JP2003333500A - Personal digest delivery system - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a personal digest delivery system for efficiently producing a user's own personal digest which fully reflects the user's preference. <P>SOLUTION: This personal digest delivery system comprises: a basic analysis part 311 which conducts a basic analysis of image metadata and calculates significance, based on the basic rules for interpreting image metadata in terms of meaning; a preference reflection part 312 which calculates an increase in the significance brought about by the user's preference, based on the image metadata and information on the user's preference, and calculates significance which reflects the increase in the significance calculated by the basic analysis part 311; and a scene selection part 313 which produces a personal digest for each user by selecting image scenes based on the significance calculated by a preference reflection part. <P>COPYRIGHT: (C)2004,JPO

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a personal digest distribution system, and more particularly, to a digest related to a sports program using video metadata that describes a video scene of a sports program such as soccer or baseball in a mobile environment. The present invention relates to a personal digest distribution system for creating and distributing to a user's mobile terminal.

[0002]

2. Description of the Related Art Recently, mobile terminals such as PDAs and mobile phones have been remarkably spread. The maximum number of next-generation mobile terminals (IMT2000) currently in trial operation is 38
Text by network bandwidth of 4 Kbps,
It is possible to receive not only still images but also moving images.
It is considered possible to implement advanced functions such as reading aloud.

Services for transmitting sports program-related information such as baseball and soccer match reports to such portable terminals have been started. According to the occurrence of events such as scoring and the end of innings, the service provider creates a character string such as "1st back, G team 3-1H team. G team reverses with T player's timely hit." Then, it is distributed to the user's mobile terminal using the Internet network or the like.

With this service, even a user who cannot watch television while moving can know the progress of the program. The information sent is not the same for all users, and it is expected that personalized information will be sent for each user, reflecting the user's preferences. For that purpose, it is desired to propose a mechanism for automatically extracting a scene and generating an explanatory character string by reflecting the user's preference.

By the way, the applicant of the present application has already proposed a method of dynamically creating a personal digest for each user and generating a personalized character string.

[0006]

[Problems to be Solved by the Invention]
When creating a personal digest, a method for efficiently creating a personal digest that sufficiently reflects the user's taste has not been proposed. For example, at present, no efficient method has been proposed for creating a personal digest unique to a user that sufficiently reflects a plurality of user preference items.

[0007] Further, conventionally, a method for creating a digest for one program has been proposed, but a digest composed of the best scenes (across programs) of a plurality of programs, a digest of the best program for the user, etc. At present, there is no proposal for a method for efficiently creating.

The present invention has been made in view of the above, and proposes a personal digest distribution system capable of efficiently creating a user-specific personal digest that sufficiently reflects the taste of the user. The purpose is to

Further, the present invention has been made in view of the above, and efficiently creates a personal digest of the best scenes in a plurality of programs, a personal digest of the best program for the user, and the like. The purpose is to propose a personal digest delivery system that can do this.

[0010]

In order to solve the above-mentioned problems, the invention according to claim 1 uses a video metadata for explaining a video scene of a sports program such as soccer or baseball in a mobile environment. In a personal digest distribution system that creates a digest about a program and distributes it to a user's mobile terminal, it is important to perform a basic analysis of the video metadata based on a basic rule for semantically interpreting the video metadata. And a basic analysis unit for calculating the degree of importance, and based on the video metadata and the user's preference information, calculates the degree of importance of the user's preference, and calculates the degree of importance of the calculated user's preference as the basic Based on the preference reflection unit that calculates the importance reflected in the importance calculated by the analysis unit and the importance calculated by the preference reflection unit And the scene selection section for creating a personal digest for each user to select the video scene,
It is characterized by having.

According to the above invention, the basic analysis unit performs basic analysis of the video metadata based on the basic rule for semantically interpreting the video metadata to calculate the importance, and the preference reflection unit is Based on the video metadata and the user preference information, the increase in the degree of importance due to the user's preference is calculated, and the calculated degree of increase in the degree of importance due to the user's preference is reflected in the degree of importance calculated by the basic analysis unit. Calculated importance,
The scene selection unit selects a video scene based on the importance calculated by the preference reflection unit and creates a personal digest for each user.

The invention according to claim 2 provides the invention according to claim 1.
In the invention according to the above, the video metadata describes characters, basic event contents, status information, etc., and the basic rule is that two or more basic events are consecutive in the metadata. It is characterized by comprising a rule of a composite event in the case of performing and a rule of weighting the basic event and the composite event.

According to the above invention, the video metadata includes:
Characters, basic event contents, status information, etc. are described, and basic rules are defined as rules for compound events when two or more basic events are consecutive in metadata, and basic events. And a rule for weighting the composite event.

The invention according to claim 3 is the same as claim 1
In the invention according to the above, the preference information is a preference item name,
Alternatively, it is characterized by being configured from the preference item name and the degree of preference of the preference item. According to the above invention, the preference information is configured by the preference item name, or the preference item name and the preference degree of the preference item.

The invention according to claim 4 is the same as claim 3
In the invention according to above, when the video metadata includes the preference item, the preference reflection unit calculates an increment of importance according to the preference of the user, using the degree of preference of the preference item as a coefficient. It is characterized by According to the above invention, when the video metadata includes a preference item, the preference reflection unit calculates the increase in the degree of importance due to the preference of the user, using the degree of preference of the preference item as a coefficient.

The invention according to claim 5 provides the invention according to claim 1.
In the invention according to the first aspect, the preference reflecting unit calculates the importance by weighting the calculated increment of the importance according to the preference of the user to the importance calculated by the basic analysis unit. According to the above invention, the preference reflecting unit calculates the increment of the importance degree according to the calculated preference of the user,
The importance calculated by weighting the importance calculated by the basic analysis unit.

The invention according to claim 6 is the same as claim 1.
In the invention according to the first aspect, the scene selection unit calculates a sum of importance levels of a plurality of pieces of video metadata describing the video scene as a scene importance level of the video scene. According to the above invention, the scene selection unit calculates the sum of the degrees of importance of the plurality of pieces of video metadata describing the video scene as the scene importance of the video scene.

The invention according to claim 7 provides the invention according to claim 6.
In the invention according to above, the scene selection unit selects a video scene in one program based on the scene importance, and creates a personal digest for each user. According to the above invention, the scene selection unit selects a video scene in one program based on the scene importance and creates a personal digest for the user.

The invention according to claim 8 is the invention according to claim 7.
In the invention according to, the scene selection unit selects a video scene from a plurality of programs based on scene importance,
It is characterized by creating a personal digest for each user. According to the above invention, the scene selection unit selects a video scene from a plurality of programs based on the scene importance, and creates a personal digest for each user.

The invention according to claim 9 is the invention according to claim 7.
In the invention according to the above, the scene selection unit calculates the program importance of each program by adding the scene importance for each program of a plurality of programs, and based on the calculated program importance, the best program for each user. Is selected, and a personal digest of the selected best program is created. According to the above invention, the scene selection unit calculates the program importance of each program by adding the scene importance for each program of the plurality of programs, and based on the calculated program importance, the best program for each user. Select to create a personal digest of the selected best show.

[0021]

BEST MODE FOR CARRYING OUT THE INVENTION Preferred embodiments of a personal digest distribution system according to the present invention will be described in detail below with reference to the drawings.

(Summary of the Invention) In the present invention, a method of creating a digest for each user by reflecting a plurality of preference items such as teams and players on the importance of each video scene based on the degree of preference and the like. suggest. Here, we propose a system that distributes personal digests in a mobile environment, where the target videos are sports videos such as baseball and soccer. The system we propose is a system that uses preference information registered by the user in advance and dynamically creates and distributes a personal digest according to the progress of the match.

Each preference information is defined by (preference item name, preference degree). Specifically, it is defined as a preference item name: team, player ID, and degree of preference (0 to 1). Then, the degree of preference is reflected in the calculation of the importance of each event. For example, if there is a person who likes G team and M player, the increase in importance due to G team love,
The degree of importance of what the M player likes is distributed based on the degree of preference, and the degree of importance that combines the preferences is calculated.
This makes it possible to calculate the importance of a certain event by reflecting a plurality of pieces of preference information.

Further, according to the present invention, the degree of importance is calculated by reflecting a plurality of pieces of preference information for each user on the video scene for each program, and the video scene is ranked across the programs based on this importance. . The top-ranked video scene is the best scene in multiple programs for each user. Further, in the present invention, the sum of importance levels of video scenes is calculated for each program, and the total sum is used as the program importance level of each program. The programs are ranked based on this program importance. The program with the highest ranking is the best game (program) for each user. A digest is generated for this best game (program) to create a best digest for each user.

(Embodiment) FIG. 1 is a schematic system configuration diagram of a personal digest distribution system of the present invention. The personal digest distribution system 100 according to the present embodiment includes a scene processing unit 200 that generates and transmits a video scene (digest candidate scene) and video metadata for explaining the video scene at a relay location of a sports program, and a scene. A video scene for each user is selected based on the video metadata, the video scene, and the preference information of each user registered in advance, which are transmitted from the processing unit 200, and a personal digest for each user is selected. A digest processing unit 300 that generates and controls access to each user, and a moving image distribution processing unit 400 that stores and distributes the personal digest created by the digest processing unit 300. The user's mobile terminal and this system are connected via a wired or wireless communication line and a network. Further, each device of the present system may be connected in a communication environment capable of moving image distribution, but here, the field processing unit 200 and the digest processing unit 300 are wirelessly connected, and the digest processing unit 30.
0 and the moving image distribution processing unit 400 are connected by a dedicated line.

Here, the digest processing section 300 does not process the program video itself, but the field processing section 200.
Since only the video scenes (also created by the on-site processing unit 200) are ranked by referring to the video metadata created by, efficient digest creation can be realized. Further, by dividing the digest processing unit 300 and the moving image distribution processing unit 400, the load of each device is distributed, and the efficiency of the moving image distribution process can be improved.

A plurality of on-site processing units 200 are arranged at various relay locations, etc., create program metadata according to the progress of the program, and edit and cut out a video scene corresponding to the video metadata. Note that a known technique can be applied to the creation of the video metadata and the cutout of the edited video scene, and thus the description thereof is omitted. Further, either the video metadata or the video scene may be created first. Further, as shown in FIG. 1, the on-site processing section 200 includes a metadata input notebook PC for creating video metadata, a metadata confirmation notebook PC for checking video metadata,
An encoding server that edits and encodes a video scene of a program and uploads it to the digest processing unit 300 is provided.

The information created by the on-site processing unit 200 is immediately sent to the digest processing unit 300 each time it is created, or is sent at the program break timing. Here, the timing of dividing the program may be, for example, in the case of a baseball program, at the end of inning or at the end of the bat.

Efficient creation of video metadata can be realized by creating and transmitting video metadata as information about a video scene. Also, by transmitting the video metadata and the video scene in real time or at the timing of program break, it becomes possible to analyze the program content without a large delay.

The video metadata created in the field processor 200 describes the video scene. Specifically, the video metadata describes information such as what kind of characters are appearing in each video scene, what they did, and what characteristics they have.

The digest of this embodiment can be expressed as a ranking for each user of the video scene. As described above, since the video metadata is created as the information for explaining the video scene, it is possible to efficiently perform the analysis process for generating the digest simply by analyzing the content of the video metadata.

The digest processing unit 300 receives the video metadata and the video metadata transmitted from the site processing unit 200, manages the access information from the user, and presents a personal digest for each user. A WWW server (registered trademark) 301 having a function, a received video metadata, a basic rule for performing a basic analysis process that semantically interprets the video metadata and calculates the importance, and a processing result of the basic analysis process , And a meta DB 302 for accumulating processing results of the taste reflection processing and the like, and a user management DB 30 for accumulating preference information relating to the taste of the user in advance.
3, an IF (interface) for inputting information to the meta DB 302 and the user management DB 303, a DB server 304 having a function of managing the information, and a moving image in which a video scene transmitted from the on-site processing unit 200 is accumulated. The storage server 305 and the digest server 306 that analyzes the importance of the video scene based on the information of the video metadata stored in the meta DB 302 are provided.

Next, the flow of data for accumulating program metadata and video scenes will be described. DB server 304
Receives the video metadata from the site processing unit 200 via the WWW server 301, stores the data in the meta DB 302, and returns the result to the site processing unit 200. On the other hand, in the field processing unit 200, it is possible to create video metadata while viewing the result of the data stored in the meta DB 302. Also, the edited and cut-out video scene is transmitted from the site processing unit 200, and the received video scene is accumulated in the video storage server 305 in the digest processing unit 300.

The digest server 306 has a structure for performing analysis processing by reading the data on the DB written by the DB server 304. Digest server 3
06 checks the update of the video metadata, and when the data is updated, immediately starts the basic analysis processing for semantically interpreting the video metadata and calculating the importance, and the analysis result (important) Is written to the meta DB 302. By such processing, the digest server 306 can analyze the video metadata without a large delay from the update of the video metadata, and obtain the importance degree by the basic analysis of the video scene. In the meta DB 302, video metadata is managed on a program basis.

Next, the data flow between the digest processing unit 300 and the moving image distribution processing unit 400 will be described. WW
When the W server 301 inputs a digest creation request from a user (mobile terminal) via the network and the moving image distribution processing unit 400, the user management DB 303
The user's preference information is read from and the digest creation request is output to the digest server 306. Upon receiving this request, the digest server 30
6 executes a preference reflection process to create a personal digest for the corresponding user. The personal digest created is a personal digest created by ranking the video scenes in the program, a personal digest created by ranking the video scenes in multiple programs, or the best program selected from multiple programs. There are personal digests created for the best programs, and these digests are created according to the user's request.

The video scenes stored in the video storage server 305 are FT of the video distribution processing unit 400.
P (File Transfer Protocol)
Every predetermined time by the server (for example, every 5 minutes)
After being uploaded, the digest processing unit 300 and the video distribution processing unit 400 are synchronized.

FIG. 2 is a functional schematic diagram functionally showing the digest processing unit 300 (mainly the digest server 306). The digest processing unit 300 controls the user screen and the user's access.
The digest processing unit 300 executes basic analysis processing,
Based on a basic rule for semantically interpreting the contents of the video metadata, a basic analysis unit 311 that performs a basic analysis of the video metadata to calculate the importance, and a preference reflection process to execute the video metadata. Based on the preference information of the user, the preference reflecting unit 312 that reflects the preference of each user to the importance calculated by the basic analysis unit 311, and the importance reflected by the preference reflecting unit 312 from the user A scene selection unit 313 that ranks important events for each and selects a video scene composed of a plurality of events (video metadata) to create a personal digest.

The basic analysis unit 311 uses the basic rules of the neutral contents, which do not depend on the user's preference, on the video metadata generated in real time as the game progresses.
Basic analysis is performed in near real time to calculate importance. Since the basic analysis by the basic analysis unit 311 is based on the description of the basic rule with neutral contents, it is a neutral analysis that does not depend on the user's preference.

The preference reflecting section 312 is activated when there is a request from the user. The preference reflecting unit 312 reads the preference information of the user from the user management DB every time there is a request from the user, and calculates the increment of the preference of the user based on the video metadata and the preference information of the user. , The increment of the user's preference is reflected in the importance analyzed by the basic analysis unit 311.

The scene selection unit 313 is a preference reflection unit 312.
Selects an important video scene for each user based on the degree of importance that reflects the user's preference, and creates a personal digest for the user.

As described above, the basic analysis of the video metadata is processed in real time by the basic analysis section 311, so that only the preference reflection section 312 and the scene selection section 313 operate when requested by the user. Just do
It is possible to reduce the load of calculating the importance for each user.

Next, a concrete method of creating a personal digest will be exemplified in the case of a baseball program, and [video metadata], [basic rule], [preference information], [importance of video scene], [personal] The digest will be described in detail in this order.

[Video Metadata] First, the field processing unit 20
The video metadata created with 0 will be described. Figure 3
[Fig. 3] is a diagram showing an example of metadata for a baseball video. The metadata shown in the figure includes "scene ID", "metadata ID", "attack team name", "defense team name",
"Main player name", "Supporting player name", "Basic event name",
It is composed of "types".

Here, "scene ID" indicates the ID of a video scene that is a digest candidate. “Metadata ID” is an identifier of each video metadata. The “attack team / defense team name” is the name of the attack team / defense team when the event described by the video metadata occurs. The “leading player name” is the name of the leading player of the event represented by the video metadata.
The name of the pitcher is usually set as the "supporting player name".
The "basic event name" is a basic event represented by the video metadata. “Type” is information for explaining the basic event in detail.

The video scene is composed of a plurality of metadata. In FIG. 3, for example, “scene ID” =
The video scene of “847” has “metadata ID” = “2
It is composed of 46100 "and" 246200 ".
“Metadata ID” = “246200” represents that the pitcher I was out in the bat at the pitcher I.

[Basic Rule] The basic analysis section 311 performs basic analysis of video metadata according to the basic rule to calculate the degree of importance. The basic rule is composed of a compound event rule when two or more basic events are continuous, and a basic event and a compound event weighting rule (importance parameter). The basic rule is the basic analysis unit 31.
Change / correction is possible by 1.

FIG. 4 shows an example of a rule for compound events. In the compound event rule shown in FIG. 4, the compound event “timely” has a basic event “progress” such as “hit”, “second base”, “third base”, followed by a basic event “ It is generated when the video metadata of "additional point" is described.

As the weighting rule of importance, importance parameters such as attack level, pitcher level, and excitement level are defined. FIG. 5 is a diagram illustrating an example of the attack level importance parameter. Descriptions of the pitcher level and excitement level importance parameters are omitted.
In the attack level importance parameter shown in FIG. 5, the importance increases by “2 * attack coefficient” at the time of adding points or timely. This attack coefficient is defined as shown in FIG. This attack coefficient increases as the inning progresses, and also increases with each turn at the cleanup, so the time and the turn order can be reflected in the value of the attack level.

For example, in the video metadata shown in FIG. 3, when the metadata ID = 246500 (G team, D team, M player, additional points, solo home run), "2"
(Additional points) * attack coefficient + 8 (home run) * attack coefficient "is the importance of the attack level based on the basic analysis. The basic rules can be easily described and can be changed or modified
Highly flexible content analysis can be realized.

[Preference Information] Next, the preference information will be described in detail. The preference information is set from the UIF for users and stored in the user management DB 303. The preference information is defined in the following format. Preference information = (preference item name, degree of preference) where preference item name: team, player ID Preference degree: 0 to 1 (the higher the number, the higher the preference)

An example of the preference information is shown below. ・ Favorite player (M player, 1.0) (K pitcher, 0.8) (G player, 0.5) ・ Favorite team (G team, 0.8) (Y team, 0.5)

In the example of the preference information described above, the user may select "M
Athletes Athletes are "1.0", K pitchers are "0.
8 ”and“ G player / player ”are favored with a degree of“ 0.5 ”, and“ G team ”is“ 1.
This indicates that the user likes "0" and "Y team" with a degree of "0.8".

FIG. 7 shows an example of a user setting screen for preference information. In the user setting screen of the preference information shown in the figure, the preference item name and the degree of preference are set. In the user setting screen of the preference information shown in FIG.
In the team setting, the degree of preference is 1.0 for the top preference item, 0.8 for the second preference item, and 0.5 for the third preference item. The degree of preference can be omitted.
In that case, the degree of preference has a default value (usually 1.
0) is set. By defining the degree of preference in this way, it is possible to set finer preference information.

The preference reflecting unit 312 calculates the increment of importance according to the user's preference based on the video metadata and the preference information. Specifically, the preference reflection unit 312 performs processing by dividing it into two types of preference, a player preference and a team preference, and determines the user's preference from the degree of preference for each level such as attack level, pitcher level, and defense level. Calculate the increment of importance by preference. Then, the importance calculated by the basic analysis unit 311 is added to or multiplied by the increment according to the user's preference to calculate the importance reflecting the user's preference. Hereinafter, a method of calculating the degree of importance that reflects the user's taste will be specifically described.

(In the case of player preference) In the case of player preference, the processing is divided depending on whether or not the player is a pitcher. In the case of the above-mentioned video metadata, whether it is a pitcher or not,
Judgment is made based on whether or not the player's name is set in the "supporting player name".

If not a pitcher If the "main player name" in the video metadata is equal to the preference item, the user's preference is reflected in the attack level and excitement level. When the favorite player is not a pitcher, the increments of attack level and excitement level preferences can be calculated by the following formula.

[0057] (Attack level) preference increment = α * preference degree Where α is the basic increment (Excitement level) preference increment = α '* degree of preference Where α'is the basic increment

The preference degree is added to the importance degree obtained by the basic analysis to calculate the importance degree reflecting the user's preference. For example, in the case of preference information = (M player, 1.0), in the case of the metadata (G team, D team, M player, additional points, solo home run) with metadata ID = 246500 shown in FIG. 3, basic analysis is performed. When the attack coefficient is 1, the importance level of the attack level by is (8 + 2) = 10 points. Further, when the basic increment α = 10, the increment of the attack level preference when the M player is favored is (10 * 1) = 1.
It becomes 0. The importance of the attack level of this metadata is 10 (basic analysis) +10 (increase of preference) = 20.

In case of pitcher When the supporting player name in the video metadata is equal to the preference item, the preference of the user is reflected in the pitcher level and the excitement level. When the favorite player is a pitcher, the increment of the pitcher level and the excitement level preference can be calculated by the following formula.

(Pitcher level) preference increment = (Pitcher level) * (β * Preference degree) (Excitement level) preference increment = (Excitement level) * (β *
However, β is the basic integration coefficient

The pitcher level increases when the player strikes out or retreats. For example, if an event such as a strikeout or a tripartite withdrawal occurs when a favorite player pitches, the value of the pitcher level at that time increases. Since (β * degree of preference) is added to the value, the pitcher level value will increase reflecting the preference. Also, since the excitement level increases when a baseman or a 4th batter bats, the value of the pitcher level is (β * The value is increased by the integrated value of the (degree).

The difference in the taste reflection between the case where the player is not the pitcher and the case where the player is the pitcher is that the taste is reflected by unconditionally adding values to the attack level and the excitement level when the player is not the pitcher. On the other hand, in the case of a pitcher, the point is to integrate the values of the pitcher level and the excitement level. That is, if the favorite player is a batter or a runner, the importance of the attack level and excitement level increases by itself, but in the case of a pitcher, an important event such as strikeout occurs and the pitcher level is higher than "0". Unless the value becomes large, the importance of the pitcher level will not increase, and the importance of excitement level will not increase unless the value of the excitement level becomes larger than "0" at full base or at bat of the 4th batter.

(In case of team preference) In case of team preference,
Processing is divided depending on whether the team is an attack team.

In the case of attack team When the attack team name in the video metadata is equal to the preference item, the preference of the user is reflected in the attack level and the excitement level. When the preference team is an attack team, the increment of preference of the attack level and the excitement level can be calculated by the following formula.

[0065] (Attack level) preference increment = γ * preference level Where γ is the basic increment (Excitement level) preference increment = γ '* degree of preference Where γ'is the basic increment

In the case of the defensive team When the defensive team name in the video metadata is equal to the preference item, the preference of the user is reflected in the pitcher level and the excitement level. When the preference team is a defensive team, the preference increments of the pitcher level and the excitement level can be calculated by the following formula.

(Pitcher level) preference increment = (Pitcher level) * (φ * Preference degree) (Excitement level) preference increment = (Excitement level) * (φ *
Degree of preference) where φ is the basic integration coefficient

As described above, by changing the increment of the preference depending on the feature of each level, the feature of the preference item, and the degree of the preference, more detailed and effective reflection of the preference can be realized. By the processing described above, it is possible to reflect the preference of each user for each event (calculation of importance).

[Calculation of Importance] The preference reflection unit 312 calculates the importance based on the importance in the video metadata analyzed by the basic analysis unit 311 and the calculated increment of the user's preference.
The importance of each video scene is calculated. Each event has a degree of importance based on a basic analysis of an attack level, a pitcher level, and an excitement level, and a preference increment for reflecting the preference for each user preference item. The importance in the video metadata can be calculated by the following formula.

Importance in video metadata = Σn + ψ
* Σn ′ + ω * Σn ″, where n: importance based on basic analysis of attack level, pitcher level, and excitement level n ′: increment of preference at each level due to player preference n ″: preference at each level due to team preference Increment ψ, ω: coefficient

For example, metadata of metadata ID = 246500 (G team, D team, M player, point addition, solo home run) and metadata of metadata ID = 246400 (G team, D team, M player, start of hit) In the case of, preference information = (M player, 1.0), (G team,
0.8), the importance of the basic analysis at each level, the increment of preference, and the importance (total) in the video metadata are as shown in FIG. However, α = 1
0, α ′ = 5, β = 10, γ = 5, γ ′ = 3, φ = 1.
It is set to 5.

Coefficients α, α ', β, γ, γ', φ, ψ, ω
Can be set arbitrarily, and by setting these coefficients to desired values, it is possible to set which of the player's preference and the team's preference is reflected by the importance. This makes it possible to flexibly calculate the importance of the video metadata that reflects the user's taste.

Since there are a plurality of pieces of video metadata that describe one video scene, the preference reflection unit 312 calculates the sum of the importance levels of a plurality of pieces of video metadata that describe the same video scene, and The scene importance of is calculated. For example, in the case of a video scene with scene ID = 847, metadata ID = 246500 and metadata I
Since the metadata of D = 246400 is described,
Sum of importance of both video metadata (23.4 + 35.4)
= 58.9) is the scene importance. This makes it possible to calculate the importance of the video scene for each user.

[Personal digest] The scene selection unit 313 sorts and uses the video scenes in one program in the descending order of scene importance of the video scenes calculated by the preference reflection unit 312, that is, in the order of importance ranking. Create a personal digest for each user that reflects their tastes. At this time, the importance of the video scenes that match the user's preference items is high, but the video scenes that are determined to be important from a neutral standpoint are ranked higher. As a result, it is possible to realize a sort in which the video scenes that are determined to be important from a neutral standpoint are ranked higher while reflecting the user's preference, and the overall flow is reflected by reflecting the user's preference. It becomes possible to create a digest that can be expressed.

Further, the scene selection unit 313 responds to the user's request from a plurality of programs (a plurality of baseball programs) based on the scene importance of the video scenes in order of the scene importance of the video scenes. Create a personal digest for users by sorting. This makes it possible to rank video scenes across programs. For example, in the case of a baseball program, for example, on a weekly, monthly, or seasonal basis, by selecting a video scene from a plurality of baseball programs within the period based on the scene importance of the video scene, It is possible to select the best video scene for users on a weekly, monthly or seasonal basis.

Further, the scene selection unit 313 calculates the sum of the scene importance levels of the video scenes for each program to obtain the program importance level in response to the user's request, and the program importance level determines the best for each user. Select a show and create a personal digest for users of this best show. This makes it possible to select the best program for each user and provide a more flexible service.

The present invention is not limited to the above-described embodiments, but can be modified as appropriate without changing the gist of the invention.

[0078]

As described above, according to the first aspect of the present invention, a digest relating to a sports program is created by using the video metadata which describes the video scene of a sports program such as soccer or baseball in a mobile environment. In a personal digest distribution system that creates and distributes to a user's mobile terminal, a basic analysis of the video metadata is performed based on a basic rule for semantically interpreting the video metadata, and the importance is calculated. Based on the basic analysis unit, the video metadata and the user's preference information, the increment of the importance according to the user's preference is calculated, and the increment of the importance according to the calculated user's preference is calculated by the basic analysis unit. Based on the importance calculated by the preference reflecting unit that calculates the importance reflected in the importance, and the preference reflecting unit, a video scene is displayed. Since it has a scene selection unit that selects and creates a personal digest for each user, it is possible to efficiently create a personal digest unique to the user that sufficiently reflects the taste of the user. It is possible to propose a personal digest distribution system.

According to the invention of claim 2, in the invention of claim 1, the video metadata is:
Characters, basic event contents, status information, etc. are described, and the basic rule is a compound event rule when two or more basic events are consecutive in the metadata, Since it consists of the basic event and the weighting rule of the composite event,
In addition to the effect of the invention according to claim 1, it becomes possible to easily create a neutral importance degree according to the importance of an event in the basic analysis unit.

According to the invention of claim 3, in the invention of claim 1, the preference information is composed of a preference item name, or the preference item name and the preference degree of the preference item. Therefore, in addition to the invention according to claim 1, it becomes possible to calculate the importance of the event according to the preference item and the degree of preference of the preference item.

According to a fourth aspect of the present invention, in the third aspect of the present invention, the preference reflecting unit is configured to include: when the video metadata includes the preference item,
Since the increase in the degree of importance due to the preference of the user is calculated using the degree of preference of the preference item as a coefficient, in addition to the effect of the invention according to claim 3, the increase in the degree of importance due to the preference of the user is calculated. It is possible to calculate easily.

According to a fifth aspect of the present invention, in the first aspect of the invention, the preference reflecting section calculates, in the basic analyzing section, an increment of the degree of importance due to the calculated user's preference. Since the importance is weighted to calculate the importance, in addition to the effect of the invention according to claim 1, it is possible to calculate the importance of the event that sufficiently reflects the taste of the user. Become.

According to a sixth aspect of the invention, in the first aspect of the invention, the scene selection unit determines the sum of the importance levels of a plurality of pieces of video metadata describing the video scene as the scene of the video scene. Since the calculation is made as the importance, in addition to the effect of the invention according to claim 1, it becomes possible to easily calculate the importance of the video scene that sufficiently reflects the taste of the user.

Further, according to the invention of claim 7, in the invention of claim 6, the scene selection unit selects a video scene in one program based on the scene importance, and Since it is decided to create the personal digest of the above, in addition to the effect of the invention according to claim 6, it is possible to efficiently create the personal digest that sufficiently reflects the user's preference consisting of the video scenes in one program. It will be possible.

[0085] According to the invention of claim 8, in the invention of claim 7, the scene selection unit selects a video scene from a plurality of programs based on the scene importance, and the scene selection section for each user. Since the personal digest of (1) is created, in addition to the effect of the invention according to claim 7, it is possible to efficiently create a personal digest composed of the best video scenes of a plurality of programs.

The invention according to claim 9 relates to claim 7
In the invention according to the above, the scene selection unit calculates the program importance of each program by adding the scene importance for each program of a plurality of programs, and based on the calculated program importance, the best program for each user. Since it is decided to create a personal digest of the selected best program, in addition to the effect of the invention according to claim 7, it is possible to efficiently create a personal digest of the best program for the user. Become.

[Brief description of drawings]

FIG. 1 is a schematic system configuration diagram of a personal digest distribution system of the present invention.

FIG. 2 is a functional schematic diagram functionally showing a digest processing unit (mainly a digest server).

FIG. 3 is a diagram showing an example of video metadata for a baseball video.

FIG. 4 is a diagram showing an example of a complex event rule in the basic rule.

FIG. 5 is a diagram showing an example of an attack level importance parameter.

FIG. 6 is a diagram for explaining an attack coefficient.

FIG. 7 is a diagram showing an example of a user setting screen of preference information.

FIG. 8 is a diagram for explaining the importance of basic analysis, the preference increment, and the importance (total) in video metadata.

[Explanation of symbols]

100 Personal digest distribution system 200 On-site processing department 300 digest processing section 400 Video distribution processing unit

Front page continuation (51) Int.Cl. ⁷ identification code FI theme code (reference) H04N 7/173 640 H04N 5/91 L H04Q 7/20 H04Q 7/04 Z 7/38 H04B 7/26 109M (72) Inventor Takashi Katooka 1-3-6 Nakamagome, Ota-ku, Tokyo Within Ricoh Co., Ltd. (72) Inventor Atsushi Iizawa 1-3-3 Nakamagome, Ota-ku, Tokyo F-term within Ricoh Co., Ltd. (reference) 5B075 ND12 NK02 NS01 PQ05 PR08 5C053 FA14 FA23 FA28 HA29 LA14 5C064 BA07 BC16 BC18 BC23 BC25 BD02 BD07 BD08 5K067 AA34 BB04 DD04 DD17 DD51 EE02 EE16 FF23 FF25 KK13

Claims (9)

[Claims] 1. A personal digest distribution system for creating a digest relating to a sports program using video metadata for explaining a video scene of a sports program such as soccer or baseball in a mobile environment and delivering the digest to a user's mobile terminal. A basic analysis unit that calculates a degree of importance by performing a basic analysis of the video metadata based on a basic rule for semantically interpreting the video metadata, and the video metadata and user preference information. Based on the calculated preference increment based on the preference of the user, and a preference reflection unit that calculates the significance reflected by the calculated importance increment of the user preference to the significance calculated by the basic analysis unit. , Select a video scene based on the degree of importance calculated by the preference reflection unit to create a personal digest for each user A personal digest distribution system comprising: 2. The video metadata describes characters, basic event contents, status information, etc., and the basic rule is that two or more basic events are consecutive in the metadata. The personal digest delivery system according to claim 1, wherein the personal digest delivery system comprises a rule of a composite event in the case of performing the event and a rule of weighting the basic event and the composite event. 3. The personal digest distribution system according to claim 1, wherein the preference information includes a preference item name, or the preference item name and a degree of preference of the preference item. 4. The preference reflecting unit calculates, when the preference item is included in the video metadata, an increment of importance according to the preference of the user using the preference degree of the preference item as a coefficient. The personal digest distribution system according to claim 3, characterized in that. 5. The preference reflecting unit calculates the importance by weighting the calculated increment of the importance according to the preference of the user to the importance calculated by the basic analysis unit. Item 1. The personal digest distribution system according to Item 1. 6. The personal according to claim 1, wherein the scene selection unit calculates a sum of importance levels of a plurality of pieces of video metadata describing the video scene as a scene importance level of the video scene. Digest distribution system. 7. The personal according to claim 6, wherein the scene selection unit selects a video scene in one program based on the scene importance and creates a personal digest for each user. Digest distribution system. 8. The personal computer according to claim 7, wherein the scene selection unit selects a video scene from a plurality of programs based on scene importance to create a personal digest for each user. Digest distribution system. 9. The scene selection unit adds the scene importance for each program of a plurality of programs to calculate the program importance of each program, and based on the calculated program importance, the best program for each user. The personal digest distribution system according to claim 7, wherein the personal digest is selected and a personal digest of the selected best program is created.