, each Detects a plurality of fade-out sections and a plurality of non-fade-out sections from the video content, selects a representative frame of each of the detected fade-out sections, and represents each of the detected non-fade-out sections Selecting a frame, and including information on a representative frame of a fade-out period for each of the plurality of fade-out periods for each of the video contents, and information on a representative frame of a non-fade-out period corresponding to the fade-out period a sequence of instructions for generating a start time detection table, extracting a similar pattern by comparing the start time detection table generated for each video content, and determining a start time of the plurality of video contents based on the similar pattern can do.

The above-described problem solving means are merely exemplary, and should not be construed as limiting the present invention. In addition to the exemplary embodiments described above, there may be additional embodiments described in the drawings and detailed description.

According to any one of the above-described problem solving means of the present invention, the present invention provides a server, a method and a computer program for determining the start time of the main part for skipping the opening section of each of a plurality of video contents included in the series. can

Through this, since the content manager does not need to directly search for the opening section of the video content, it is possible to reduce wasted manpower on unnecessary tasks and improve work productivity.

1 is a block diagram of a start time determination server according to an embodiment of the present invention.
FIG. 2 is a diagram for explaining the necessity of determining a start time for a main part of each video content according to an embodiment of the present invention.
3A to 3C are diagrams for explaining a method of generating a start time detection table according to an embodiment of the present invention.
4A to 4E are diagrams for explaining a method of generating a similar pattern table and a method of determining a start time of a plurality of video contents using the similar pattern table, according to an embodiment of the present invention.
5A to 5D are diagrams for explaining a method of generating a similar position table and a method of determining a start time of a plurality of video contents using the similar position table, according to an embodiment of the present invention.
6 is a flowchart illustrating a method of extracting a start time of each main part of a plurality of video contents included in a series according to an embodiment of the present invention.

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS Hereinafter, embodiments of the present invention will be described in detail with reference to the accompanying drawings so that those skilled in the art can easily carry out the present invention. However, the present invention may be embodied in several different forms and is not limited to the embodiments described herein. And in order to clearly explain the present invention in the drawings, parts irrelevant to the description are omitted, and similar reference numerals are attached to similar parts throughout the specification.

Throughout the specification, when a part is "connected" with another part, this includes not only the case of being "directly connected" but also the case of being "electrically connected" with another element interposed therebetween. . Also, when a part "includes" a certain component, it means that other components may be further included, rather than excluding other components, unless otherwise stated.

In this specification, a "part" includes a unit realized by hardware, a unit realized by software, and a unit realized using both. In addition, one unit may be implemented using two or more hardware, and two or more units may be implemented by one hardware.

Some of the operations or functions described as being performed by the terminal or device in this specification may be instead performed by a server connected to the terminal or device. Similarly, some of the operations or functions described as being performed by the server may also be performed in a terminal or device connected to the corresponding server.

Hereinafter, detailed contents for carrying out the present invention will be described with reference to the accompanying configuration diagram or process flow diagram.

FIG. 2 is a diagram for explaining the necessity of determining a start time for a main part of each video content.

Referring to FIG. 2 , a person can intuitively recognize whether each section of video content is an age information section, a publisher section, a title section, and the like, but the computer cannot recognize these sections.

Even if the computer detects a fade-out/fade-in section in the video content, the detected fade-out/fade-in section can be a standard for dividing an image, but it is not possible to recognize which section the divided section is. can't

In addition, since the fade-out/fade-in section may not exist in the age information section, the publisher section, and the title section, it may be difficult to use the fade-out/fade-in section as a separator for dividing the specific section.

For example, similar points of a fade-out/fade-in section may exist in several places between a plurality of video contents based on a series, such as the second episode and third episode, Rangse-hwan, shown in FIG. 2 . Such similarities may exist before and after the start of the video content, and it is not known which of these similarities is a section before the start of the main story. In addition, when data structures of a plurality of video contents are different from each other before the start of the main story, it is difficult to determine the start point of the main story.

For this reason, there is a need for a method of detecting the start time of the main part of each video content.

1 is a block diagram of a start time determination server 10 according to an embodiment of the present invention.

Referring to FIG. 1 , the start time determination server 10 includes a detection unit 100 , a representative frame selector 110 , a table generation unit 120 , a pattern extraction unit 130 , and a start time determination unit 140 . can do. However, the start time determination server 10 shown in FIG. 1 is only one implementation example of the present invention, and various modifications are possible based on the components shown in FIG. 1 .

Hereinafter, FIG. 1 will be described with reference to FIGS. 3A to 5D.

The detector 100 may detect a plurality of fade-out sections and a plurality of non-fade-out sections from each video content included in the series.

The representative frame selector 110 may select a representative frame of each of a plurality of fade-out sections detected for each video content, and select a representative frame of each of a plurality of detected non-fade-out sections.

The detection unit 100 extracts a vector value from each of a plurality of frames included in each video content included in the series, calculates a similarity between a vector value of each extracted frame and a vector value of a predefined black frame, and calculates A frame having a similarity equal to or greater than a preset threshold may be detected as a first frame belonging to a fade-out period.

The representative frame selector 110 selects, for each video content, a frame having the highest similarity with a predefined black frame among a first frame belonging to a fade-out period for each of a plurality of fade-out periods as a representative frame of the fade-out period. can

Referring to FIG. 3A , the detection unit 100 extracts image characteristic information from each of a plurality of frames 30 included in the first video content by using image convolution, and extracts the extracted image characteristic information. It can be converted to a vector value.

Also, the detector 100 may calculate a similarity (cosine similarity) between a plurality of frame-by-frame vector values included in the first video content and a predefined vector value of the black frame 32 . Also, the detector 100 may detect, as a fade-out frame, a frame having a calculated similarity equal to or greater than a preset threshold (eg, 0.985) among a plurality of frames included in the first piece of video content.

The table generator 120 generates index information for a first frame belonging to a plurality of fade-out sections for each video content included in the series, a similarity between the first frame and a predefined black frame, and a first frame in the video content. A fade-out frame table 34 for each video content may be generated based on the location information of .

For example, the table generating unit 120 includes index information for each frame included in the first part video content, a degree of similarity between each frame and a predefined black frame, and location information of each frame in the first part video content. A frame similarity table can be created.

The detector 100 may compare adjacent frames (eg, a frame positioned at index i and a frame positioned at index i+1) based on a frame similarity table corresponding to the first video content.

The detector 100 may classify a section in which index information for adjacent frames differs by more than a reference frame number (eg, two frames) as a fade-out section.

The detector 100 may detect a fade-out section by comparing index information for all adjacent frames based on a frame similarity table corresponding to the first video content.

At this time, since frames belonging to each detected fade-out period are consecutive frames, the detected fade-out period must be converted into a single value.

For this transformation, the table generating unit 120 performs descending sorting based on the similarity between a frame belonging to each detected fade-out section and a predefined black frame, selects a representative frame for each fade-out section, and A fade-out frame table corresponding to one piece of video content may be created.

The detection unit 100 is located in a section before a preset time from the location information of each first frame for each index information for a first frame belonging to a plurality of fade-out sections included in the fade-out frame table 34 for each video content. A frame may be detected as a second frame belonging to a non-fade-out period.

The representative frame selector 110 has the highest similarity with the vector value of the predefined black frame among the second frames belonging to the non-fade-out section for each non-fade-out section corresponding to each of the plurality of fade-out sections for each video content. A low frame may be selected as a representative frame of the non-fading-out period.

Referring to FIG. 3B , the detection unit 100 performs a predetermined time (eg, 2 seconds) before a predetermined time (eg, 2 seconds) from the position information of the representative frame 301 of each fade-out section included in the fade-out frame table corresponding to the first video content. A frame 303 located in a section of (eg, a 20 frame section corresponding to about 2 seconds before because it is 10 frames per second) may be determined as a non-fading-out frame. Here, 20 frames, which is a threshold value for determining the non-fade-out section including the non-fading-out frame, may be a heuristic value determined through observation of video content included in a plurality of series.

The table generator 120 may generate a non-fade-out frame table for a non-fade-out section by using the frame similarity table for the first video content.

Specifically, the table generating unit 120 sorts in ascending order the similarity between the frames belonging to each non-fade-out period and the predefined black frame for each non-fade-out period corresponding to each fade-out period, and among the frames sorted in ascending order. A non-fade-out frame table corresponding to the first video content may be generated by selecting a frame having the lowest similarity between predefined black frames as a representative frame of the non-fading-out section.

The table generator 120 may generate a start time detection table for each video content by using a fade-out frame table and a non-fade-out frame table corresponding to the video content for each video content.

Specifically, the table generating unit 120 generates information on a representative frame of a fade-out section for each of a plurality of fade-out sections for each video content, and information on a representative frame of a non-fade-out section corresponding to the fade-out section. It is possible to create a starting point detection table that includes. Here, the start time detection table includes index information on the representative frame of the fade-out period, index information on the representative frame of the non-fade-out period corresponding to the fade-out period, and the degree of similarity between the fade-out period and the non-fade-out period. can Also, the non-fade-out section corresponding to the fade-out section may mean a non-fade-out section immediately before a specific fade-out section.

The pattern extractor 130 may extract a similar pattern by comparing the start time detection table generated for each video content.

Referring to FIG. 3C , the pattern extracting unit 130 compares the first video content, the second video content, and the third video content, respectively, start time detection tables, and compares the first video content and the second video content. and a first similar pattern section, a second similar pattern section, and a third similar pattern section can be extracted from the third video content.

The start time determiner 140 may determine the start time of the plurality of video contents based on the extracted similar pattern.

Meanwhile, although it is known that the image feature of the representative frame of the fade-out section included in the start time detection table for each video content is a black screen, the image feature of the representative frame of the non-fade-out section is different for each frame, so the plurality of video content In order to check a similar pattern between the frames, it is necessary to check whether the frames of the non-fade-out section are the same image.

To this end, in order to check whether the frames in the non-fade-out section are the same image, the VGG16 CNN-based DISTS (Deep Image Structure and Texture Similarity) algorithm is used to check the similarity pattern table through the start point detection table for each video content. create

Specifically, the table generating unit 120 may generate a similar pattern table by comparing representative frames of non-fade-out sections included in a plurality of start time detection tables for each video content.

Referring to FIG. 4A , the table generator 120 includes a representative frame of a non-fade-out section included in the start time detection table 42 of the second video content and the start time detection table 44 of the third video content. ), a similar pattern table 400 for the second part and the third part video content may be generated by comparing the representative frames of the non-fade-out section.

For example, the table generating unit 120 may generate the similar pattern table 400 including the same or similar representative frames among the representative frames included in the start time detection table 42 and the start time detection table 44 . can Here, the same or similar representative frames may mean representative frames having a similarity of less than or equal to a preset threshold (eg, 0.12). Here, the similar pattern table 400 may include a plurality of representative frame sets including the same or similar representative frames.

In addition, referring to FIG. 4B , the table generating unit 120 includes a representative frame of a non-fade-out section included in the start time detection table 40 of the first video content, and the second and third video content. Comparing the representative frames of the non-fade-out section included in the similar pattern table 400 for and a similar pattern table 402 for the first edition, the second edition, and the third edition video content by using the similar pattern table 400 for the second edition and the third part video content. Similarly, the table generator 120 may generate the similar pattern table 402 including the same or similar representative frames among the representative frames included in the start time detection table 40 and the similar pattern table 400 .

The start time determiner 140 may determine the start time of the plurality of video contents based on the similar pattern table.

For example, referring to FIG. 4C , the number of representative frame sets included in the first similar pattern table 404 for the second and third video content is two or more, and the first and second editions are and the number of representative frame sets included in the second similar pattern table 408 for the third video content is one, and the similar pattern 406 in the first similar pattern table 404 and the second similar pattern table 408 , 410), the start time determiner 140 sets frame information corresponding to the representative frame set included in the second similar pattern table 408 to the first, second, and third video contents, respectively. can be determined as the starting point of the main part of

For example, referring to FIG. 4D , the number of representative frame sets included in the third similar pattern table 414 for the second and third video content is two or more, and the first and second editions are and when the number of representative frame sets included in the fourth similar pattern table 416 for the third video content is two or more, the start time determining unit 140 determines the first, second, and third video content Frame information corresponding to the last representative frame set 418 of the fourth similar pattern table 416 for content may be determined as the start time of each main part of the first part, the second part, and the third part video content.

The start time determiner 140 may determine the start time of the main part of the additional video content added to the series based on the start time of the main part determined from the plurality of video contents included in the series.

For example, referring to FIG. 4E , when the fourth video content is added to the first, second, and third video content constituting the series, the detection unit 100 detects the first and second episodes. The reference frame 422 corresponding to the reference value in the fourth video content may be determined based on the frame information 420 having the largest value among frame information at the start time of the main part determined in the part and the third part video content. For example, the detector 100 may determine, as the reference frame 422 , a frame having a similarity to the frame information 420 and a predetermined threshold or more among frames of the fourth video content.

In addition, the detection unit 100 includes an image of a frame (eg, the previous 5 frames) located in a preset previous section from the position of the reference frame 422 corresponding to the reference value in the fourth video content and the reference frame corresponding to the reference value. The image of 422 is first compared to perform image similarity comparison, and the image of the frame (for example, one frame thereafter) located in the interval after a preset from the position of the reference frame 422 corresponding to the reference value and the image corresponding to the reference value Image similarity comparison may be performed by comparing images of the reference frame 422 . In this case, the VGG16 CNN-based DISTS algorithm is used to compare the image similarity.

When the image similarity value is equal to or less than a preset threshold (eg, 0.12), the start time determiner 140 may determine a section including a frame having a corresponding image similarity value as the start time of the main part of the fourth video content.

Meanwhile, when the table generating unit 120 cannot generate the similar pattern table, the table generating unit 120 may generate the similar position table based on the distance difference between the representative frames of the fade-out section for each of the plurality of video contents.

For example, referring to FIG. 5A , the table generating unit 120 records the representative frame included in the fade-out frame table 500 of the second video content and the fade-out frame table 502 of the third video content. The distance difference (ie, similar position) between the included representative frames is calculated, and the similar positions of the second and third edition video content are calculated using representative frames having a distance difference within a predetermined threshold (eg, 200). Table 504 may be created.

Referring to FIG. 5B , the table generator 120 includes a representative frame included in the fade-out frame table 506 of the first video content and a representative frame included in the fade-out frame table 500 of the second video content. A first similar position can be calculated between the livers.

In addition, the table generating unit 120 is configured to set a second similar position between the representative frame included in the fade-out frame table 500 of the second video content and the representative frame included in the fade-out frame table 502 of the third video content. can be calculated.

Also, the table generating unit 120 is configured to perform a third similar position between the representative frame included in the fade-out frame table 506 of the first video content and the representative frame included in the fade-out frame table 502 of the third video content. can be calculated.

The table generating unit 120 may generate a similar position table 508 of the first part, the second part, and the third part video content based on the calculated first similar position, the second similar position, and the third similar position. have.

The start time determiner 140 may determine the start time of the main part of the plurality of video contents based on the similar position table.

For example, referring to FIG. 5C , when there is one representative frame set included in the similar position table 508 of the first, second, and third video contents, the start time determiner 140 may Frame information 510 of the representative frame set corresponding to the last index among the plurality of representative frame sets included in the similar position table 504 of the second and third video contents and the first, second, and third pieces Frame information 512 corresponding to the first index among frame information of the representative frame set included in the similar position table 508 of the part video content is the start time of the main part of the first part, the second part, and the third part video content. can be determined as

Meanwhile, referring to FIG. 5D , when the start time of the main part cannot be determined using the similar position table of a plurality of video contents, the detection unit 100 uses the similar position table of the second and third video contents. A frame corresponding to a median value (eg, 693.5=(695+692)/2) of the frame information before the start of the main part of the second part and the third part video content is determined by determining the start time of the main part of the first part video content (eg, 693 frames) may be determined as the reference frame. Here, the reference frame is used to detect the start time of the main part of the first part video content.

In addition, the detection unit 100 detects a fade-out frame first encountered while proceeding in the first direction (the 0 frame direction) and the second direction (the last frame direction) from the position of the reference frame corresponding to the reference value in the first video content. can

The start time determiner 140 may determine a frame having the smallest distance value among the detected fade-out frame and the distance value of the reference frame corresponding to the reference value as the start time of the main part of the first video content.

Meanwhile, for those skilled in the art, each of the detection unit 100, the representative frame selection unit 110, the table generation unit 120, the pattern extraction unit 130, and the start time determination unit 140 may be implemented separately, or at least one of them It will be fully understood that this may be integrated and implemented.

6 is a flowchart illustrating a method of determining a start time of each of a plurality of video contents included in a series according to an embodiment of the present invention.

Referring to FIG. 6 , in step S601 , the start time determination server 10 may detect a plurality of fade-out sections and a plurality of non-fade-out sections from each video content.

In step S603, the start time determination server 10 may select a representative frame of each of the plurality of detected fade-out sections, and select a representative frame of each of the detected non-fade-out sections.

In step S605, the start time determination server 10 provides information on a representative frame of a fade-out section for each of a plurality of fade-out sections for each video content, and information on a representative frame of a non-fade-out section corresponding to the fade-out section. It is possible to create a starting point detection table including

In step S607, the start time determination server 10 may extract a similar pattern by comparing the start time detection table generated for each video content.

In step S609, the start time determination server 10 may determine the start time of the plurality of video contents based on the similar pattern.

In the above description, steps S601 to S609 may be further divided into additional steps or combined into fewer steps, according to an embodiment of the present invention. In addition, some steps may be omitted if necessary, and the order between the steps may be changed.

An embodiment of the present invention may also be implemented in the form of a recording medium including instructions executable by a computer, such as a program module executed by a computer. Computer-readable media can be any available media that can be accessed by a computer and includes both volatile and nonvolatile media, removable and non-removable media. Also, computer-readable media may include all computer storage media. Computer storage media includes both volatile and nonvolatile, 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.

The foregoing description of the present invention is for illustration, and those of ordinary skill in the art to which the present invention pertains can understand that it can be easily modified into other specific forms without changing the technical spirit or essential features of the present invention. will be. Therefore, it should be understood that the embodiments described above are illustrative in all respects and not restrictive. For example, each component described as a single type may be implemented in a distributed manner, and likewise components described as distributed may also be implemented in a combined form.

The scope of the present invention is indicated by the following claims rather than the detailed description, and all changes or modifications derived from the meaning and scope of the claims and their equivalent concepts should be construed as being included in the scope of the present invention. .

10: server to determine when to start
100: detection unit
110: representative frame selection unit
120: table creation unit
130: pattern extraction unit
140: start time determining unit

Claims (19)

In the server for determining the start time of each main part of a plurality of video content included in the series,
a detection unit configured to detect a plurality of fade-out sections and a plurality of non-fade-out sections from each video content;
a representative frame selector for selecting a representative frame of each of the plurality of detected fade-out periods and selecting a representative frame of each of the detected non-fading-out periods;
For each video content, a start time detection table including information on a representative frame of a fade-out period for each of the plurality of fade-out periods and information on a representative frame of a non-fade-out period corresponding to the fade-out period is generated. table creation unit;
a pattern extraction unit for extracting similar patterns by comparing the start time detection tables generated for each video content; and
A start time determiner configured to determine a start time of the plurality of video contents based on the similar pattern
Which will include, the server.
The method of claim 1,
The detection unit extracts a vector value from each of a plurality of frames included in each video content, calculates a similarity between a vector value of each extracted frame and a vector value of a predefined black frame, and the calculated similarity is The server, which detects a frame having more than a set threshold as a first frame belonging to the fade-out period.
3. The method of claim 2,
The start time detection table includes index information on a representative frame of the fade-out period, index information on a representative frame of a non-fade-out period corresponding to the fade-out period, and a second interval between the fade-out period and the non-fade-out period. 1 comprising a degree of similarity.
4. The method of claim 3,
The table generating unit includes index information on the first frame belonging to the plurality of fade-out sections, a second degree of similarity between the first frame and the predefined black frame, and location information of the first frame in the video content. and generating a fade-out frame table based on the server.
5. The method of claim 4,
The detector detects a frame located in a section before a predetermined time from the position information of the first frame for each index information of the first frame included in the fade-out frame table as the second frame belonging to the non-fade-out section What to do, server.
6. The method of claim 5,
The server, wherein the representative frame selector selects a frame having the highest second similarity among the first frames belonging to the fade-out period for each of the plurality of fade-out periods as the representative frame of the fade-out period.
6. The method of claim 5,
The representative frame selector selects a frame having the lowest similarity with the vector value of the predefined black frame among the second frames belonging to the non-fade-out section for each non-fade-out section corresponding to each of the plurality of fade-out sections. What is selected as a representative frame of the fade-out section, the server.
The method of claim 1,
The table generating unit generates a similar pattern table by comparing the representative frames of the non-fade-out section included in the start time detection table for each of the plurality of video contents,
The server, wherein the start time determiner determines the start time of the plurality of video content based on the similar pattern table.
9. The method of claim 8,
When the table generating unit cannot generate the similarity pattern table, the table generating unit generates a similarity position table based on a distance difference between representative frames of the fade-out period for each of the plurality of video contents;
The server, wherein the start time determiner determines the start time of the plurality of video content based on the similar position table.
In the method of determining the start time of the main part of each of a plurality of video content included in the series,
detecting a plurality of fade-out sections and a plurality of non-fade-out sections from each video content;
selecting a representative frame of each of the plurality of detected fade-out periods and selecting a representative frame of each of the detected non-fading-out periods;
For each video content, a start time detection table including information on a representative frame of a fade-out period for each of the plurality of fade-out periods and information on a representative frame of a non-fade-out period corresponding to the fade-out period is generated. to do;
extracting a similar pattern by comparing the start time detection table generated for each video content; and
determining a start time of the plurality of video contents based on the similar pattern
A method of determining the start time of the main story, including.
11. The method of claim 10,
The detecting step
extracting a vector value from each of a plurality of frames included in each of the video contents;
calculating a similarity between the extracted vector value of each frame and a predefined vector value of the black frame; and
and detecting a frame having the calculated similarity greater than or equal to a preset threshold as a first frame belonging to the fade-out period.
12. The method of claim 11,
The start time detection table includes index information on a representative frame of the fade-out period, index information on a representative frame of a non-fade-out period corresponding to the fade-out period, and a second interval between the fade-out period and the non-fade-out period. 1 The method of determining the start time of the main story, including the degree of similarity.
13. The method of claim 12,
Fade out based on index information of the first frame belonging to the plurality of fade-out periods, a second degree of similarity between the first frame and the predefined black frame, and location information of the first frame in the video content. The method further comprising the step of generating a frame table, the main story start time determination method.
14. The method of claim 13,
Detecting a frame located in a section before a preset time from the position information of the first frame as a second frame belonging to the non-fade-out section for each index information of the first frame included in the fade-out frame table; Further comprising, a method of determining the start time of the main story.
15. The method of claim 14,
The step of selecting the representative frame is
and selecting a frame having the highest second similarity among the first frames belonging to the fade-out period for each of the plurality of fade-out periods as a representative frame of the fade-out period.
15. The method of claim 14,
The step of selecting the representative frame is
A frame having the lowest similarity with the vector value of the predefined black frame among the second frames belonging to the non-fade-out period for each non-fade-out period corresponding to each of the plurality of fade-out periods is a representative of the non-fade-out period The method of determining the starting point of the main story, comprising the step of selecting a frame.
11. The method of claim 10,
generating a similar pattern table by comparing representative frames of the non-fade-out section included in the start time detection table for each of the plurality of video contents; and
The method of determining the start time of the main story, the start time determining unit further comprising the step of determining a start time of the plurality of video contents based on the similar pattern table.
18. The method of claim 17,
generating a similarity position table based on a distance difference between representative frames of the fade-out period for each of the plurality of video contents when the similar pattern table cannot be generated; and
The method of determining the start time of the main story, the start time determining unit further comprising the step of determining a start time of the plurality of video contents based on the similar position table.
A computer program stored in a computer-readable recording medium including a sequence of instructions for extracting a start time of each main part of a plurality of video contents included in a series, the computer program comprising:
When the computer program is executed by a computing device,
Detecting a plurality of fade-out sections and a plurality of non-fading-out sections from each video content,
selecting a representative frame of each of the detected plurality of fade-out periods, selecting a representative frame of each of the detected non-fading-out periods,
For each video content, a start time detection table including information on a representative frame of a fade-out period for each of the plurality of fade-out periods and information on a representative frame of a non-fade-out period corresponding to the fade-out period is generated. do,
A similar pattern is extracted by comparing the start time detection table generated for each video content,
A computer program stored in a computer-readable medium, comprising a sequence of instructions for determining a start time of the plurality of video contents based on the similar pattern.

Images