JP2000023062A - Digest production system - Google Patents

Abstract

PROBLEM TO BE SOLVED: To efficiently produce a digest image from video information such as a video obtd. by recording a program based on audio information that synchronizes with the video information or a telop video in the video. SOLUTION: A telop extracting part 104 extracts a telop video from a program video fetched by a video fetching part 101 and on the other hand, an audio information extracting part 105 extracts a characteristic image of a characteristic scene based on, e.g. a volume value of the audio information that is synchronized with the program video and is added. A combination processing part 106 extracts a video part that newly becomes characteristic due to the step wise combination of characteristics parts which correspond to the extracted telop video and audio information and a digest producing part 108 produces a digest video based on the extracted video.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a digest creation method for creating a digest video from video information,
In particular, the present invention relates to a digest creation method suitable for use in an automatic recording system for automatically recording a broadcast program.

[0002]

2. Description of the Related Art In recent years, some CS broadcasts have begun in addition to terrestrial and BS broadcasts, so that more programs can be viewed. In the future, the number of channels is expected to increase with the digitalization of broadcasting. As a result, an environment for users to select a program that suits their tastes and preferences from more diversified channels is being prepared. At the same time as increasing the number of channels, a technical environment for high quality such as HD-TV is being prepared.

[0003]

However, with the increase in the number of channels, it is difficult to cover what programs are being broadcast. As a result, it is difficult to determine whether there is a program that one wants to watch. It has become difficult to grasp. For this reason, there has been a problem that the user misses the program he or she is interested in without even knowing that it is being broadcast.

[0004] Similarly, when recording a broadcasted program, there is a problem that a program to be watched must be selected from a large number of programs, so that the user often misses or forgets to record.

[0005] As described above, there are various problems in selecting program recording in accordance with personal tastes and preferences for the multi-channel era. In order to watch the program you want to watch,
Conventionally, there is a problem in that it is necessary to spend time for recording a program or to fast-forward manually to watch the program again from the beginning, which takes time. If the program was not the one I wanted to watch with interest, I wasted a lot of time.

[0006] The present invention has been made in consideration of the above circumstances, and its object is to obtain video information such as video recorded in a program.
An object of the present invention is to provide a digest creation system capable of efficiently creating a digest video based on sound information synchronized with the video information or a telop video in the video.

[0007]

SUMMARY OF THE INVENTION A digest creation system according to the present invention includes a sound information extracting means for extracting a characteristic image or a characteristic scene of video information based on sound information added in synchronization with the video information. And a digest creation unit for creating a digest video based on the feature image or feature scene extracted by the sound information extraction unit. In such a configuration, the characteristic image or the characteristic scene of the video information is extracted based on the sound information added in synchronization with the video information. By creating a digest from the above, it is possible to easily create a digest video including a characteristic portion related to the sound information.

Here, the characteristic image or characteristic scene is extracted by the sound information extracting means using the volume value of the sound information added in synchronization with the video information, and the volume value is continuously equal to or greater than the set value. It is only necessary to detect a state in which the time (time zone) becomes equal to or longer than the set time. For example, an image at the start of the state is extracted as a feature image, or a continuous image in the state (time zone) is extracted as a feature scene. do it.

In addition to the volume value of the sound information, the type of the sound information is classified into voice, sound effect, and music sound, and corresponds to, for example, a characteristic sound specified by the user or according to the taste of the user. It is also possible to extract an image portion to be processed as a feature image. In particular, for voices, attributes such as adults, children, men, and women are identified, and for sound effects and music, corresponding sound patterns are prepared in advance, so that the user can specify or match the tastes of the user. Image portion corresponding to the characteristic sound (eg, if the drive is a hobby user, a video scene corresponding to the running sound of the vehicle, etc.)
Is extracted at a finer level and cut out on the time axis, so that it can be edited into a digest video according to the user's taste and taste.

Further, the present invention comprises telop extracting means for extracting a telop video including character information from the video information, and digest generating means for generating a digest video based on the telop video extracted by the telop extracting means. It is characterized by having.

In such a configuration, since a telop video including character information is extracted from the video information, a digest video is created by using the telop video, whereby a digest video including a characteristic portion related to the telop is created. Can be easily created.

Here, considering that the region on the video screen where the telop is displayed is a region specific to the type of the video or the type of the telop, the time of the image data in the region is targeted for the region. If a telop image is extracted by detecting a change on the axis, efficient telop image extraction can be realized.

When the video information is a program video, it is necessary to identify two types of telops, such as breaking news telops used for news programs and the like, and enhanced telops used for variety programs and the like. Both the telops have different characteristics, such that the flash telop is displayed in a predetermined area at the top or bottom of the screen, and the emphasized telop is displayed in a large character over the entire screen in a specific conspicuous color. Therefore, for the flash telop, by checking the change in the luminance information on the time axis in the upper and lower predetermined areas on the program video screen, the time point at which the change is large is determined at the start time or the end time of the flash telop. Can be detected as On the other hand, for highlighted telops, the change state on the time axis of the color information is checked for the entire program video screen, and the state where the color separated from the past extracted colors on the time axis exceeds the set ratio is detected. By doing so, it can be detected as the start point or the end point of the emphasized telop.

The present invention further comprises a combination processing means for extracting a new characteristic image by combining the characteristic image or the characteristic scene extracted by the sound information extracting means and the telop image extracted by the telop extracting means, The digest creating means creates a digest video based on the characteristic video extracted by the combination processing means.

In such a configuration, a high-quality digest image can be created by creating a digest image by combining two types of features. In particular, by using both features in stages, for example, determine the time range of the feature scene, which is feature video information by telop display,
By extracting a video scene whose volume value is equal to or higher than the set value within the time range, a higher-quality digest video can be created.

In addition, by replacing the extracted video material for creating a digest on the time axis by the above-described digest creating means, a digest video more useful to the user can be created.

In the case of handling a digitized video, video information corresponding to the extracted features is obtained as a still image, and grouped according to each desired feature condition, so that a digest video having a wider application range can be obtained. Can be created.

[0018]

Embodiments of the present invention will be described below with reference to the drawings.

FIG. 1 is a block diagram of an information system having a digest creation system according to an embodiment of the present invention.

The digest creation system 10 records a TV program video on a VTR tape as a magnetic tape medium by the TV program reservation system 20, and performs a digest creation process on the video (video video) recorded on the VTR tape. The result is output to the TV program reservation system 20 as a digest video.

The TV program reservation system 20 operates in cooperation with the digest creation system 10. The TV program reservation system 20 receives a real video of a TV (terrestrial broadcast, satellite broadcast, etc.) and TV program information (TV program guide), and performs program video recording processing along with a TV program set by a user. is there.

By using the above two systems 10 and 20 in combination, an efficient TV program reservation system can be realized.

For example, when a user views a large amount of recorded video, it is inefficient and wastes time to watch the same time as the recorded time. Therefore, it is expected that the necessity of creating a digest video for a recorded video will increase more and more for a multi-channel in the future digital age. In such an era, the meaning of creating a digest video is that in addition to shortening the time so that the contents of the program video can be understood, there is a need for digest videos suitable for so-called personal information, including personal tastes and preferences. Thinking about the importance of gender.

FIG. 2 is a block diagram showing the configuration of the digest creation system 10.

In FIG. 2, the image capturing unit 101 includes:
(With TV program reservation system 20 in FIG. 1) Normal TV
Video information recorded from video, satellite transmission, cable television, or the like is captured, and the entire program video is stored in a video storage unit 103 of the video information management unit 102. Video information management unit 1
02 performs overall control management for creating a digest video with respect to the video information stored in the video storage unit 103.

The telop extraction unit 104 is one functional element for creating a digest video. The telop extraction unit 104 extracts a program video corresponding to a program video point (elapsed time counted from the beginning of the program video) in which the telop is displayed in the program video. The extracted still image (frame image) and a plurality of continuous moving images are extracted.

[0027] The sound information extracting unit 105 includes the telop extracting unit 1.
Similar to 04, it is one functional element for creating a digest video, and extracts a characteristic part that summarizes a program video using sound information existing in synchronization with the program video.

The combination processing unit 106 uses the telop extraction unit 104 and the sound information extraction unit 105 in a stepwise combination, and uses the feature portion of the program video extracted by telop extraction and the program video extracted based on the sound information. A digest video having a new characteristic part having the characteristic part is extracted.

Extraction information control unit (information switching unit) 107
Switches the telop extraction unit 104, the sound information extraction unit 105, and the combination processing unit 106 as appropriate, and provides the digest creation unit 108 with digest video creation.

The digest creation unit 108 creates a final digest video based on the output of the extraction information control unit 107 by using personal information including the individual's hobbies and preferences obtained by the personal information processing unit 109. Management unit 102
To be stored.

The digest information output unit 110 outputs and displays the digest video stored in the video storage unit 103 to another system (here, the TV program reservation system 20 in FIG. 1).

The user interface unit 111 serves as an interface with the user, and transfers personal information input by the user to the personal information processing unit 109 and transfers control information according to the user's instruction to the video information management unit 102. .

Next, the operation of the digest creation system 10 configured as described above will be described.

First, sound information (specifically, volume value) by the sound information extraction unit 105 in the digest creation system 10
Extraction of feature points (extraction of feature points or feature scenes) will be described with reference to the operation explanatory diagram of FIG. 3 and the flowchart of FIG.

Here, the sumo wrestling program video recorded on the recording tape (VTR tape) TP for 2 hours by the TV program reservation system 20 in FIG. And stored in the video storage unit 103. The time range R1 in the recording tape TP shown in FIG. 3 indicates one approach.

A graph obtained by tracing the sound volume value in R1 and forming a graph is shown in the R2 area. In this approach, the first frame images of the program video when the sound volume level (audio power level) becomes equal to or more than a predetermined fixed sound value (audio power level) Th1 (or exceeds Th1) are taken as D1, D2, and D3 images, respectively. Extract.

As for the status of the efforts, the audience was excited,
That is, considering that there are three scenes in the continuous time zone of the loud cheering sound, it is possible to continuously extract the program video corresponding to the time zone in which the sound volume is equal to or more than the certain value Th1. Become. When sound information is used, extracting as a temporally continuous feature scene can be a more effective feature extraction.

Therefore, the sound information extracting unit 1 in the present embodiment
05 checks the volume value of the sound information synchronized with the program video from the beginning of the recording tape TP (step S1), and when it is detected that the volume value is greater than Th1, the volume value is greater than Th1. It is checked whether or not the state continues for a predetermined time Th2 or more from that point in time, that is, whether or not a period (time zone) in which the volume value continuously becomes Th1 or more is Th2 or more (step S3). . And the sound information extraction unit 10
5 performs the next video extraction only when the state in which the volume value is equal to or greater than Th1 continues for the time Th2 or more.

First, the sound information extraction unit 105 determines whether to extract as a still image or a continuous image in accordance with the user's designation information input to the video information management unit 102 via the user interface unit 111 (step). S
4). If a still image is specified, the sound information extraction unit 105 sets the first frame image of the program video whose volume value is equal to or greater than Th1 (as a feature point based on the volume value).
Extract (step S5). On the other hand, if the continuous image is designated, the sound information extracting unit 105 sets the volume value to Th1.
The program video in the time zone described above is extracted (as a characteristic scene based on the volume value) (step S6).

The sound information extracting unit 105 repeats the above operation up to the end position of the program video recorded on the recording tape TP (step S7). That is, in the example of FIG. 3, by performing the process for the time zone R1 of each approach in the recording tape TP for the entire two hours of the recording tape TP, the characteristic points or the characteristic scenes based on the volume values can be obtained. Extract.

The monitor M1 on the operation screen in FIG. 3 displays the recording tape TP tape as it is, and the monitor M2 displays the extracted characteristic scene (here, the sound information extraction unit 105).
Is used to display a digest video based on the volume feature extracted by the above.

FIG. 5 shows a video of the sumo wrestling video program described with reference to FIG.
The video scene displaying the telop is extracted by the telop extraction unit 10
FIG. 4 is a diagram for explaining a method of extraction by No. 4.

In FIG. 5, the time range R1 of the recording tape TP shows one approach. In the example of FIG.
1, three scenes as telop images, that is, image D1
1, D12 and D13 are extracted. FIG. 6 shows a specific example of these images (telop images) D11, D12, and D13.
(A), (b) and (c) show. Image D1 in FIG.
1 is a telop image of the opponent's wrestler name displayed after calling the wrestler, image D12 of FIG. 6 (b) is a telop image of the past match record, and image D13 of FIG. Since the display telop pattern is fixed for the sumo wrestling program video, which is the telop image that is the definitive telop image when receiving the video, the 1
It is also possible to extract R1 which is one working time range.

The extraction of a telop image (screen) by the telop extraction unit 104 is basically performed as follows. First, the number of pixels for which the difference information between adjacent frame images in the time direction of the entire video screen is equal to or greater than a certain value is counted, and an image whose count value is equal to or more than a certain number is detected. Note that the difference information between adjacent frame images in the time direction for the entire video screen is binarized by a constant value,
An image in which the appearance frequency of the edge data is equal to or higher than a certain value may be detected.

In the above example of sumo wrestling, the area on the screen where the telop is displayed is fixed as shown in FIG. In such a case, by setting and checking a mask area in advance, the extraction accuracy can be significantly increased. The area on the screen where the telop is displayed is often defined by each program. Therefore, for example, it is also possible to treat the information as one of the user's hobby and taste information, that is, to set a unique mask area for each user's favorite program.

Also, the combination processing unit 106 uses a combination of a feature scene extraction method based on the sound volume value by the sound information extraction unit 105 and a feature video extraction method based on the telop extraction by the telop extraction unit 104. More efficient digest program video can be created. In other words, in the case of the above-mentioned sumo wrestling, the time range of each approach is determined based on the characteristic video information by telop display, and video scenes in which the volume value is equal to or greater than the set value Th1 are extracted within the time range. A higher quality digest video can be created.

The operation of extracting a telop video by the telop extraction unit 104 and creating a digest video is described in detail below. The details of the operation of extracting a telop video for breaking news such as news and the emphasis often used in variety programs are shown. The case of extracting a telop video will be described separately.

FIG. 7A shows a mask area set for extracting a flash telop image such as news. Here, a horizontally long rectangular area 31 at the top of the screen or a horizontally long rectangular area 32 at the bottom of the screen is set as a detection area (mask area) on the television screen 30.
Alternatively, by checking the temporal change of the image in 32, the telop image for the bulletin is extracted.

Hereinafter, the process of extracting a telop image for a flash report by the telop extraction unit 104 will be described with reference to the flowchart of FIG.

First, the telop extraction unit 104 sequentially extracts the preceding and succeeding video frames on the time axis from the beginning of the program video recorded on the recording tape TP, and extracts the pixels at the same position in the mask area (31 or 32) of both frame images. Multi-value difference information (difference images) composed of, for example, differences between the luminance data is obtained (steps S11 to S13).

Next, the telop extraction unit 104 counts the number of pixels having a certain threshold or more for the difference information (differential image) of the mask area (31 or 32) as it is in the multi-valued image (step S15). , A count value of which is determined in the area (determined by the total number of pixels in the mask area) (or a certain ratio to the total number of pixels in the mask area)
If this is the case, it is determined that a flash telop video such as news has been extracted (step S15).

With this determination method alone, extraction is performed only at the time when the telop video is displayed and when the telop display is completed. Therefore, in the present embodiment, the following method is applied.

That is, upon detecting the time point at which the telop image is displayed, the telop extraction unit 104 holds the immediately preceding frame image, that is, the frame image immediately before the telop image is displayed (steps S16 and S17), and Is turned on (step S18), and the current frame image (here, the first telop image) extracted in step S12 is extracted as the telop image (step S18). Step S19).

Thereafter, the telop extraction unit 104 extracts the subsequent frame image (step S12), and
The difference information from the frame image immediately before displaying the telop stored in step 17 is obtained (step S13). In this way, by obtaining difference information from the frame image immediately before displaying the telop, as long as the telop display continues, the number of pixels equal to or larger than the set threshold in the difference information (difference image) is continuously constant. The value is larger than the number (steps S14 and S15). Therefore, in this state, when the telop video display start flag is ON, the current frame image is extracted as the telop image (steps S16 and S19), so that all the frame images during the period in which the telop is displayed are displayed. Can be extracted as a telop image.

When the number of pixels equal to or larger than the set threshold value is less than the predetermined number, the telop extraction unit 104 turns off the telop video display start flag (step S21), and holds the current frame image (step S22). The next frame image is taken out (step S12), and the processing from step S13 is repeated. As described above, when the number of pixels equal to or larger than the set threshold value does not reach the certain number, the display returns to the normal display, and the number of pixels equal to or larger than the set threshold value in the difference information decreases.

FIG. 7B shows a mask area set to extract an emphasized telop image frequently used in variety programs and the like. Here, a rectangular area 33 of almost the entire screen is set as a detection area (mask area) on the television screen 30, and a temporal change of an image of the area 33 is checked to extract an emphasized telop video. The feature of the emphasized telop image is that the telop image is played in large characters and in a conspicuous color over the entire screen.

Therefore, in the present embodiment, as shown in the flow chart of FIG. 9, similar to the extraction of the telop image for the bulletin, the difference processing on the time axis of the image in the area 33 (steps S21, S22 to S22). In addition to performing S24), the following processing is performed.

First, when the number of pixels equal to or larger than the set threshold value in the difference information (difference image) of the mask area (33) becomes equal to or larger than a predetermined number, the immediately preceding frame image, that is, the frame image immediately before the telop is displayed. Is held and the telop video display start flag is turned ON (steps S26 to S28)
Is the same as the previous case of extracting the telop video for the bulletin. The difference from the extraction of the telop image for the flash report is that the current frame image at that point in time is a period in which the number of pixels equal to or greater than the set threshold in the difference information (difference image) is continuously a large value equal to or greater than a certain number Is extracted as a highlighted telop video candidate, and color frequency information in the frame image is obtained (step S29). In this case, color extraction can be performed stably by using hue values as color information.

The telop extraction unit 104 turns off the telop video display start flag when the number of pixels equal to or larger than the set threshold value does not reach the predetermined number and returns to the normal display, and holds the current frame image (step S30, S3
1) Compare the color frequencies of the respective emphasized telop video candidates that have been extracted up to that point, and compare them with the mask area (check area) 33.
A color having a fixed ratio is searched for (step S32). Then, when there is a color having a fixed ratio in the area 33, the telop extraction unit 104 determines that the video is the emphasized telop video and stores the video (step S34). on the other hand,
If there is no color having a certain ratio in the area 33, the telop extracting unit 104 discards the relevant video as not being an enhanced telop video.

Here, as the digest video, if the video is only the emphasized telop video, the original program video content may be difficult to understand. Therefore, by extracting a video at a predetermined time after the highlighted telop video is displayed, it is possible to make the content expressed by the emphasized telop correspond to the video.

Next, using the user's personal information provided through the user interface unit 111 with the video material extracted for digest creation, the digest creation unit 10
The method of replacing the data on the time axis at 8 will be described with reference to FIG.

First, for example, in FIG.
As the viewing style of the cooking program video 40 of 0, when the user needs to actually cook, he often wants to first see the video of the cooking name and recipe information, and refers to the cooking process as necessary. I do.

However, in the digest video normally created by the method described above, the order of the characteristic scenes is as shown in FIG. 10 in that the telop image D21 of the "cooking program title" and the telop image D22 of the "recipe" Then, the telop image D23 of “finished” is arranged in this order, and the user needs to watch the entire cooking program video 40 (here, 10 minutes).

On the other hand, the digest creation unit 108
In consideration of the user's personal information (preference), arrow 4
As shown in FIG. 1, the temporal order of the extracted characteristic scenes is changed, and for example, an image D23 of the “finished” telop displaying the name of the dish and a telop image D22 displaying the “recipe” information are arranged in this order. Arrange video scenes during the cooking process. The video scene of the cooking process is temporally sampled even if a section of the time L sandwiched between the “recipe” telop image D22 and the “finished” telop image 23 is extracted as it is. No problem.

Although the digest creation system 10 used in combination with the TV program reservation system 20 has been described above, the present invention is not limited to this.

For example, as shown in FIG. 11, the digest creation system 10 may be used by being incorporated in a system (content provider side system) 50 for providing contents such as TV program video. That is, the digest creation system 1 is used by the content provider system 50 to create a digest video of the program video requested by the user system 60.
It is also possible to use 0.

In the configuration shown in FIG. 11, the user system 60 requests the program video that the user wants to view from the content provider system 50. At this time, when the requested program video is provided from the content provider system 50, if the system 50 is configured to send all of the program video from the beginning to the user system 60, the capacity is large. It takes time to transfer the video, and there is a problem that the user really wants to watch the program video.

On the other hand, in the video-on-demand system of FIG. 11, the digest creation system 1 incorporated in the content provider system 50
With 0, a digest video of the program video specified by the user can be created in accordance with the user's request and can be provided to the user-side system 60. With this, the user first grasps the rough contents of the target program video in the digest video, and if the user wants to watch the program video really, he requests the transfer of the original video of the program video again. Good.

By the way, when the digest video is considered as one summary showing the contents of the program video, the expression of the summary of the video program clustered on the time axis toward the digitization of the video which is expected to spread in the future. However, the information need not be continuous on a similar time axis. So, for example, DVD
It is also possible to extract a movie content for use as a still image in a video corresponding to the feature information and paste it into a movie introduction scenario map prepared in advance to use it as a summary showing the content of the movie. Of course, if you are interested in the summary, you can also watch the original video.

Further, as applied in the above embodiment, each point video on the time axis, which is a feature of the program video, is extracted using the telop information and the sound information, and the digest program which is continuous on the time axis is extracted. Besides creating a point program on the time axis without editing it on the original continuous time axis, but by arbitrarily rearranging and editing it on the time axis, It allows you to see the result video immediately. As an effective example of the editing by such rearrangement, there is application to a sports program video such as baseball or soccer. Here, it is possible to think of a situation in which the result of the game is viewed first, and then the content is slowly viewed later.

The digest creation system 1 described above
0, or a processing procedure executed by the digest creation system 10, particularly a processing procedure executed by the telop extraction unit 104 and the sound information extraction unit 105, uses a computer as a set of the functional element group. A program for causing the computer to execute the processing procedure or a program for causing the computer to function.
This is realized by loading a recording medium such as a D-ROM into a computer and reading and executing the program. This program may be loaded into a computer through a communication medium such as a communication line.

[0072]

As described above in detail, according to the present invention, for video information, especially for a recorded program video, information compression on a simple time axis such as a normal VTR device, video recognition, etc. Regardless of the method that requires processing time, scenes and scenes that are characteristic of the program video are extracted using telops and sound information representing the attributes of the program video, and the digest video is extracted. Can be edited flexibly.

[Brief description of the drawings]

FIG. 1 is a block diagram of an information system configured by combining a digest creation system according to an embodiment of the present invention with a TV program reservation system.

FIG. 2 is a block diagram showing a configuration of a digest creation system 10 in FIG.

FIG. 3 is a view for explaining feature extraction based on sound information (volume value) by a sound information extraction unit 105;

FIG. 4 is a flowchart showing a processing procedure of feature extraction based on sound information (volume value) by a sound information extraction unit 105;

FIG. 5 is a view for explaining extraction of a telop video from a program video by a telop extraction unit 104;

FIG. 6 is a diagram showing an example of a telop video extracted from a sumo wrestling program video.

FIG. 7 is a diagram showing a mask area used for telop image extraction by the telop extraction unit 104, for each of the preliminary notification telop image extraction and the enhanced telop image extraction.

FIG. 8 is a diagram showing a processing procedure for extracting a telop video for a flash report by a telop extraction unit 104;

FIG. 9 is a diagram showing a processing procedure for extracting an enhanced telop video by a telop extraction unit 104;

FIG. 10 is a view for explaining an operation of replacing a video material extracted for digest creation with a digest creation unit on a time axis.

FIG. 11 is a block diagram showing a configuration example in which a digest creation system according to an embodiment of the present invention is applied to a video-on-demand system.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 10 ... Digest creation system 20 ... TV program reservation system 50 ... Content provider side system 102 ... Video information management part 104 ... Telop extraction part 105 ... Sound information extraction part 106 ... Combination processing part 107 ... Extraction information control part 108 ... Digest creation part 109 ... Personal Information Processing Department

 ──────────────────────────────────────────────────続 き Continued on the front page (72) Inventor Tetsuya Abe 70, Yanagimachi, Kochi-ku, Kawasaki, Kanagawa Prefecture Inside the Toshiba Yanagicho Plant Co., Ltd. (72) Inventor Kazunori Konda 70, Yanagicho, Saiwai-ku, Kawasaki City, Kanagawa Prefecture

Claims (6)

[Claims] 1. A sound information extracting means for extracting a characteristic image or a characteristic scene of the video information based on the sound information added in synchronization with the video information, and a characteristic extracted by the sound information extracting means. A digest creation system, comprising: digest creation means for creating a digest video based on an image or a feature scene. 2. A telop extracting means for extracting a telop video including character information from video information, and a digest generating means for generating a digest video based on the telop video extracted by the telop extracting means. Digest creation system featuring. 3. A sound information extracting means for extracting a characteristic image or a characteristic scene of the video information based on the sound information added in synchronization with the video information, and a telop video including character information from the video information. A telop extraction unit for extracting a characteristic image or a characteristic scene extracted by the sound information extraction unit, and a telop video extracted by the telop extraction unit, and a combination processing unit for extracting a new characteristic video; A digest creation system, comprising: digest creation means for creating a digest video based on the feature video extracted by the combination processing means. 4. The method according to claim 1, wherein the sound information extracting unit is configured to set a time period during which a volume value of the sound information added in synchronization with the video information continuously exceeds a predetermined set value to a predetermined set time or more. The digest creation according to claim 1 or 3, wherein a feature image of the video information at a corresponding time point or a feature scene of the video information in a corresponding time zone is extracted by detecting a state of the digest. system. 5. The telop extraction means detects a change on the time axis of image data in the region of the type of the video on the video screen or a region specific to the type of the telop video. The digest creation system according to claim 2 or 3, wherein a telop image is extracted. 6. The video information to be subjected to telop video extraction by the telop extraction means is a program video, and the telop extraction means targets a predetermined area in an upper part and a lower part on a program video screen on a time axis. Bulletin telop extraction means for extracting bulletin telop images by checking changes in luminance information, and checking the change state on the time axis of color information for the entire program video screen and extracting it in the past on the time axis 6. The digest creation system according to claim 5, further comprising: an enhanced telop extracting means for extracting an enhanced telop image by detecting a state in which a color separated from the set color is equal to or greater than a set ratio.