JP2009153144A - Event importance adjustment method and device for carrying out the method - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide an event importance adjustment device capable of extracting a scene reflecting the flow of a video content. <P>SOLUTION: An event importance adjustment device includes a storing unit for accumulating an event correlation pattern table indicative of a trigger event that indicates each of events in a video content and becomes the basis of processing start and adjustments for importance of other events, a correlation event to adjust importance thereof on the basis of the trigger event, and an importance adjustment point determined by combining the trigger event and the correlation event. It is determined, by receiving event information with importance containing the event and its importance and state information of the video content, whether the event of the event information is the trigger event or not on the basis of the table. In a case where the event is the trigger event, based on the table, a corresponding correlation event is extracted from the event information, and importance of the extracted correlation event is adjusted using the importance adjustment point. <P>COPYRIGHT: (C)2009,JPO&INPIT

Description

The present invention relates to a method for adjusting an importance level when an event with a high importance level is extracted from video content and creating a digest, and an apparatus and program for executing the method. The importance is adjusted so that an event reflecting the flow of content can be extracted.

The broadcasting station creates a digest from the video of the sports broadcast and broadcasts it as a “highlight scene so far” in the middle of the broadcast, or scratches and introduces the content of the game with sports news of the day. A sport game is a drama without a scenario, and it is not known when an important scene will appear. Therefore, when manually editing a digest, it takes a lot of time and labor to select an important scene. Attempts have been made to automatically extract this important scene based on changes in video features and cheer of the supporter, but this method accurately extracts important events that determine the flow of the game. Is difficult.

Further, Patent Document 1 below describes a digest creation device that digitizes individual events and situations in baseball based on preset contents, analyzes changes in the numerical values, and selects events to be extracted. Yes. The index creator inputs the video with the index added to this device. The index is set in detail for various events such as game start, batting start, pitching, strike, ball, first hit, second hit, and the like.

The digest creation device holds a table in which numerical information for each index, weight information added to the numerical information, and the like are set, and generates numerical video information that expresses the content of the video information by changing the numerical value. Then, an event at the peak point of this numerical change is extracted to create a digest. When this index information is multiplexed with sports video information and broadcast by digital broadcasting, the receiving device equipped with the digest creation device must store the sports broadcast video once and then generate and display the digest. Can do.
JP 2000-261754 A

However, the method described in Patent Document 1 calculates the importance of the event based only on the event and the situation at that time, so that there is a problem that the importance cannot be determined by looking at the flow of the game. . If events are extracted using the importance calculated only by the event and the situation at that time, without considering the “flow of the game”, for example, when events with high importance are dense in a certain period , Only “events that occur in a short period of time” are extracted, and in games where scoring events such as “reversal” occur many times, only “events with similar contents” are extracted. , A digest that is local and biased in content or overlaps in content will be created.

Further, the method described in Patent Document 1 has a problem that the value of the previous event cannot be re-evaluated by the later event because the connection between events is not taken into consideration. For example, hits and four balls that occurred just before the home run should be evaluated higher than normal hits and four balls in order to contribute to increasing the score at the time of the home run. The method cannot reevaluate these events.

The present invention solves such conventional problems, and extracts events reflecting the flow of video content, and also extracts correlated events, whereby a scene reflecting the flow of video content is obtained. It is an object of the present invention to provide an event importance level adjustment method that can be extracted from the program, and to provide an apparatus and a program for implementing the method.

The event importance level adjusting device according to the present invention is an event indicating individual events in video content, and serves as a trigger to start processing, and serves as a basis for adjusting the importance level of other events, and based on the trigger event. An event correlation pattern information storage unit that accumulates an event correlation pattern table in which a correlation event whose importance is adjusted and an importance adjustment point determined by a combination of the trigger event and the correlation event are indicated; Receiving event information with importance including the event and importance assigned to the event, and status information indicating the status of the video content, and attaching the importance based on the event correlation pattern table Determine whether the event included in the event information is the trigger event When the event is the trigger event, a corresponding correlation event is extracted from the event information with importance based on the event correlation pattern table, and the importance adjustment point for the importance of the extracted correlation event And an event correlation importance level adjusting unit that adjusts using the.

Therefore, it is possible to avoid time concentration of events used for the digest and duplication of contents, and it is possible to automatically produce a digest reflecting the flow of video content.

Further, the present invention uses the event information with importance indicating the importance given to the event indicating each event in the video content, and the status information indicating the status of the video content, and uses the importance. Event importance adjustment method for adjusting the importance of the event of the attached event information, which is a trigger for starting the processing and is a basis for adjusting the importance of other events, and is important based on the trigger event Using the event correlation pattern table in which the correlation event to be adjusted in degree and the importance adjustment point determined by the combination of the trigger event and the correlation event are used, the event included in the situation information is It is determined whether the event is the trigger event, and when the event is the trigger event, the event is performed. Based on the bets correlation pattern table, from the importance-event information, extracts the correlation event applicable, and subject matter to be adjusted using the importance level adjustment point for the importance of the extracted the correlation event.

As a result, an “event group” having a correlation with an event that is characteristic in expressing the flow of the video content can be easily adopted in the digest, and a digest that reflects the flow of the video content can be automatically produced.

Further, the present invention uses an event indicating an individual event in video content and event information with importance including importance assigned to the event, and status information indicating the status of the video content, A trigger that is a trigger for starting processing and is used as a basis for adjusting the importance of other events in a computer built in a device that adjusts the importance of the event information with importance, and the importance based on the trigger event The event included in the event information with importance using the event correlation pattern table in which the correlation event to be adjusted and the importance adjustment point determined by the combination of the trigger event and the correlation event are indicated. Is the trigger event, and the event is the trigger event In order to execute a procedure for extracting a corresponding correlation event from the event information with importance based on the event correlation pattern table and adjusting the importance of the extracted correlation event using the importance adjustment point The program is summarized as follows.

As a result, an “event group” having a correlation with an event that is characteristic in expressing the flow of the video content can be easily adopted in the digest, and a digest that reflects the flow of the video content can be automatically produced.

The event importance level adjusting method and the event importance level adjusting apparatus that implements the method according to the present invention increase the importance level of important events in video content so as to avoid temporal concentration of events and duplication of contents. In addition, considering the connection between events, the importance of the events can be adjusted so that correlated events are collected and adopted in the digest. Therefore, an event reflecting the flow of video content can be extracted, and an appropriate digest can be automatically created.

Embodiments according to the present invention will be described below. In each embodiment, baseball is described as an example of video content to be used, but other sports video content may be used, and the video content can be applied to video content of other various genres.

(First embodiment)
In the event importance level adjustment method according to the first embodiment of the present invention, the flow of the video content is “confirmed” according to the game pattern set based on “the presence / absence of scores of both teams” and “the transition”. It is characterized by extracting events that are “start”, “approval”, “turn”, and “conclusion”, and increasing the importance of the corresponding events. FIG. 1 shows the configuration of an event importance level adjusting apparatus 10 that implements this method. The apparatus 10 includes a receiving unit 11 that receives event information with importance added to event information indicating the content of the event, and situation information indicating the status of a score and a runner, and the received importance Stores the storage unit 12 that stores event information and status information, and the digest scenario information that sets the method for determining the event corresponding to “start”, “approval”, “turn”, and “conclusion”, and the importance increase amount of the corresponding event Storage unit 13, scenario importance level adjusting unit 14 that adjusts the importance level of event information with importance level based on digest scenario information, and output unit that outputs event information with level of importance level adjusted 17. The output destination is not particularly limited, such as a monitor, a storage device, or a printer.

The scenario importance level adjustment unit 14 is realized by a computer built in the event importance level adjustment device 10 executing a program for setting operations.

The event information and the situation information are created by an information producer who viewed the video of the baseball game, or automatically created from the video of the game. For this event information, for example, the event information with importance assigned with the importance by the digest creation device described in Patent Document 1 and the situation information are input to the reception unit 11 of the event importance adjustment device 10. Is done. As shown in FIG. 2, the event information with importance includes the “event name”, the “occurrence time” of the event, the “actor” of the event, the “actor team” indicating the execution team of the event, and the value of the event. The event name is classified as “event name”, “event attribute” indicating whether or not a score has been generated, and the result of the event. “Event result” indicating the type of the event.

Event names include, for example, “hit”, “double hit”, “home run”, “sacrifice fly”, “infield goro”, “throwing”, “dead ball”, “checking success”, “stolen success”, “strike”, “ball”, “four-ball”, “pitcher” The names of events such as “change” and “melee” are shown. In the event result, for example, the first event of the action team with an event attribute of “Yes” is indicated as “preemptive”, and when the event attribute is “Yes” and win / loss is reversed. “Reverse” for the event, “Yes” for the event attribute, “Same” for the event when the score is the same, “Yes” for the event attribute, and “Lose” from “Tie” to “Win” Shows “Katsuetsu” for the event when it changes, “Score” for an event that has none of the above, “Yes” event attribute, and “Attack” for an event that does not have any of the above It is.

As shown in FIG. 3, the situation information includes information on the score and the runner. As score information, the score of “Home Team”, the score of “Away Team”, and the time when the score situation is reached are displayed. `` Start time '', `` End time '' indicating the time when the scoring situation is lost is indicated, Runner situation is a value indicating `` Decoy situation '', `` Start time '' indicating the time when the addictive situation is reached, An “end time” indicating the time when the situation is no longer in the heel state is displayed.

In addition, as shown in FIG. 4, the digest scenario information has a score (when both teams are present, one team is present, the other team is absent, both teams are absent), both teams have points. “Same / Reverse / Score” (“/” represents “and / or”) (Same / Reverse / Succeeded, Tie / Reversal / No Strict). There are five scenario patterns (scenario pattern determination table) according to the preemptive position (first half, second half) when the other team does not. In other words, the game pattern is divided into five patterns based on “the presence / absence of scores of both teams” and “the transition”, and in each pattern, events necessary to express the flow of the game are determined, and the event The condition necessary for extracting the event is set as the “event extraction condition”. Further, the “event extraction order” is set so that the important events are acquired in order from the determined events.

Then, for each scenario pattern, the search order of the event corresponding to “start”, “approval”, “turn” and “conclusion” is set, and the search condition of the event and the importance of the obtained event are adjusted. Point is set (event extraction adjustment table). For example, if both teams have points and there is a tie / reversal / win-win, the event is searched by “pattern 1”. In "Pattern 1"
“Start”: Event when the game is first moved “Accept”: Offense and defense of both teams “Rotation”: Event that determined the game “Yui”: An event representing a counterattack of the losing team. In order to extract events suitable for the above contents, in “pattern 1”, the corresponding events are searched in the order of “turn”, “start”, “approval”, and “conclusion”. For the “roll” event, the team from which the game starts (start) to “end of the digest creation target section” (hereinafter referred to as “current”) is set as the search target section, and the winning team or the team with the same score Search for the last reverse / tie / winover event and add +10 to the importance of the event. For the “start” event, set the search target section from the start of the game to the occurrence time of the “turn” event, search for the first event in the first team, and add +8 to the importance of the corresponding event . For the “accept” event, set the search target section from the occurrence time of the “start” event to the occurrence time of the “turn” event, and search for events with high importance without being limited by the team. +7 is added to the importance of the event to be performed. For “Yu” events, search from the time when the “Turnover” event occurred until the current time is set as the search target section, and search for events with high importance in the opposite team to “Turnover”. Add +8 to importance.

If both teams have points and there is no tie / reverse / no win, the event is searched by “pattern 2”. In “Pattern 2”, each event of the settlement
“Start”: Event that determines the game “Acceptance”: Offense and defense of both teams “Turn”: Event that represents the catching up of the losing team “Yui”: Additional points of the winning team, or of the losing team Think of it as a counterattack event. In order to extract events suitable for the above contents, in “pattern 2”, the corresponding events are searched in the order of “start”, “turn”, “approval”, and “conclusion”. For the “start” event, the period from the start of the game (start) to the present is set as the search target section, the preemptive event in the preemptive team is searched, and +8 is added to the importance of the corresponding event. For the event of “turn”, the time from the occurrence time of the “start” event to the present is set as the search target section, the last scoring event in the team opposite to “start” is searched, and +10 is added to the corresponding event . For the “accept” event, set the search target section from the occurrence time of the “start” event to the occurrence time of the “turn” event, and search for events with high importance without being limited by the team. +6 is added to the importance of the event to be performed. For “Yu” events, search from the time when the “Turn” event occurred until the present to the search target section, search for events with high importance without limiting the team, and the importance of the corresponding event Add +7 to

If the score is only for one team and it is preempted in the first half, the event is searched by “pattern 3”. In “Pattern 3”, the events of the start and end
“Start”: Event that determines the game “Acceptance”: Offense and defense of both teams “Rotation”: Event that represents the counterattack of the losing team “Yui”: Additional points of the winning team, or of the losing team Think of it as a counterattack event. In order to extract events suitable for the above contents, in “pattern 3”, the corresponding events are searched in the order of “start”, “turn”, “approval”, and “conclusion”. For the “start” event, the period from the start of the game (start) to the present is set as the search target section, the preemptive event in the preemptive team is searched, and +8 is added to the importance of the corresponding event. For the event of “turn”, set the search target section from the time of occurrence of the “start” event to the present, search for the event with high importance in the team opposite to “start”, and add +10 to the corresponding event Add. For the “accept” event, set the search target section from the occurrence time of the “start” event to the occurrence time of the “turn” event, and search for events with high importance without being limited by the team. +5 is added to the importance of the event to be performed. For “Yu” events, search from the time when the “Turn” event occurred until the present to the search target section, search for events with high importance without limiting the team, and the importance of the corresponding event Add +5 to

If the score is only in one team and it is preempted in the second half, the event is searched by “pattern 4”. In "Pattern 4"
“Start”: Attack and defense of both teams “Acceptance”: Attack and defense of both teams “Turn”: Event that determined the game “Yui”: Consider an event representing a counterattack of the losing team. In order to extract an event suitable for the above contents, in “pattern 4”, the corresponding event is searched in the order of “turn”, “approval”, “start”, and “conclusion”. For the event of “turn”, the period from the start of the game (start) to the present is set as the search target section, the first event in the first team is searched, and +8 is added to the importance of the corresponding event. For the “acceptance” event, the period from the start of the game (start) to the time of occurrence of the “turn” event is set as the search target section, the team is not limited and searches for events with high importance, and the corresponding event Add +5 to importance. For the “start” event, the period from the start of the game (start) to the occurrence time of the “approval” event is set as the search target section, the team is not limited and searches for high importance events, and the corresponding event Add +6 to importance. For “Yu” events, search from the time when the “Turnover” event occurred until the current time is set as the search target section, and search for events with high importance in the opposite team to “Turnover”. Add +5 to importance.

If both teams have no score, the event is searched by “pattern 5”. In “Pattern 5”, the events of the start and end are considered as events representing the offense and defense of both teams, and in order to extract them, the corresponding events are searched in the order of “turn”, “approval”, “start”, and “end”. For the event of “turn”, the period from the start of the game (start) to the present is set as the search target section, the team is not limited and a high-importance event is searched, and +7 is added to the importance of the corresponding event. For the “acceptance” event, the period from the start of the game (start) to the time of occurrence of the “turn” event is set as the search target section, the team is not limited and searches for events with high importance, and the corresponding event Add +5 to importance. For the “start” event, the period from the start of the game (start) to the occurrence time of the “approval” event is set as the search target section, the team is not limited and searches for high importance events, and the corresponding event Add +6 to importance. For “Yu” events, search from the time when the “Turn” event occurred until the present to the search target section, search for events with high importance without limiting the team, and the importance of the corresponding event Add +5 to

The flowchart of FIG. 5 shows the processing procedure of the event importance level adjusting device 10. The receiving unit 11 receives the event information with importance and the situation information, and accumulates them in the accumulating unit 12 (step 1). The scenario importance level adjustment unit 14 acquires the digest scenario information from the digest scenario information storage unit 13 at a designated timing (step 2), and based on the digest scenario information, the event importance level included in the event information with importance level Is recalculated (step 3).

The processing in the scenario importance adjustment unit 14 is performed in detail according to the procedure shown in FIG. That is, information for applying the “scenario pattern determination table” of the digest scenario information is read from the acquired event information with importance and situation information, and the scenario pattern is determined according to this table (step 31). Next, “extraction order” information of the determined scenario pattern is acquired (step 32), and according to the “extraction order”, event extraction of digest scenario information of “start”, “approval”, “turn”, and “conclusion” is performed. An adjustment table is acquired (step 33), and the corresponding event is extracted based on the event extraction table (search target section, search target team, search target event information) indicated in the event extraction adjustment table (step 33). 34). Next, the importance value of the extracted event is adjusted according to the event importance adjustment point (P) indicated in the event extraction adjustment table of the digest scenario information (step 36). The procedure of steps 33 to 35 is repeated until all events corresponding to “start”, “approval”, “turn” and “conclusion” are extracted (step 36). The output unit 17 outputs event information with importance obtained by adjusting the value of event importance by the scenario importance adjustment unit 14 (step 4).

Fig. 7 HYPERLINK "http://www6.ipdl.inpit.go.jp/Tokujitu/tjitemdrw.ipdl?N0000=234&N0500=4E#N/;%3e7?:?:86///&N0001=10&N0552=9&N0553=000009 "\ t" tjitemdrw "~ Figure 11 shows the results of" Start "," Confirm "," Turn ", and" Conclusion "in the case of a game development corresponding to each scenario pattern using the information of the scoreboards of Team A and Team B It shows the corresponding event or where to select it. Here, it is assumed that digests are created in seven tables. FIG. 7 shows a case where the game development corresponds to the scenario pattern 1, the event of the 5th table that determined the game is selected as the “turn” event, and the first event that the game moved first is “ The event with the highest priority in the range from the 2nd table to the 4th is selected as the event of “Start”, and the event of “Yu” is It is selected as an event that shows the counterattack of the team losing the most important event in the sixth round.

FIG. 8 shows a case where the game development corresponds to scenario pattern 2 and the event in the first table that selected the game is selected as the “start” event, and the event in the back of the fourth event showing the follow-up of the losing team Is selected as the event of “turn”, the event of “approval” is selected as the most important event in the range from the first to the fourth time, and the event of “conclusion” is 5 times In the range from the front to the sixth round, the event with the highest importance is selected when the winning team raises an additional point or when the losing team strikes back.

FIG. 9 shows a case in which the game development corresponds to scenario pattern 3, and the first event that determined the game is selected as the “start” event, and it is lost from each of the two to six tables. The event that shows the counterattack of a certain team is selected as an event of “turn” (here, the 5th table is determined to be a “turn” event based on the assumption that an event with high importance exists in the 5th table) The event with the highest importance is selected from the 2nd table to the 4th back, and the “Yu” event is from the 5th back to the 6th back. Among them, a highly important event is selected when a winning team raises an additional point or when a losing team strikes back.

FIG. 10 shows a case where the game development corresponds to scenario pattern 4, and the event of the 5th time that determined the game is selected as the “turn” event, and the range from the first to the fourth time is within the range. An event with a high degree of importance is selected as an event of “start” and “acceptance”, and an event with a high degree of importance in the 6th table indicating the counterattack of the losing team is selected as an event of “conclusion”. Also, FIG. 11 shows a case where the game development corresponds to scenario pattern 5, and the events of high importance are “start”, “approval”, “change”, “consequence” within the range from the first table to the sixth back. Is selected as the event corresponding to "."

  In this way, by comparing the game development to a scenario, searching for the event corresponding to the “start”, “approval”, “turn”, and “conclusion” of the scenario, and adjusting the importance of the corresponding event, the flow of the match It is possible to automatically create a reflected digest. In this method, an event extraction range that considers the flow of the game is set, and the events are picked up by being distributed in the game, so that it is possible to improve the extraction of only the events that occurred in a short period of time (" Solving the problem of time crowding). In addition, because meaningful combinations of events are extracted in a form that fits the “consolidation”, it is possible to improve that only events with similar contents are extracted redundantly (solving the problem of “duplication of contents”) ). Therefore, by using this adjusted event information with importance, it is possible to create a digest that can give an overview of the flow of the game, rather than a local digest.

(Second Embodiment)
The event importance level adjustment method according to the second embodiment of the present invention is characterized in that the importance level of an event is adjusted in consideration of the correlation of events. FIG. 12 shows the configuration of the event importance level adjusting apparatus 10 that implements this method. The apparatus 10 includes a receiving unit 11 that receives event information with importance added to event information indicating the content of the event, and situation information indicating the status of a score and a runner, and the received importance Storage unit 12 for storing attached event information and situation information, a storage unit 15 for storing an event correlation pattern table indicating importance adjustment points of correlated events, and an importance level based on the event correlation pattern table An event correlation importance level adjustment unit 16 that adjusts the importance level of the attached event information, and an output unit 17 that outputs the event information with the importance level that has been adjusted in importance level. The event correlation importance adjustment unit 16 is realized by a computer built in the event importance adjustment device 10 executing a program for setting operations.

The data structure of the event information with importance and the situation information input to the receiving unit 11 is the same as that of the first embodiment (FIGS. 2 and 3). In the event correlation pattern table, as shown in FIG. 13, the “trigger event name” of the trigger event that triggers the re-evaluation of the event importance, and the search range of the event to be re-evaluated related to the trigger event are displayed. “Correlation range” to indicate, “Correlation target event” to indicate the event name related to the trigger event, and “Correlation event importance adjustment point” for adjusting the importance of the correlation target event.

For example, if the home run is a trigger event, the shorter of the period from the start of inning to the trigger event occurrence time, or the period from the start time of the dredging situation that the runner first encountered to the trigger event occurrence time, whichever is shorter When a 3 strike of a correlated event is searched from the correlation range, +7 is added to the event importance, and when a boke of the correlated event is searched from the correlation range, +5 is added to the event importance. If the hit is a trigger event, the immediately preceding event is set as the correlation range, and if sacrificing is performed immediately before, +6 is added to the importance of the event, and the sending bunt is performed immediately before When that happens, add +3 to the event importance.

The flowchart of FIG. 14 shows the processing procedure of the event importance level adjusting device 10. The receiving unit 11 receives the event information with importance and the situation information, and accumulates them in the accumulating unit 12 (step 1). The event correlation importance level adjustment unit 16 refers to the event correlation pattern table stored in the event correlation pattern table storage unit 15 and acquires information on the trigger event set therein (step 5). Then, from the received event information, it is identified whether or not the event is a trigger event (step 6). If the event is a trigger event, a correlation target event is extracted and its importance is adjusted. "Event extraction adjustment process" is executed (step 7).

This correlation event extraction adjustment process is performed in detail according to the procedure shown in FIG.

In other words, the correlation target search range is specified by the “correlation range” information in the event correlation pattern table (step 71), and the “correlation target event” (trigger event) set in the event correlation pattern table in the correlation range. (Events affected by) are searched (step 72). If there is no correlation target event in this correlation range, the process ends. If it exists (step 73), the “correlation event importance adjustment table” in the event correlation pattern table is acquired (step 74) and extracted. The importance level of the event is adjusted based on the correlation event importance level adjustment table (step 75). The output unit 17 outputs event information with importance obtained by adjusting the value of event importance by the event correlation importance adjustment unit 16 (step 8).

16 and 17 show specific examples of correlation event extraction adjustment processing. FIG. 16 shows the events and runner situations indicated in the event information with importance and the situation information in time series (p indicates the event importance). Event information with importance is accumulated in the order of (1) to (6). Since “home run” is set as a trigger event in the event correlation pattern table, the correlation event extraction adjustment process is started when “(6) home run” is received. First, “correlation range” information of the event correlation pattern table is acquired. Since the runners have been wiped out during double play, the event that the runners encountered for the first time is “(4) Four Balls”, and the correlation range is “(4) Four Balls” to “(6) Home Runs” as shown in FIG. Until "."

Based on the “correlation target event” information in the event correlation pattern table, events of “correlation range” are extracted. In this correlation range, there are “(3) four balls” and “(5) hits” of correlation target events, so these two events are extracted, and the “correlation event importance adjustment table” of the event correlation pattern table 7 is added to the event importance (p = 4) of “(3) Four Spheres”, and the event importance of “(3) Four Spheres” is adjusted to “11”. Similarly, 7 is added to the event importance (“P = 5”) of “(5) hit”, and the event importance of “(5) hit” is adjusted to “12”. As a result of adjusting the event importance level based on the event correlation, the importance levels of the respective events are as shown in FIG.

As described above, the event importance level adjusting device 10 re-evaluates the value of the event in the specified range based on the event correlation pattern table when the event serving as the trigger occurs, and at the moment when the event occurs. Event importance can be adjusted not only in terms of evaluation, but also in terms of whether it is a live (connected) event. In this method, the value of the previous event is re-evaluated by the later event, and the event importance is adjusted in consideration of the connection of the event. Therefore, by using this adjusted event information with importance, , It is possible to automatically create a digest considering the flow of the game.

(Third embodiment)
The third embodiment of the present invention is characterized in that event importance is adjusted by combining the methods of the first embodiment and the second embodiment. FIG. 18 shows the configuration of the event importance level adjusting apparatus 10 that implements this method. The apparatus 10 includes a receiving unit 11 that receives event information with importance added to event information indicating the content of the event, and situation information indicating the status of a score and a runner, and the received importance Accumulating unit 12 for accumulating event information and status information, a storage unit 13 for storing digest scenario information, and a scenario importance adjusting unit 14 for adjusting the importance of event information with importance based on the digest scenario information The event correlation pattern table is stored, the event correlation importance level adjustment unit 16 that adjusts the importance level of the event information with importance level based on the event correlation pattern table, and the importance level adjustment. And an output unit 17 that outputs event information with importance.

The data structure of the event information with importance and the situation information input to the receiving unit 11 is the same as that of the first embodiment (FIGS. 2 and 3). The digest scenario information is the same as that described in the first embodiment (FIG. 4), and the event correlation pattern table is the same as that described in the second embodiment (FIG. 13).

The flowchart of FIG. 19 shows the processing procedure of the event importance level adjusting device 10. The receiving unit 11 receives the event information with importance and the situation information, and accumulates them in the accumulating unit 12 (step 1). The scenario importance level adjustment unit 14 acquires the digest scenario information from the digest scenario information storage unit 13 at a designated timing (step 2), and based on the digest scenario information, the event importance level included in the event information with importance level Is recalculated (step 3).

As described in the first embodiment, the process of the scenario importance level adjustment unit 14 is performed according to the procedure shown in FIG. 6, and the scenario importance level adjustment unit 14 performs “start”, “approval”, “transfer”, “concatenation”. Are extracted, and the importance of those events is adjusted.

The event correlation importance level adjustment unit 16 includes situation information and event information with importance level in which the importance levels of events corresponding to “start”, “approval”, “turn”, and “conclusion” are adjusted by the scenario importance level adjustment unit 14. Obtain (step 4). Further, the event correlation pattern table stored in the event correlation pattern table storage unit 15 is referred to, and the trigger event information set therein is acquired (step 5). Then, it is identified whether or not the event corresponding to “start”, “approval”, “turn” and “conclusion” is a trigger event (step 6). If the event is not a trigger event, the obtained event information with importance is output. 17 to output. If the event corresponding to “start”, “approval”, “turn”, “conclusion” is a trigger event, execute the “correlation event extraction adjustment process” to extract the correlation target event and adjust the importance of the obtained event (Step 7).

This correlation event extraction adjustment process is performed according to the procedure shown in FIG. 15, as described in the second embodiment. The event correlation importance adjustment unit 16 outputs event information with importance obtained by adjusting the value of the event importance by the correlation event extraction adjustment process to the output unit 17, and the output unit 17 includes the scenario importance adjustment unit 14 and the event correlation. Event information with importance whose importance is adjusted by the importance adjusting unit 16 is output (step 8).

20 to 23 show specific examples of processing in this apparatus. FIG. 20 shows the event and score information indicated in the event information with importance and the situation information in time series (p indicates the event importance). Further, FIG. 21 displays scoreboard information, and the numbers (1) to (6) indicate innings in which events of high importance shown in FIG. 20 have occurred. Based on the event extraction adjustment table, the scenario importance adjustment unit 14 extracts the next event as each event of “start”, “turn”, “approval”, and “conclusion”, and sets the importance of the event as follows: adjust.
“Start” event: “(1) A Preemptive HR” event Adjusted from 10 to 18 importance “A” event: “(3) A score H” event Adjusted from 8 to 14 “Change” event: “(5 ) B scoring H ”event Adjusted severity 6 → 16“ Yu ”event:“ (6) A scoring HR ”event Severity 5 → Adjusted to 12 The inning where each of the events occurred is displayed. The event correlation importance adjustment unit 16 selects an event that can be a trigger event from the events corresponding to “start”, “approval”, “turn”, and “conclusion” based on the event correlation pattern table. The next event is selected as the trigger event.

“Start” event: “(1) A preemptive HR” event “Yui” event: “(6) A score HR” event Then, the event correlation importance adjustment unit 16 performs correlation event extraction adjustment processing, and Adjust the importance of correlated events.

Correlated events of “(1) A Preemptive HR”:
“H (1)” event severity +7
“H (2)” event severity +7
Correlation target event of “(6) A score HR” “3BH (8)” event Importance level +7
FIG. 23 shows the importance of each event adjusted by the scenario importance adjusting unit 14 and the event correlation importance adjusting unit 16.

As described above, in the event importance level adjusting device 10, according to the game pattern set based on “the presence / absence of scores of both teams” and “the transition thereof”, the flow of the game and the event correlation It is possible to adjust the importance considering both. In this case, since the event corresponding to “calling and rolling” becomes a trigger event, the importance of the correlation event related to “calling and rolling” is also increased, and the effect of setting “calling and rolling” is compared with the first embodiment. And get bigger. This is not just the adjustment of the event importance of the “consolidation of events” event alone, but also the adjustment of the importance of events that have a correlation with the event. It is because it becomes possible to influence.

(Fourth embodiment)
In the first to third embodiments, the digest scenario information is set based on “the presence / absence of scores of both teams” and “the transition”. In the fourth embodiment of the present invention, the pattern of the game is distinguished into seven patterns based on “the presence / absence of the difference between the scores of both teams” and “the transition”, and the flow of the game is expressed in each pattern. It is characterized in that a necessary event is determined, and a condition necessary for extracting the event is set as an “event extraction condition”. In addition, an “event extraction order” is set so that important events are acquired in order from the determined events. It is also characterized in that the events to be selected for expressing the flow of the game are not limited to four points of “start”, “approval”, “turn”, and “conclusion”.

When the game patterns are differentiated based on “whether there is a difference in the scores between the two teams” and “the transition”, for example, if either team is overwhelmingly dominant, both teams will be playing a seesaw game. There are various patterns such as when both teams do not score. In addition, important events for expressing the flow of a game vary depending on the pattern of the game. For example, for a game in which either team is overwhelmingly superior, the “batter scoring scene” is an important event, and it is preferable to extract a plurality of these events in the digest scene. Also, in a game where both teams play a seesaw game, the “reverse scene” is an important event, and it is preferable that a plurality of these events be extracted in the digest scene. Furthermore, for a game in which neither team gets a score, a “pitcher's good pitching scene” is an important event, and it is preferable to extract a plurality of these events.

FIG. 24 shows the configuration of the event importance level adjusting apparatus 10 that implements this method. The apparatus 10 includes a receiving unit 11 that receives event information with importance added to event information indicating the content of the event, and situation information showing the status of the score and the runner, and the received importance attached An accumulation unit 12 that accumulates event information and situation information, a storage unit 13 that stores digest scenario information, a scenario importance adjustment unit 14 that adjusts the importance of event information with importance based on the digest scenario information, and A storage unit 15 in which an event correlation pattern table is stored, an event correlation importance level adjustment unit 16 that adjusts the importance level of event information with importance level based on the event correlation pattern table, and an importance level adjusted importance level And an output unit 17 that outputs event information with time.

The data structure of the event information with importance and the situation information input to the receiving unit 11 is substantially the same as that of the first embodiment (FIGS. 2 and 3). However, as shown in FIG. 25, the event information with importance in the present embodiment has an event identifier, and the situation information has a score information identifier and a runner situation identifier as shown in FIG. For example, the score information identifier of the situation information is given whenever the score changes, and the runner situation identifier is given every time the runner comes out (FIG. 27).

In addition, as shown in FIG. 28, the digest scenario information of the present embodiment includes pattern numbers (1 to 7) divided for each game content, scoring status (scoring of winning team / scoring of losing team), scoring difference, scoring The transition of the difference, the number of reverse rotations, the number of tie points, and the like are associated (scenario pattern determination table). The game content includes, for example, overwhelming victory, great victory, complete victory, hot win, close battle, sticky victory, score draw, and the like, and pattern numbers 1 to 7 correspond to these. An event to be selected for expressing the flow of the game for each pattern, an extraction order thereof, a search target section, and an event importance adjustment point are set (event adjustment extraction table). For example, from the scoring status of both teams, the score difference is “7 points or more (large)”, the transition of the score difference is “4 to 6 points or more by the middle,” the number of reversals is “0 after the middle,” and the number of ties is “ In the case of “0 times since the middle stage”, “Pattern 1 (Overwhelming victory)” is extracted, and the event is searched. In “Pattern 1”, the events to be searched are
"Scene 1": Winning team scoring event (final score)
"Scene 2": Winning team scoring event (additional points)
“Scene 3”: Strike event for losing team, strikeout event “Scene 4”: Scoring event for winning team (additional points to push)
“Scene 5”: Set as a losing team's hitting and strikeout event.

In order to extract events suitable for the above contents, “pattern 1” is searched for corresponding events in the order of “scene 1”, “scene 4”, “scene 2”, “scene 3”, and “scene 5”. First, the search is made from “events that have decided the game” or “events with high event importance” depending on whether there is a difference in scores or their transition. That is, the search is performed from points that are considered to be most important in the game, and this concept is derived and set from the extraction order shown in the description of the first to third embodiments. And, for example, as a means of determining the order of extracting events according to the game pattern (score difference and its transition, etc.) According to the game pattern, “Ki 1”, “Ki 2”, “Ki 1” “Ki 1” “Ki 2” “Ki 3” “Yi 1” are determined, and “Ki” and “Tai” candidates Is set to extract multiple.

For the event of “Scene 1”, the period from the start of the game (start) to the present is set as the search target section, the score event that becomes the final score in the winning team is searched, and +1000 is added to the importance of the corresponding event. For the event of “Scene 4”, from the time when the event of “Scene 1” occurs until the current time is set as a search target section, a scoring event that becomes an additional point of uselessness in the winning team is searched from behind, and the corresponding event is determined. Add +1000. For the event of “Scene 2”, from the time of occurrence of the event of “Scene 1” to the current time is set as the search target section, a score event that becomes an additional point in the winning team is searched from before, and +3000 is added to the corresponding event. Add. For the “Scene 3” event, the period from the start of the game (start) to the occurrence time of the “Scene 4” event is set as the search target section. Add +2000. Also, for the event of “Scene 5”, from the time of occurrence of the event of “Scene 4” to the current time is set as the search target section, the hit and strikeout events in the losing team are searched from behind, and +1000 is assigned to the corresponding event. Add.

Also, in the event correlation pattern table of the present embodiment, as shown in FIG. 32, “scene number”, “target team” corresponding to the scene of the event selected based on the event extraction adjustment table, and the above selected “Correlation target event” indicating the event name related to the event, “Correlation range” indicating the search range of the correlation target event, and “Correlation target event importance adjustment point” for adjusting the importance of the correlation target event are set. Has been.

For example, when pattern 1 is selected using the scenario pattern determination table of the digest scenario information and the event of scene 1 is selected using the event extraction table, the correlation event in the event correlation pattern table of pattern 1 corresponding to this is searched. Then, there are “hits, advance” and “advance / home-in”. When these are searched in the range shown in “Correlation Range” and “Hit, Progress” is found, +400 is added to the event importance, and “Progress / Home-In” is present +100 is added to the event importance.

The flowchart of FIG. 30 shows the processing procedure of the event importance level adjusting device 10. The receiving unit 11 receives the event information with importance and the situation information, and accumulates them in the accumulating unit 12 (step 11). The scenario importance level adjustment unit 14 acquires the accumulated event information with importance level and the situation information at a designated timing (step 12), and inputs them to the digest scenario information storage unit 13. The digest scenario information storage unit 13 selects a pattern number from the “scenario pattern determination table” based on these pieces of information, acquires an “event extraction adjustment table” of the selected pattern number, and uses this as the scenario importance adjustment unit 14. (Step 13). Then, the scenario importance level adjustment unit 14 selects an event (scenario event) to be adjusted based on the “event extraction adjustment table”, and adjusts the importance level of the selected scenario event (step 14).

Details of processing in the scenario importance adjustment unit 14 will be described with reference to FIG. In the scenario importance level adjustment unit 14, information for applying the “scenario pattern determination table” of the digest scenario information from the acquired event information with importance level and the situation information, for example, score status, score difference, transition of score difference, Extract the number of reverse rotations, the number of tie points, etc. From the situation information, information such as a score difference and its transition is extracted from the score situation, and information such as the number of inversions and tie points is extracted from the event information with importance (FIGS. 27 and 34). Based on these pieces of information, the digest scenario information storage unit 13 determines the pattern number, obtains an “event extraction adjustment table” for the pattern number, and inputs this to the scenario importance adjustment unit 14 (step 22). . Subsequently, corresponding events are extracted according to the extraction order, target team, target event, and search target section indicated in the “event extraction adjustment table” (step 23). If there is no corresponding event, the next event is searched. For example, in FIG. 29, an event of “scene 3” is searched, and when there is no losing team strikeout or strikeout event, the next “scene 5” event is searched. Next, the importance value of the extracted event is adjusted according to the event importance adjustment point indicated in the event extraction adjustment table (step 24), and the scene number indicated in the event extraction adjustment table is assigned to the event. (FIG. 35). By extracting the scenario event (step 23) and adjusting its importance (step 24), the process is repeated according to the extraction order shown in the “event extraction adjustment table” (step 25), thereby suppressing the points of the game, The flow of the game can be expressed.

The scenario importance adjustment unit 14 that has performed the above processing inputs the event information with importance (FIG. 35) adjusted using the digest scenario information and the pattern number to the event correlation importance adjustment unit 16 (step 15) The event correlation importance adjustment unit 16 inputs the pattern number to the event correlation pattern table storage unit 15. The event correlation pattern table storage unit 15 determines the pattern number, acquires the “event correlation pattern table” of the pattern number, and inputs it to the event correlation importance level adjustment unit 16 (step 16). 16 selects a correlation event related to the scenario event to be adjusted by the scenario importance level adjustment unit 14 based on the “event correlation pattern table”, and adjusts the importance level of the selected scenario event (step 17). ).

Details of processing in the event correlation importance level adjustment unit 16 will be described with reference to FIG. The event correlation importance adjustment unit 16 inputs the acquired pattern number to the event correlation pattern table storage unit 15 to acquire the “event correlation pattern table” of the pattern number (step 31). Subsequently, the corresponding correlation event is extracted according to the scene number, the target team, the target event, and the correlation range indicated in the “event correlation pattern table” (step 32). Next, the importance value of the extracted event is adjusted according to the event importance adjustment point shown in the “event correlation pattern table” (step 33), and the scene number shown in the event correlation pattern table is included in the event. (FIG. 36). By extracting the correlation event (step 32) and adjusting the importance thereof (step 33), the process is repeated according to the extraction order shown in the “event correlation pattern table” (step 34), so that an important event in the game, And importance level adjustment considering both correlation events. By using the event correlation pattern table, the importance level of the event related to the event adjusted by the scenario importance level adjustment unit 14 is also increased, so that the effect of setting the game flow is further increased. This is not just the adjustment of the event importance of a single scenario event, but also the adjustment of the importance of events that have a correlation with the event, so that multiple important events that differ depending on the game pattern are extracted. That is, an important scene in a game can be accurately extracted.

The event correlation importance adjustment unit 16 that has performed the above processing inputs the adjusted event information with importance (FIG. 36) to the output unit 17, and the output unit 17 outputs this (step 18).

As described above, the event importance level adjusting apparatus 10 first determines the game pattern (pattern 1) using the input event information with importance level (FIG. 34) and the situation information, and determines the pattern of the game pattern. Use the event extraction adjustment table to select important events in the game. Then, the selected event is identified as a part of the important scene to be extracted, and then a correlation event related to the identified event is selected using the event correlation pattern table of the game pattern. In this way, according to the game pattern set based on the “presence / absence of scores between both teams” and “its transition”, the importance level can be adjusted by taking into account both the flow of the game and the event correlation. As shown in FIG. 37, important scenes can be extracted accurately.

The event importance level adjusting device of the present invention can be used in combination with a device that automatically creates a digest from video content. For example, an editing device that edits a digest video of a sports game at a broadcasting station, or a received video of sports It can be incorporated into a household device or the like for creating a digest, or used as a front device for those devices.

In addition, the event importance level adjusting device of the present invention can be applied to video content of other various genres without limiting the video content to sports such as baseball, soccer, tennis, golf, swimming, marathon and the like. For example, it can be applied to a movie, a drama, or news. In other words, a pattern is determined for a certain genre based on predetermined information (“score and its transition” or “score difference and its transition” when the genre is sport), and the flow of video content is expressed for each pattern. By managing a table that shows the digest scenario information that sets the extraction conditions necessary to extract the events and the order in which the events are extracted, by determining the necessary events for the event, Can be extracted. Note that by setting and managing the correlation pattern information of each event, an event having a correlation with an important event in the video content can be extracted.

  In each of the above-described embodiments, the event importance level adjusting device is input with the event information with importance level and the situation information. However, the importance level calculation unit that calculates the importance level from the event information is adjusted with the event importance level adjustment. The event importance level adjusting device may be provided in the device, supplying event information and status information to the event importance level adjusting device, and generating event information with importance level from the event information may be performed inside the event importance level adjusting device. .

The block diagram which shows the structure of the event importance adjustment apparatus in the 1st Embodiment of this invention. Data configuration diagram of event information with importance input to the event importance adjustment device according to the first embodiment of the present invention The data block diagram of the status information input into the event importance adjustment apparatus in the 1st Embodiment of this invention The figure which shows the data structure of the digest scenario information in the 1st Embodiment of this invention. The flowchart which shows the process sequence of the event importance adjustment apparatus in the 1st Embodiment of this invention. The flowchart which shows the procedure of the event extraction adjustment process in the 1st Embodiment of this invention. The figure which shows the position or the range of the incoming and outgoing roll set by the event extraction adjustment process in the 1st Embodiment of this invention (in the case of scenario pattern 1) The figure which shows the position or the range of the incoming and outgoing roll set by the event extraction adjustment process in the 1st Embodiment of this invention (in the case of scenario pattern 2) The figure which shows the position or range of the starting and rolling connection set by the event extraction adjustment process in the 1st Embodiment of this invention (in the case of scenario pattern 3) The figure which shows the position or range of the incoming and outgoing roll set by the event extraction adjustment process in the 1st Embodiment of this invention (in the case of scenario pattern 4) The figure which shows the position or range of the incoming and outgoing roll set by the event extraction adjustment process in the 1st Embodiment of this invention (in the case of scenario pattern 5) The block diagram which shows the structure of the event importance adjustment apparatus in the 2nd Embodiment of this invention. The figure which shows the data structure of the event correlation pattern information in the 2nd Embodiment of this invention. The flowchart which shows the process sequence of the event importance adjustment apparatus in the 2nd Embodiment of this invention. The flowchart which shows the procedure of the correlation event extraction adjustment process in the 2nd Embodiment of this invention. The figure explaining the specific example of the correlation event extraction adjustment process in the 2nd Embodiment of this invention (The figure which shows the event before adjustment) The figure explaining the specific example of the correlation event extraction adjustment process in the 2nd Embodiment of this invention (The figure which shows the event importance after adjustment) The block diagram which shows the structure of the event importance adjustment apparatus in the 3rd Embodiment of this invention. The flowchart which shows the process sequence of the event importance adjustment apparatus in the 3rd Embodiment of this invention. The figure explaining the specific example of the event importance adjustment process in the 3rd Embodiment of this invention (The figure which shows the event importance before adjustment) The figure explaining the specific example of the event importance adjustment process in the 3rd Embodiment of this invention (The figure which shows the relationship between the display of a scoreboard, and an event) The figure explaining the specific example of the event importance degree adjustment process in the 3rd Embodiment of this invention (The figure which shows the relationship between the display of a score board and the event of a start-and-following conclusion) The figure explaining the specific example of the event importance adjustment process in the 3rd Embodiment of this invention (The figure which shows the event importance after adjustment) The block diagram which shows the structure of the event importance adjustment apparatus in the 4th Embodiment of this invention. Data configuration diagram of event information with importance input to the event importance adjustment device according to the fourth embodiment of the present invention The data block diagram of the situation information input into the event importance adjustment apparatus in the 4th Embodiment of this invention The figure which shows the data structure which provided the identifier about the status information in the 4th Embodiment of this invention. The figure which shows the data structure of the scenario pattern determination table of digest scenario information in the 4th Embodiment of this invention. The figure which shows the data structure of the event extraction adjustment table of digest scenario information in the 4th Embodiment of this invention. The flowchart which shows the process sequence of the event importance adjustment apparatus in the 4th Embodiment of this invention. The flowchart which shows the procedure of the event extraction adjustment process in the 4th Embodiment of this invention. The figure which shows the data structure of the event correlation pattern information in the 4th Embodiment of this invention. The flowchart which shows the procedure of the correlation event extraction adjustment process in the 4th Embodiment of this invention. Data configuration diagram of event information (input data) with importance input to the scenario importance adjustment unit in the fourth embodiment of the present invention Data structure diagram of event information (intermediate data) with importance output from the scenario importance adjustment unit in the fourth embodiment of the present invention Data configuration diagram of event information (output data) with importance output from the event correlation importance adjustment unit according to the fourth embodiment of the present invention The figure which shows the various scenes extracted by the event extraction adjustment process and correlation event extraction adjustment process in the 4th Embodiment of this invention

Explanation of symbols

DESCRIPTION OF SYMBOLS 10 Event importance adjustment apparatus 11 Receiving part 12 Event information with importance and status information storage part 13 Digest scenario information storage part 14 Scenario importance degree adjustment part 15 Event correlation pattern information storage part 16 Event correlation importance degree adjustment part 17 Output part

Claims (3)

An event indicating an individual event in video content, a trigger for starting processing, a trigger event that is a basis for adjusting the importance of other events, and a correlation event in which the importance is adjusted based on the trigger event When,
An importance adjustment point determined by a combination of the trigger event and the correlation event;
An event correlation pattern information storage unit that accumulates an event correlation pattern table indicated by;
Receiving event information with importance including the event and importance assigned to the event, and status information indicating the status of the video content;
Based on the event correlation pattern table, it is determined whether an event included in the event information with importance is the trigger event, and when the event is the trigger event,
Based on the event correlation pattern table, extract the relevant correlation event from the event information with importance,
An event correlation importance level adjustment unit that adjusts the importance level of the extracted correlation event using the importance level adjustment point;
An event importance adjustment device comprising: The event of the event information with importance using the event information with importance indicating the importance assigned to the event indicating each event in the video content and the situation information indicating the status of the video content An event importance adjustment method for adjusting the importance of
A trigger event that triggers the start of processing and serves as a basis for adjusting the importance of other events; a correlation event that adjusts the importance based on the trigger event;
An importance adjustment point determined by a combination of the trigger event and the correlation event;
Using the event correlation pattern table with
It is determined whether the event included in the situation information is the trigger event, and when the event is the trigger event,
Based on the event correlation pattern table, extract the relevant correlation event from the event information with importance,
An event importance adjustment method for adjusting the importance of the extracted correlation event using the importance adjustment point. Using the event information with importance including the event indicating the individual phenomenon in the video content and the importance assigned to the event, and the status information indicating the status of the video content, the event information with importance The computer built into the device that adjusts the importance of
A trigger event that triggers the start of processing and serves as a basis for adjusting the importance of other events; a correlation event that adjusts the importance based on the trigger event;
Using the event correlation pattern table showing the importance adjustment point determined by the combination of the trigger event and the correlation event,
It is determined whether an event included in the event information with importance is the trigger event, and when the event is the trigger event,
Based on the event correlation pattern table, extract the relevant correlation event from the event information with importance,
A program for executing a procedure for adjusting the importance level of the extracted correlation event using the importance level adjustment point.

Images