JP2014006844A - Video recording apparatus, information processing method, and recording medium - Google Patents

Abstract

PROBLEM TO BE SOLVED: To allow a successful search for a content to be recommended to a user, from recorded video contents.SOLUTION: A characteristic keyword is extracted with respect to each of plurality of scenes into which a stored video content is divided, and scores are calculated in accordance with a frequency of the extracted keyword in the scene. The calculated scores of keywords of respective video contents are accumulated per keyword to obtain cumulative values. A scene meeting a prescribed condition out of scenes from which keywords of which the obtained cumulative values of scores are equal to or larger than a threshold have been extracted or a video content including the scene is selected as a scene or a video content to be recommended.

Description

  The present disclosure relates to a video recording apparatus that stores video content such as a broadcast program, an information processing method applied to the video recording apparatus, and a recording medium on which a program for executing the information processing method is recorded.

Video recording devices that store video content such as broadcast programs have been developed and put into practical use. For example, various video recording apparatuses using a hard disk as a recording medium have been commercialized. This type of video recording apparatus tends to increase the capacity of programs that can be stored with the recent increase in capacity of hard disks. For example, a video recording apparatus having a hard disk with a recording capacity of 2 Tbytes can record a program exceeding 1000 hours.
As a video recording device, a device having a plurality of tuners and capable of simultaneously recording broadcast programs of a plurality of channels has been developed.

Some video recording devices continuously record all programs broadcast on a plurality of channels by utilizing such a large recording medium capacity and a simultaneous recording function for broadcasting programs on a plurality of channels. is there. For example, as a video recording apparatus, there is one in which a program broadcast on a maximum of 8 channels is continuously recorded, and at least a recording program of about one week is left on a recording medium. In the following description, such a video recording apparatus is referred to as a multi-channel constant recording type video recording apparatus.
The multi-channel continuous recording type video recording device does not require a user to perform a recording reservation operation in advance, and has the effect of reliably watching a missed program, while the number of recorded programs is enormous, There is a problem that it takes time to search for a recorded program that the user wants to watch.

  In Patent Document 1, when a video recording apparatus reproduces a broadcast program recorded on a recording medium, the keyword is detected from subtitles of the reproduced program, the detected keyword is accumulated, and the user's preference is recorded. There is a description about obtaining a reflected user evaluation value. In the video recording apparatus described in Patent Document 1, a recorded program having a keyword with a large user evaluation value is recommended as a program that matches the user's preference.

JP 2007-281856 A

  The method described in Patent Document 1 obtains a user evaluation value of a keyword from a program reproduced by the user himself and recommends a recorded program based on the user evaluation value. That is, the method described in Patent Document 1 searches for a program having a content similar to a program reproduced by the user in the past, a program of the same genre as a program reproduced in the past, or the like.

  However, in the case of the multi-channel always-recording type video recording apparatus as described above, since the recorded program is all the programs to be broadcast, the recommended recording program using a specific evaluation value based on the user viewing history is recommended. However, it is not always possible to recommend an appropriate program. That is, in the case of a multi-channel continuous recording type video recording apparatus, it is necessary to recommend a program to the user from all the programs of various genres that have been broadcast, and development of a more advanced method is desired. It was.

  Here, the multi-channel always-recording type video recording apparatus has been described as an example. However, even when the video recording apparatus recommends video content such as a recorded program even if the type is not always-recording type, It is hoped that more advanced recommendations can be made.

  An object of the present disclosure is to provide a video recording apparatus, an information processing method, and a recording medium that can favorably find video content recommended to a user from recorded video content such as a program.

The video recording apparatus according to the present disclosure includes a receiving unit that receives video content, and a content storage unit that stores video content obtained by the receiving unit. Then, a feature extraction unit, a cumulative calculation unit, and a selection unit are provided as processing units for recommending the video content stored by the content storage unit.
The feature extraction unit extracts a keyword that is a feature for each scene obtained by dividing each video content accumulated by the content accumulation unit into a plurality of scenes, and calculates a score according to the frequency of the extracted keyword in the scene.
The accumulating unit obtains an accumulated value by accumulating the keyword score of each video content calculated by the feature extracting unit for each keyword.
The selection unit recommends a scene or video that recommends a scene that matches a predetermined condition or a video content including the scene among scenes from which a keyword having a cumulative score obtained by the cumulative calculation unit is extracted with a threshold value or more. Select as content.

The information processing method of the present disclosure performs a score calculation process, a cumulative calculation process, and a selection process.
In the score calculation process, a keyword that is characteristic for each scene obtained by dividing the accumulated video content into a plurality of scenes is extracted, and a score is calculated according to the frequency of the extracted keyword in the scene.
In the cumulative calculation process, the score of each video content keyword calculated in the score calculation process is cumulatively added for each keyword to obtain a cumulative value.
The selection process is a scene or video that recommends a scene that matches a predetermined condition or video content including the scene among scenes from which a keyword having a cumulative score obtained by the cumulative calculation process is extracted with a threshold value or more. Process to select as content.

The recording medium of this indication is a recording medium which recorded the program for making a computer perform, and makes a computer perform the following procedures (a)-(c).
(A) A score calculation procedure for extracting a keyword that is characteristic for each scene obtained by dividing the accumulated video content into a plurality of scenes and calculating a score according to the frequency of the extracted keyword in the scene.
(B) A cumulative calculation procedure in which the score of each video content keyword calculated in the score calculation procedure is cumulatively added for each keyword to obtain a cumulative value.
(C) A scene or video content that recommends a scene that matches a predetermined condition or a video content including the scene among scenes from which keywords having a cumulative score obtained by the cumulative calculation procedure are extracted from a threshold value or more are extracted. As a selection procedure.

  According to the present disclosure, since a cumulative value of scores corresponding to the frequency of keywords detected for each scene of video content accumulated by the content accumulation unit is obtained, topics commonly handled by a plurality of video programs can be obtained. , Detected as a cumulative value of high score. For this reason, an apparatus including a content storage unit can take out a scene that deals with the current topic from the stored video content, and the scene of the current topic (or the scene) Video content) can be recommended to the user.

  According to the present disclosure, it is possible to take out a scene (or video content including the scene) of the current topic from the accumulated video content and recommend it to the user. Therefore, even when a large number of video contents are stored, such as a multi-channel continuous recording type video recording apparatus, it is possible to recommend appropriate video contents to the user.

It is a block diagram which shows the structural example of the video recording apparatus by one embodiment of this invention. It is a flowchart which shows the process example by one embodiment of this invention. It is explanatory drawing which shows the example of the scene division | segmentation process by one embodiment of this invention. It is explanatory drawing which shows the example of the score calculation process of the keyword by one embodiment of this invention. It is explanatory drawing which shows the example of the process which determines a recommendation scene from the cumulative value of each scene by one embodiment of this invention. It is a figure which shows the example of the display screen of the recommendation scene by one embodiment of this invention. It is a flowchart which shows the process example (modification 1: the example which determines a recommendation program) by one embodiment of this invention. It is explanatory drawing which shows the example of the process which determines a recommendation program from the cumulative value of each scene by the modification 1 of one embodiment of this invention. It is a figure which shows the example of the display screen of the recommended program by the modification 1 of one embodiment of this invention. It is a figure which shows the example of the display screen of the recommendation scene and the recommendation program by the modification 2 of one embodiment of this invention. It is a flowchart which shows the process example (Modification 3: Example which calculates a score by weighting) by one embodiment of this invention. It is a flowchart which shows the process example (Modification 4: Example which performs digest reproduction | regeneration of a recommendation scene) by one embodiment of this invention. It is explanatory drawing which shows the digest reproduction | regeneration state by the modification 4 of one embodiment of this invention. It is a flowchart which shows the process example (modification 5: the example which transfers a recommendation scene to another terminal device) by one embodiment of this invention. It is explanatory drawing which shows the transfer state by the modification 5 of one embodiment of this invention.

Examples of a video recording apparatus, an information processing method, and a recording medium according to an embodiment of the present disclosure will be described in the following order with reference to the drawings.
1. Configuration example of video recording device (FIG. 1)
2. Example of processing flow (FIGS. 2 to 6)
3. Modification 1: Example of determining recommended programs (FIGS. 7 to 9)
4). Variation 2: Example of Displaying Recommended Scenes and Recommended Programs (FIG. 10)
5. Modification 3: Example of calculating a score by weighting (FIG. 11)
6). Modification 4: Example of digest playback of a recommended scene (FIGS. 12 and 13)
7). Variation 5: Example of transferring a recommended scene to another terminal device (FIGS. 14 and 15)
8). Other variations

[1. Configuration example of video recording apparatus]
FIG. 1 is a diagram illustrating a configuration example of a video recording apparatus 100 according to an embodiment of the present disclosure.
The video recording apparatus 100 includes a large-capacity recording medium and accumulates a large number of video contents (broadcast programs, movies, moving images acquired from moving image sites, etc.). A display device 300 is connected to the video recording device 100. When the video recording device 100 plays back the stored video content, the display device 300 displays the video of the played video content, and a sound (not shown) provided in the display device 300 outputs the sound.
Note that the video recording device 100 and the display device 300 may be integrated.

  The video recording apparatus 100 includes a tuner 102 to which an antenna 101 is connected. The tuner 102 simultaneously receives television broadcast signals of a plurality of channels. For example, the tuner 102 can simultaneously receive eight channels. Broadcast waves that can be received simultaneously may be either terrestrial or satellite broadcast waves.

  The television broadcast signal received by the tuner 102 is supplied to the recording processing unit 104, and the signal subjected to the recording process by the recording processing unit 104 is recorded on the recording medium 120 as video content. A hard disk drive, a flash memory, or the like is used for the recording medium 120 that is a storage unit for video content. As the recording medium 120, a recording medium having a very large capacity (for example, about several T bytes) is used.

In the recording medium 120, video content obtained by the tuner 102 receiving a plurality of channels is constantly recorded. For example, the tuner 102 of the video recording apparatus 100 always receives a plurality of channels specified in advance by a user operation or the like. Then, video programs of a plurality of channels received by the tuner 102 are always recorded on the recording medium 120.
The recording medium 120 stores the recorded video program for a predetermined period such as one week.

  When video content is recorded on the recording medium 120, the system controller 110 determines the program title, genre, outline of the content, etc. from the electronic program information added to the broadcast wave of each channel. Then, under the control of the system controller 110, information such as the title is recorded on the recording medium 120 as additional information of the video content. When text data such as subtitles is added to the video content, the text data is also recorded on the recording medium 120.

The video program recorded on the recording medium 120 is recorded in a state of being divided into a plurality of scenes according to the content of the video program. When a video program is divided into a plurality of scenes, for example, the timing at which a large change in video or audio is detected is used as a scene break. Alternatively, the timing at which a commercial is inserted into a program may be a scene break.
Further, the recording medium 120 includes an area for recording data of the analysis result of the video content recorded by the recording medium 120 itself. The process of analyzing the video content is executed by the image / audio feature extraction unit 111 under the control of the system controller 110.

  In addition, the video recording apparatus 100 includes a disk drive unit 130, and a disk-shaped recording medium 131 such as an optical disk mounted on the disk drive unit 130 can record video content.

  The video recording apparatus 100 includes a network interface unit 103 and can be connected to the Internet 90 via a predetermined network. Video content received from the outside by the network interface unit 103 is recorded on the recording medium 120 or the disk-shaped recording medium 131.

The reproduction processing unit 108 reproduces the video content recorded on the recording medium 120 to obtain image data and audio data. Image data and audio data obtained by this reproduction processing are supplied to the external display device 300 via the display processing unit 109. The display panel of the display device 300 displays an image based on the supplied image data. Audio based on the supplied audio data is output from a speaker included in the display device 300.
When the information processing unit 105 performs a content list creation process, the display processing unit 109 creates image data for displaying the content list and supplies the created image data to the display device 300. To do. A list indicating recommended titles and recommended scenes, which will be described later, is also created by the information processing unit 105, and image data for displaying the list is created by the display processing unit 109.

  In addition, the video recording apparatus 100 can perform data transfer with the terminal apparatus 200 through the terminal interface unit 106. The terminal interface unit 106 can transfer data by connecting to the terminal device 200 through communication using a wired cable or wireless communication. The terminal device 200 connected to the video recording device 100 is a mobile phone terminal device having a video content playback function called a smartphone. Alternatively, the terminal device 200 may be a terminal device having a video content reproduction function, which includes a relatively large display panel called a tablet terminal device.

The image / sound feature extraction unit 111 extracts features from the video content image or sound recorded in the recording medium 120 and acquires keywords from the extracted features. For example, the image / sound feature extraction unit 111 converts a character in an image such as a caption or a telop added to the image into text data, and detects a keyword. In speech feature extraction processing, speech data is converted into text data by speech recognition processing, and keywords are detected.
When there is information (program guide information, subtitles, etc.) added to the video content, the image / audio feature extraction unit 111 obtains text data from the added information and detects a keyword. Also good.
When the image / sound feature extraction unit 111 detects a keyword from text data, for example, dictionary data prepared in advance in the video recording apparatus 100 is referred to. This dictionary data may be sequentially updated to the latest data by communication with an external server.

  The keyword data extracted by the image / audio feature extraction unit 111 is supplied to the information processing unit 105. The information processing unit 105 calculates a keyword score for each scene of each video content. The keyword score for each scene of each video content calculated by the information processing unit 105 is recorded in the recording medium 120 as data of the video content analysis result. A specific processing example in which the information processing unit 105 calculates a keyword score for each scene will be described later.

  Furthermore, the information processing unit 105 selects a scene or video content recommended for the user from the video content recorded on the recording medium 120 based on the calculated score of each keyword. Then, under the control of the system controller 110, the information processing unit 105 performs a process of creating a list of selected scenes or video contents.

  Each processing unit in the video recording apparatus 100 exchanges control commands and the like with the system controller 110 via the control bus 141 and transfers data via the data bus 142. A memory 107 is connected to the control bus 141 and the data bus 142, and data is temporarily stored in the memory 107 when the system controller 110 or the like performs processing. The memory 107 stores a program executed by the system controller 110. Then, when the video recording apparatus 100 is activated, the system controller 110 executes the program stored in the memory 107, so that the video recording apparatus 100 operates as instructed by the program. The process of recommending a scene or video content described in the example of the present embodiment is also performed by the system controller 110 executing a program stored in the memory 107.

[2. Example of processing flow]
Next, processing for selecting a recommended scene or video content from the video content accumulated by the video recording apparatus of the present disclosure will be described with reference to FIGS. Here, processing for recommending a broadcast program recorded on the recording medium 120 or a part of scenes in the program is performed. In the following description, video content is referred to as a program.

FIG. 2 is a flowchart illustrating an example of a process for selecting a scene or program to be recommended.
First, the information processing unit 105 extracts a target program broadcast within a designated period from programs stored in the recording medium 120 (step S11). For example, the information processing unit 105 extracts a target program broadcast within one day (24 hours). The target program is, for example, a program of a specific genre. Alternatively, the target program may be all programs broadcast within the corresponding time.

If the extracted target program is not divided into scenes, the information processing unit 105 divides the extracted program for each scene (step S12). If the extracted target program has already been divided into scenes at the time of recording on the recording medium 120, this step S12 is omitted.
Then, the image / sound feature extraction unit 111 converts the image and sound in each scene of the extracted target program into text (step S13). For example, the text conversion is performed by recognizing characters in the image such as subtitles and telops added to the image. The speech is converted into text by performing speech recognition processing on the speech data. If the program data has additional information (caption information, program guide information, etc.), text data is obtained from the additional information. In addition, when performing this text, it is preferable to remove the commercial part from the text object.

  When the text data of each scene is acquired, the information processing unit 105 extracts a keyword that is characteristic for each scene from the text (step S14). At the time of taking out this keyword, for example, the TFIDF method is used. The TFIDF method is a method of extracting a keyword that is characteristic for each scene using two indices, an index called TF (term frequency) and an index called IDF (inverse document frequency). Specifically, in the TFIDF method, the frequency with which a certain keyword appears in a scene and the appearance frequency with respect to all of a plurality of scenes for which the keyword is searched are calculated. In the TFIDF method, a keyword that has a high frequency of appearance in a specific scene and a low frequency of appearance in another scene is determined as a high score keyword. Note that the use of the TFIDF method is one example, and a keyword may be extracted from each scene using another statistical method to calculate a score.

In this way, the information processing unit 105 extracts a keyword that is characteristic in each scene, and obtains a score for the keyword. At this time, for example, a score having a certain threshold value or more is taken out. Alternatively, for each scene, a predetermined number is extracted from those having a high score.
Then, the information processing unit 105 obtains the average value of the same keyword for each broadcasting station (channel) on which the program of each scene is broadcast from the keyword score thus extracted (step S15). When the average value of each keyword for each broadcasting station is obtained, the information processing unit 105 calculates an average value of each broadcasting station and a cumulative value obtained by adding the same keywords together (step S16).

  When the cumulative value of the average value of the keywords is calculated, the information processing unit 105 determines whether there is a cumulative value that is equal to or greater than a predetermined threshold value and one broadcast station (step S17). . In this determination, when it is determined that there is a keyword detected from only one broadcasting station, the keyword is a keyword that is a topic only in a specific broadcasting station, and is thus deleted from the candidate keywords ( Step S18).

Then, after the keyword with one broadcast station is not detected in step S17 and after the keyword with only one broadcast station is deleted in step S18, the information processing unit 105 has a keyword having a cumulative value equal to or greater than the threshold. A list of scenes in which is detected is listed (step S19).
After acquiring a list of scenes in which keywords having a cumulative value equal to or greater than the threshold value are detected in this way, the information processing unit 105 determines whether or not the same scene is extracted from a plurality of keywords in the acquired list. Judgment is made (step S20). If it is determined that there is a keyword in the same scene, the keyword with the lower score accumulation value is deleted from the candidate list (step S21).

Further, the information processing unit 105 determines whether or not there is a plurality of scenes for one keyword in the list of scenes extracted for each keyword (step S22). In this determination, when there are a plurality of scenes, the information processing unit 105 selects the longest scene and excludes other scenes (second and subsequent scenes) from the list. (Step S23). The information processing unit 105 determines the scene remaining in the list as a recommended scene in this way (step S24).
Then, the information processing unit 105 creates a list notification screen for notifying the user of the confirmed recommended scene.

Next, a specific example of the processing described with reference to the flowchart of FIG. 2 will be described with reference to FIGS.
FIG. 3 is a diagram illustrating an example of processing in which the information processing unit 105 extracts a target program from programs stored in the recording medium 120. In this example, the information processing unit 105 extracts a program whose genre is news from programs broadcast within one day. For example, as shown in FIG. 3A, from a daily program guide, program A broadcasted by broadcast station X, programs C and E broadcasted by broadcast station B, and programs B and D broadcasted by broadcast station Z are news. The information processing unit 105 takes out each program. Then, as illustrated in FIG. 3B, the information processing unit 105 divides each extracted program for each scene. This scene division may be performed when the program is recorded on the recording medium 120.

FIG. 4 is a diagram illustrating an example in which a keyword score is calculated based on text extracted from each scene of a program.
When the data of the scenes 1, 2, 3,... Obtained by dividing the program A as shown in FIG. 4A are acquired, the audio / video feature extraction unit 111 converts each scene into text as shown in FIG. 4B. FIG. 4B shows an example in which the audio of program A is converted into text.
Then, as illustrated in FIG. 4C, the information processing unit 105 calculates a score by the TFIDF method from the text of each scene. For example, assume that a large number of keywords “major league” are detected in scene 2, a keyword “large snow” is detected in scene 3, and a keyword “consumption tax” is detected in other scenes. For each keyword, as shown in FIG. 4C, a score is calculated based on the frequency in each scene.

  The example shown in FIG. 4C is a keyword score of program A, but the score of each program keyword is averaged for each broadcasting station (channel). For example, the scores of “major league”, “snowfall”, and “consumption tax” described above are averaged with the scores of the same keyword detected from another news program of the same broadcasting station X.

FIG. 5 is a diagram illustrating an example of determining a recommended scene from an average score of keywords for each broadcasting station.
First, as shown in FIG. 5A, it is assumed that an average score for each keyword is obtained individually at each broadcasting station X, Y,. At this time, the information processing unit 105 obtains a list of scores of cumulative addition values illustrated in FIG. 5B in which the average score of each broadcasting station is added for each keyword. This list of cumulative addition value scores has cumulative addition values of scores for each keyword and the number of broadcast stations where the corresponding keyword is detected from the program. The list of cumulative addition value scores shown in FIG. 5B is created for a cumulative addition value that is equal to or greater than a specific threshold value, for example.

In the list of cumulative addition value scores shown in FIG. 5B, the number of broadcast stations in which each keyword appears is shown. For example, the keywords “major league”, “large snow”, “minimum temperature”, and “high school baseball” were detected from two broadcast station programs, respectively, and the keyword “consumption tax” was detected from only one broadcast station program. Is shown.
At this time, the information processing unit 105 deletes the keyword detected from only the program of one broadcasting station from the score list of the cumulative addition value. For example, in the example of FIG. 5B, the keyword “consumption tax” is deleted.
By deleting such a keyword detected from only one broadcasting station, a topic handled by only one broadcasting station is deleted from the keyword.

Next, as illustrated in FIG. 5C, the information processing unit 105 lists the scenes in which each keyword described in the list of accumulated added value scores is detected. The broadcast time of each scene is added to this list.
For example, it is indicated that the keyword “major league” corresponds to scene 2 of program A and scene 3 of program B. Further, it is shown that the keyword “Daisetsu” corresponds to scene 4 of program B and scene 2 of program E. Further, it is shown that the keyword “minimum temperature” corresponds to scene 4 of program B and scene 3 of program D.

In the state shown in FIG. 5C, since the keyword “large snow” and the keyword “minimum temperature” correspond to a common scene (scene 4 of program B), the keyword with the lower keyword cumulative score is selected from the list. Deleted. That is, in the cumulative score shown in FIG. 5B, the keyword “heavy snow” has a higher value than the keyword “minimum temperature”, so the keyword “minimum temperature” is selected for deletion.
Therefore, as shown in FIG. 5D, the keyword “minimum temperature” is deleted from the list. In this example, all scenes corresponding to the keyword “minimum temperature” are deleted. That is, not only the scene 4 of the program B matching the keyword “Daisetsu” but also the scene 3 of the program D is deleted from the list.
By performing such deletion, it is prevented that keywords having similar meanings are included in the candidate list.

  Furthermore, as illustrated in FIG. 5E, the information processing unit 105 selects a scene having the longest broadcast time for each keyword from the candidate list. For example, for the keyword “major league”, scene 2 of program A and scene 4 of program B are listed, but scene 2 of program A with a long broadcast time is selected and left in the list. Also for the keyword “Daisetsu”, scene 4 of program B with a long broadcast time is selected and left in the list.

  In this way, the scene remaining in the list becomes the recommended scene. For example, in the example of FIG. 5E, scene 2 of program A is selected for the keyword “major league”, scene 4 of program B is selected for the keyword “large snow”, and scene 3 of program E is selected for the keyword “high school baseball”. Suppose you are chosen.

FIG. 6 shows an example in which the display device 300 displays a screen 310 for notifying the user of the recommended scene determined in this way.
In the example of FIG. 6, the window 311 indicating the scene 2 of the program A, the window 312 indicating the scene 4 of the program B, and the window indicating the scene 3 of the program E, which are the three scenes determined in the list of FIG. 313 is displayed. In each of the windows 311, 312, and 313, the program title of each scene, the broadcast date and time, the length of the scene, and the outline of the program contents are described. As the outline of the program content, for example, the head part of the text of each scene is used.
In addition, thumbnail images 311a, 312b, and 313c as representative images of the respective scenes are displayed in the respective windows 311, 312, and 313.

  When the system controller 110 detects that a user operation for selecting any of the windows 311, 312, 313 has been performed in a state where the screen 310 shown in FIG. 6 is displayed, the selected scene is reproduced. For example, when the window 311 is selected by a user operation, the system controller 110 gives an instruction to reproduce the scene 2 of the program A, and the reproduction processing unit 108 performs reproduction processing.

  Based on the text in the recorded program, a scene to be recommended is determined, and a list showing the recommended scene is displayed. You will be able to recommend automatically. In other words, the frequency of each keyword scene is averaged for each broadcasting station (channel), and the average value is cumulatively added for each channel, so that the cumulative score addition value is broadcasted by a plurality of broadcasting stations. The higher the topic, the higher. Therefore, a topic that is a topic from a program broadcast within a specific period such as one day can be obtained by recommending a scene with a higher cumulative addition value for a topic handled in many programs. Can be recommended automatically.

  In this case, even if the cumulative addition value of the score is high, if it is a keyword detected only from a program of one broadcasting station, it is excluded from the recommended scene so that only a specific broadcasting station can handle it. Topics are excluded. Moreover, when there are different keywords detected from the same scene, it is possible to prevent a plurality of similar keywords from becoming a plurality of candidates by leaving only one keyword. By performing these processes, it is possible to select a recommended scene with high accuracy.

  In particular, as in the example of the embodiment of the present disclosure, when it is applied to the multi-channel constant recording type video recording apparatus 100, all programs broadcast on the respective channels are recorded, which is a hot topic. It will be possible to recommend appropriate matters based on keywords.

[3. Modification 1: Example of determining recommended program]
In the example of the embodiment described above, the recommended scene is determined, but a recommended program (video program) may be determined from the candidate scenes.
The flowchart of FIG. 7 is a diagram illustrating an example of processing when determining the recommended program. In the flowchart of FIG. 7, the same step numbers are assigned to the same processes as those in the flowchart of FIG.

The processing in the flowchart of FIG. 7 will be described. In step S20, the information processing unit 105 determines whether or not the same scene is extracted with a plurality of keywords, and deletes the keywords extracted from the same scene in step S21. The process is the same as that shown in FIG.
Then, when a plurality of keywords are not extracted from the same scene in step S20 and after deleting the keywords in step S21, the information processing unit 105 counts the number of recommended scenes for each program (step S25). At this time, the information processing unit 105 targets, for example, the top N-th scene (N is an arbitrary integer) in the list. Alternatively, the information processing unit 105 may target all scenes included in the list.

Then, after summing up the number of recommended scenes for each program, a program having a higher recommended scene number is selected as a recommended program (step S26).
Then, the information processing unit 105 creates a list notification screen for notifying the user of the confirmed recommended program.

FIG. 8 is a diagram showing an example of determining a recommended program by calculating a keyword score based on text extracted from each scene of a program. 8A to 8D are the same as FIGS. 5A to 5D and will not be described.
Then, the information processing unit 105 deletes one of the two keywords acquired from the same scene as shown in FIG. 8D, and then totals the number of candidate scenes for each program as shown in FIG. 8E.
That is, for example, in the example of FIG. 8D, the information processing unit 105 determines that the program B and the program E have two scenes in the list, and the two scenes are candidate programs B and E as the recommended programs. .

FIG. 9 shows an example in which the display device 300 displays a screen 320 for notifying the user of the recommended scene determined in this way.
In the example of FIG. 9, the screen 320 includes a window 321 indicating the title and broadcast date and time of program B, which are two programs determined in the list of FIG. 8E, and a window 322 indicating the title and broadcast date and time of program E. Is displayed. In addition, as an order when displaying a plurality of programs on the screen 320, for example, the one with the higher average score of keywords is set upward. Alternatively, the order may be determined by other processing, such as making the program having the longer length upward.

  When the system controller 110 detects that a user operation for selecting any one of the windows 321 and 322 is performed with the screen 320 shown in FIG. 9 being displayed, the reproduction processing unit 108 displays the selected program. Reproduce. For example, when the window 321 is selected by a user operation, the system controller 110 instructs the reproduction processing unit 108 to reproduce the program B.

  By displaying what is recommended in units of programs (units of video programs) in this way, the video recording apparatus 100 recommends programs that deal with a lot of topical information and makes recommendations in units of programs favorably. Can do.

  In the example of FIG. 9, a program in which a plurality of scenes are candidates is set as a recommended program in the list. On the other hand, all programs included in the list may be displayed in order by ranking with some processing.

[4. Modification 2: Example of Displaying Recommended Scene and Recommended Program]
The list display of recommended scenes shown in FIG. 6 and the list display of recommended programs shown in FIG. 9 may be performed simultaneously.
That is, as shown in FIG. 10, on the screen 330, the window 311 showing the scene 2 of the program A, the window 312 showing the scene 4 of the program B, which are the three scenes determined in the list of FIG. A window 313 showing scene 3 of E is displayed. Further, on the screen 330, a window 321 indicating the title and broadcast date and time of the program B, which are two programs determined in the list of FIG. 8E, and a window 322 indicating the title and broadcast date and time of the program E are simultaneously displayed.

  By displaying such a screen 330, when the user selects any of the windows 311 to 313, a scene corresponding to each window is reproduced. When the user selects one of the windows 321 and 322, a program corresponding to each window is reproduced.

[5. Modification 3: Example of calculating score by weighting]
When determining the score of a keyword to determine a recommended scene or program, some weighting may be performed.
The flowchart of FIG. 11 is a diagram illustrating an example of processing when determining the recommended program. In the flowchart of FIG. 11, the same step numbers are assigned to the same processes as those in the flowchart of FIG.

  The processing of the flowchart of FIG. 11 will be described. In step S14, the information processing unit 105 extracts the keyword that is characteristic for each scene from the text and calculates the score for each keyword, and is the same as the flowchart of FIG. It is. Then, the information processing unit 105 performs a weighting process on the extracted keyword score (step S27).

  The weighting process here is performed, for example, by reflecting the past usage state of the video recording apparatus 100. Specifically, in the case of matching with a keyword detected from a program reproduced in the past, the information processing unit 105 performs a process of multiplying a coefficient that slightly increases the score value. Alternatively, for a keyword detected from a program of a favorite genre registered in advance by the user, the information processing unit 105 performs a process of multiplying a coefficient that slightly increases the score value. Furthermore, a weighting process for lowering the score may be performed on keywords detected from programs of a genre that have been reproduced in the past.

Then, the information processing unit 105 obtains the average value of the same keyword for each broadcasting station (channel) on which the program of each scene is broadcast from the score subjected to the weighting process in step S27 (step S15). The processing after step S15 is the same as the flowchart of FIG.
The flowchart of FIG. 11 is applied to the process of determining the recommended scene shown in the flowchart of FIG. 2, but the same weighting process may be performed when determining the recommended program shown in FIG.

  As shown in the flowchart of FIG. 11, the score processing is performed after the keyword scores are weighted, so that the video recording apparatus 100 reflects the user's usage state and displays a recommended scene or a recommended program. I can decide.

[6. Modification 4: Example of digest playback of a recommended scene]
In the example of the embodiment described above, the recommended scene or the recommended program selected by the user from the list screen is reproduced. When performing playback based on this user operation, digest playback such as fast-forward playback may be performed.
The flowchart of FIG. 12 is a diagram illustrating a processing example in this case. The system controller 110 of the video recording apparatus 100 determines whether or not there is a digest playback selection operation in a state where the list screen indicating the recommended scene is displayed (step S31). If it is determined that there is an operation for selecting digest playback, the system controller 110 sends an instruction to the playback processing unit 108, and fast-forwards and plays back each scene displayed on the list screen (step S32).

FIG. 13 is a diagram showing an example of a screen 310 ′ for notifying the user of a recommended scene in this case. On the screen 310 ′, a window 311 showing the scene 2 of the program A, a window 312 showing the scene 4 of the program B, and a window 313 showing the scene 3 of the program E are displayed. Further, a button 314 for instructing digest reproduction is displayed on the screen 310 ′.
When the system controller 110 detects that the user has performed an operation of selecting the button 314, as shown in the lower side of FIG. 13, the scene 2 of the program A, the scene 4 of the program B, and the program E Next, the fast forward playback PB1 is performed. Here, the fast forward reproduction is performed at a speed at which the contents of the scene can be understood, for example, at a speed about twice the normal speed.

By providing such a digest playback button 314, the user can view the content of the recommended scene in a short time.
In the example of FIG. 13, the scene recommendation screen is applied. However, the program recommendation screen may include a button for instructing digest playback.

[7. Modification 5: Example of transferring a recommended scene to another terminal device]
The video recording apparatus 100 illustrated in FIG. 1 includes a terminal interface unit 106 that is connected to a terminal apparatus 200 such as a smartphone. The video recording device 100 may transfer the recommended scene or the recommended program to the external terminal device 200 via the terminal interface unit 106.

  The flowchart in FIG. 14 is a diagram illustrating a processing example in this case. The system controller 110 of the video recording apparatus 100 determines whether there is a batch transfer operation in a state where the list screen indicating the recommended scene is displayed (step S41). If it is determined that there is a batch transfer operation, the system controller 110 reads all scene data displayed on the list screen from the recording medium 120 and transmits the data from the terminal interface unit 106 to the terminal device 200 (step S42). .

  If it is determined in step S41 that there is no batch transfer operation, the system controller 110 determines whether there is any scene transfer operation (step S43). Here, when it is determined that there is a scene-unit transfer operation, the system controller 110 reads the data of the designated scene from the recording medium 120 and transmits the data from the terminal interface unit 106 to the terminal device 200 (step S44).

FIG. 15 is a diagram showing an example of a screen 310 ″ for notifying a user of a recommended scene in this case. On the screen 310 ″, a window 311 indicating the scene 2 of the program A and a window indicating the scene 4 of the program B are shown. 312 and a window 313 showing the scene 3 of the program E are displayed. Further, on the screen 310 ″, a button 315 for instructing batch transfer and three buttons 311b, 312b, and 313b for instructing transfer of each scene are displayed.
When the system controller 110 detects that the user has performed an operation of selecting the button 315, the three scenes (the scene 2 of the program A, the scene 4 of the program B, the scene 4 of the program A) are displayed. Transfer the data of scene 3) of program E. In addition, when the system controller 110 detects that the user has performed an operation of selecting the button 311b, the system controller 110 transfers the data of the scene 2 of the program A indicated by the window 311. Similarly, the operation of the button 312b or 313b causes the system controller 110 to transfer the data of the scene 4 of the program B or the scene 3 of the program E.

  By transferring to the external terminal device 200 in this way, for example, the display panel 201 included in the terminal device 200 shown in FIG. 15 can display the transferred scene or program.

[8. Other variations]
In the flowchart of FIG. 2, the recommended scenes and program candidates are limited to programs of a specific genre such as news specified in advance. On the other hand, the video recording apparatus 100 may search for all programs recorded in the recording medium 120 within a certain period as recommended scenes or program candidates.
Further, the process (TFIDF method) for obtaining the keyword score for each scene described in the process of the flowchart of FIG. 2 is an example, and the video recording apparatus 100 may obtain the keyword score by other processes.

  In the example of the embodiment described above, the video recording apparatus 100 is an apparatus that receives and accumulates broadcast signals. On the other hand, the processing of the present disclosure may be applied to a recording apparatus that can store other video content.

In the example of the embodiment described above, an example in which the video recording apparatus 100 performs processing has been described. On the other hand, for example, a computer (software) that performs the processing procedure of the present disclosure is installed in a computer device that performs various data processing so that the computer device analyzes an image and searches stored video content. It may be. The program may be prepared by storing it in various storage media, for example.
Further, the configurations and processes described in the claims of the present invention are not limited to the above-described embodiments. It will be understood by those skilled in the art that various modifications, combinations, and other embodiments can be made without departing from the spirit of the present invention.

In addition, this indication can also take the following structures.
(1)
A receiver for receiving video content;
A content accumulation unit for accumulating video content obtained by the reception unit;
A feature extraction unit that extracts a keyword that is characteristic for each scene obtained by dividing each video content accumulated by the content accumulation unit, and calculates a score according to the frequency of the extracted keyword in the scene;
A cumulative operation unit that cumulatively adds the keyword score of each video content calculated by the feature extraction unit for each keyword to obtain a cumulative value;
Among the scenes in which keywords having a cumulative score obtained by the cumulative calculation unit extracted from a threshold value or more are extracted, a scene that matches a predetermined condition or a video content including the scene is selected as a recommended scene or video content. A video recording apparatus comprising a selection unit.
(2)
The receiving unit is a receiving unit that receives a plurality of channels simultaneously,
The accumulation calculation unit calculates an average score of each keyword of each channel received by the reception unit, and accumulates and accumulates the average score of each keyword of each channel for all the channels for each keyword. A value is obtained. The video recording apparatus according to (1).
(3)
The selection unit excludes a keyword extracted from only video content of one channel among keywords whose cumulative value obtained by the cumulative calculation unit is equal to or greater than a threshold from candidates for determining a recommended scene or video content. The video recording device according to (1) or (2).
(4)
The selection unit selects a scene that matches the predetermined condition for each keyword, and when the same scene is selected with a different keyword, recommends a keyword having a lower cumulative score value. Alternatively, the video recording apparatus according to any one of (1) to (3), wherein the video recording apparatus is excluded from candidates for determining video content.
(5)
The process of selecting a scene that matches the predetermined condition is a process of selecting a scene having the longest time among scenes from which the corresponding keyword is extracted. The process according to any one of (1) to (4), Video recording device.
(6)
The process of calculating the score according to the frequency of the extracted keyword in the scene by the feature extraction unit is a process of calculating the score based on two frequency indicators of the word appearance frequency and the reverse document frequency. The video recording device according to any one of (1) to (5).
(7)
When calculating the score according to the frequency of the extracted keyword in the scene, the feature extraction unit weights the score based on the image or sound in the scene. Any of (1) to (6) 2. A video recording apparatus according to claim 1.
(8)
The selection unit determines the number of scenes from which keywords having a cumulative score value equal to or greater than a threshold are present in one video content, and selects video content having a large number of scenes as recommended video content. The video recording device according to any one of to (7).
(9)
The video recording apparatus according to any one of (1) to (8), further including a reproduction unit that performs digest reproduction of the scene or video content selected by the selection unit by a predetermined operation.
(10)
The video recording apparatus according to any one of (1) to (9), further including a transfer unit that transfers the scene or video content selected by the selection unit to an external device by a predetermined operation.
(11)
A score calculation process for extracting a keyword that is characteristic for each scene obtained by dividing the accumulated video content into a plurality of scenes, and calculating a score according to the frequency of the extracted keyword in the scene;
Cumulative calculation processing for obtaining a cumulative value by accumulating the score of each video content keyword calculated in the score calculation processing for each keyword;
Among scenes from which keywords having a cumulative score value obtained by the cumulative calculation process are extracted that are equal to or greater than a threshold, a scene that meets a predetermined condition or a video content including the scene is selected as a recommended scene or video content. An information processing method including a selection process.
(12)
A score calculation procedure for extracting a keyword that is characteristic for each scene obtained by dividing the accumulated video content into a plurality of scenes, and calculating a score according to the frequency of the extracted keyword in the scene;
A cumulative calculation procedure for accumulating the score of each video content keyword calculated in the score calculation procedure for each keyword to obtain a cumulative value;
Among scenes from which keywords having a cumulative score obtained by the cumulative calculation procedure are extracted that are equal to or greater than a threshold, a scene that matches a predetermined condition or a video content including the scene is selected as a recommended scene or video content. The selection procedure,
A recording medium that records a program to be executed by a computer.

  DESCRIPTION OF SYMBOLS 90 ... Internet 100 ... Video recording apparatus 101 ... Antenna 102 ... Tuner 103 ... Network interface part 104 ... Recording processing part 105 ... Information processing part 106 ... Terminal interface part 107 ... Memory 108 ... Reproduction processing 109: Display processing unit, 110: System controller, 111: Image / audio feature extraction unit, 120 ... Recording medium, 130 ... Disk drive unit, 141 ... Control bus, 142 ... Data bus, 200 ... Terminal device, 201 ... Display Panel, 300 ... display device, 310, 310 ', 310 ", 320, 330 ... screen

Claims (12)

A receiver for receiving video content;
A content accumulation unit for accumulating video content obtained by the reception unit;
A feature extraction unit that extracts a keyword that is characteristic for each scene obtained by dividing each video content accumulated by the content accumulation unit, and calculates a score according to the frequency of the extracted keyword in the scene;
A cumulative operation unit that cumulatively adds the keyword score of each video content calculated by the feature extraction unit for each keyword to obtain a cumulative value;
Among the scenes in which keywords having a cumulative score obtained by the cumulative calculation unit extracted from a threshold value or more are extracted, a scene that matches a predetermined condition or a video content including the scene is selected as a recommended scene or video content. A video recording apparatus comprising a selection unit. The receiving unit is a receiving unit that receives a plurality of channels simultaneously,
The accumulation calculation unit calculates an average score of each keyword of each channel received by the reception unit, and accumulates and accumulates the average score of each keyword of each channel for all the channels for each keyword. The video recording device according to claim 1, wherein a value is obtained. The selection unit excludes a keyword extracted from only video content of one channel among keywords whose cumulative value obtained by the cumulative calculation unit is equal to or greater than a threshold from candidates for determining a recommended scene or video content. Item 3. A video recording apparatus according to Item 2. The selection unit selects a scene that matches the predetermined condition for each keyword, and when the same scene is selected with a different keyword, recommends a keyword having a lower cumulative score value. The video recording apparatus according to claim 3, wherein the video recording apparatus is excluded from candidates for determining video content. The video recording apparatus according to claim 1, wherein the process of selecting a scene that matches the predetermined condition is a process of selecting a scene having the longest time among scenes from which corresponding keywords are extracted. The process of calculating the score according to the frequency of the extracted keyword in the scene by the feature extraction unit is a process of calculating the score based on two frequency indicators of a word appearance frequency and a reverse document frequency. Item 2. A video recording apparatus according to Item 1. The video recording apparatus according to claim 6, wherein the feature extraction unit weights the score based on an image or sound in the scene when calculating the score according to the frequency of the extracted keyword in the scene. 2. The selection unit determines the number of scenes in which one or more scenes from which a cumulative score value is equal to or greater than a threshold is present in one video content, and selects video content having a large number of scenes as recommended video content. Video recording device. The video recording apparatus according to claim 1, further comprising: a reproduction unit that performs digest reproduction of the scene or video content selected by the selection unit by a predetermined operation. The video recording apparatus according to claim 1, further comprising: a transfer unit that transfers the scene or video content selected by the selection unit to an external device by a predetermined operation. A score calculation process for extracting a keyword that is characteristic for each scene obtained by dividing the accumulated video content into a plurality of scenes, and calculating a score according to the frequency of the extracted keyword in the scene;
Cumulative calculation processing for obtaining a cumulative value by accumulating the score of each video content keyword calculated in the score calculation processing for each keyword;
Among scenes from which keywords having a cumulative score value obtained by the cumulative calculation process are extracted that are equal to or greater than a threshold, a scene that meets a predetermined condition or a video content including the scene is selected as a recommended scene or video content. An information processing method including a selection process. A score calculation procedure for extracting a keyword that is characteristic for each scene obtained by dividing the accumulated video content into a plurality of scenes, and calculating a score according to the frequency of the extracted keyword in the scene;
A cumulative calculation procedure for accumulating the score of each video content keyword calculated in the score calculation procedure for each keyword to obtain a cumulative value;
Among scenes from which keywords having a cumulative score obtained by the cumulative calculation procedure are extracted that are equal to or greater than a threshold, a scene that matches a predetermined condition or a video content including the scene is selected as a recommended scene or video content. The selection procedure,
A recording medium that records a program to be executed by a computer.

Images