JP2013229873A - Candidate selection device, candidate selection method, and computer program - Google Patents

Abstract

PROBLEM TO BE SOLVED: To present a more accurate input candidate to a user who inputs a character string depending on a reproduced video.SOLUTION: A character string is selected as a user's input candidate in accordance with an elapsed time from the start of reproduction of video data reproduced on an output device.

Description

  The present invention relates to a technique for assisting character string input.

  In recent years, streaming distribution of moving images to personal computers and portable terminals has been performed. There has also been proposed a technique for accepting posting of a comment corresponding to the playback time of a video (video playback time) and displaying the video and the comment in accordance with the video playback time. For example, Patent Literature 1 discloses a technique for linking live distribution with a communication function for users. In this technique, live video to be distributed and message information (comments) transmitted from a chat participant's terminal are displayed in synchronization with a designated layout on the same screen.

  A chat participant inputs a comment by operating an input device such as a keyboard, a touch panel, or a button. In recent years, several techniques for assisting the input of character strings have been proposed. For example, Patent Literature 2 discloses a technique for generating a supplement candidate character string and presenting it as a supplement candidate character string for a character string being input.

JP 2004-015750 A JP 2007-058444 A

  In the technique disclosed in Patent Document 2, the input assistance process does not change according to the reproduced moving image. However, the character string that the user intends to input as a comment may change depending on the moving image being played. For this reason, in the technique disclosed in Patent Document 2, the character string that the user intends to input may not be appropriately presented.

  In view of the above circumstances, an object of the present invention is to provide a technique for presenting input candidates with higher accuracy to a user who inputs a character string in accordance with a reproduced moving image.

  One aspect of the present invention is a candidate selection device including a selection unit that selects a character string as a user input candidate in accordance with an elapsed time from the start of reproduction of moving image data reproduced on an output device.

  One aspect of the present invention is the above-described candidate selecting device, which is obtained by the coefficient obtaining unit that obtains a coefficient that changes according to the elapsed time from the start of reproduction of the moving image data that is reproduced on the output device, and the coefficient obtaining unit. A score calculation unit that calculates a score of a character string based on the coefficient, and the selection unit selects the character string having a high score as a character string of a user input candidate.

  One aspect of the present invention is the candidate selection device described above, wherein the selection unit displays the input candidate selected in the past and the input candidate selected at the present time on the output device, and has already been displayed in a certain past. Do not select the selected input candidate.

  One aspect of the present invention is a candidate selection method including a selection step of selecting a character string as a user input candidate in accordance with an elapsed time from the start of reproduction of moving image data reproduced on an output device.

  One aspect of the present invention is a computer program for causing a computer to execute a selection step of selecting a character string as a user input candidate in accordance with an elapsed time from the start of reproduction of moving image data reproduced on an output device.

  According to the present invention, it is possible to present input candidates with higher accuracy to a user who inputs a character string according to a reproduced moving image.

1 is a system configuration diagram illustrating a system configuration of a moving image distribution system 1. FIG. 3 is a schematic block diagram illustrating a functional configuration of the first embodiment (terminal device 100a) of the terminal device 100. FIG. It is a figure which shows the outline of an input candidate table. 6 is a diagram illustrating a specific example of a screen 20 displayed by an output unit 102. FIG. It is a flowchart which shows the flow of operation | movement of the terminal device 100a. It is a figure which shows the screen 20 displayed by the process of the selection part 107 comprised as a modification. It is a schematic block diagram showing the function structure of 2nd embodiment (terminal device 100b) of the terminal device 100. FIG. It is a figure which shows the specific example of coefficient information. It is a figure which shows the specific example of a parameter table. It is a schematic block diagram showing the function structure of the coefficient information generation apparatus 600 which produces | generates coefficient information. It is a figure which shows the outline of a comment category table. It is a figure which shows the outline of a comment table. 3 is a schematic block diagram illustrating a functional configuration of a comment classification device 700. FIG. It is a figure which shows the outline of a category definition table. It is a figure which shows the outline of a moving image genre table. It is a figure which shows the concept of a co-occurrence index. It is a schematic block diagram showing the function structure of 3rd embodiment (terminal device 100c) of the terminal device 100. It is a figure which shows the outline of the input candidate table used in 3rd embodiment. It is a figure which shows the specific example of the screen 20 displayed by the output part 102 in 3rd embodiment. 5 is a schematic block diagram illustrating a functional configuration of an input candidate table generation device 800. FIG. 6 is a diagram illustrating an example of clustering processing of a cluster generation unit 802. FIG.

[Outline]
Hereinafter, the terminal device 100 which is one Embodiment of this invention is demonstrated. The terminal device 100 selects a character string as a user input candidate according to the elapsed time (moving image playback time) from the start of playback of the moving image data played back on the output device. Then, the terminal device 100 provides the user with the character string of the selected input candidate displayed on the output device. The user can easily input a character string to the terminal device 100 by selecting the displayed input candidate character string. The input candidate character string provided to the user changes according to the elapsed time (moving image reproduction time) from the start of reproduction of the reproduced moving image data. Therefore, input candidates corresponding to the contents of the moving image data being reproduced are provided. Therefore, it becomes possible to present input candidates with higher accuracy to a user who inputs a character string according to a reproduced moving image. The movie playback time represents the elapsed time when playback is started from the beginning of the movie data.

[System configuration]
FIG. 1 is a system configuration diagram illustrating a system configuration of the moving image distribution system 1. The moving image distribution system 1 includes a terminal device 100, an input candidate server 200, a moving image distribution server 300, a comment server 400, and a network 500. Each device is connected to be communicable with each other via a network 500. Note that the moving image distribution system 1 may include a plurality of devices.

  The terminal device 100 is a device operated by a user, and receives distribution of moving image data from the moving image distribution server 300. The terminal device 100 is a communicable information processing device such as a mobile phone, a PDA (Personal Digital Assistant), a smartphone, a tablet personal computer, a notebook personal computer, a stationary personal computer, a car navigation device, a television receiver, and a fixed telephone. Device. The terminal device 100 selects a character string as a user input candidate according to the elapsed time (moving image playback time) from the start of playback of the moving image data played back on the output device. Then, the terminal device 100 provides the user with the character string of the selected input candidate displayed on the output device.

  The input candidate server 200 stores an input candidate table in advance. The input candidate table is a table in which a moving image reproduction time is associated with a character string that is likely to be input by the user during the moving image reproduction time. The input candidate server 200 transmits an input candidate table to the terminal device 100 via the network 500.

The moving image distribution server 300 stores moving image data in advance. The moving image distribution server 300 distributes moving image data to the terminal device 100 via the network 500.
The comment server 400 distributes the comment table to the terminal device 100 via the network 500. The comment table is a table in which character information (comments) posted from the terminal device 100 receiving distribution of moving image data is associated with the moving image playback time at the time the comment is posted. In the terminal device 100, a comment is displayed according to the moving image playback time of the moving image data.

[First embodiment]
FIG. 2 is a schematic block diagram illustrating a functional configuration of the first embodiment (terminal device 100a) of the terminal device 100. The terminal device 100a includes a CPU (Central Processing Unit), a memory, an auxiliary storage device, and the like connected by a bus, and executes a program. By executing the program, the terminal device 100a functions as a device including an input unit 101, an output unit 102, a communication unit 103, a moving image playback unit 104, a comment control unit 105, a storage unit 106, and a selection unit 107. Note that all or part of each function of the terminal device 100a may be realized using hardware such as an application specific integrated circuit (ASIC), a programmable logic device (PLD), or a field programmable gate array (FPGA). The program executed by the terminal device 100a may be recorded on a computer-readable recording medium. The computer-readable recording medium is, for example, a portable medium such as a flexible disk, a magneto-optical disk, a ROM, a CD-ROM, or a storage device such as a hard disk built in the computer system. The program executed by the terminal device 100a may be transmitted via a telecommunication line.

  The input unit 101 is configured using an existing input device such as a keyboard, a pointing device (such as a mouse or a tablet), a button, or a touch panel. The input unit 101 is operated by the user when inputting a user instruction to the terminal device 100a. The input unit 101 may be an interface for connecting the input device to the terminal device 100a. In this case, the input unit 101 inputs an input signal generated in response to a user input in the input device to the terminal device 100a.

  The output unit 102 is a display device such as a CRT (Cathode Ray Tube) display, a liquid crystal display, or an organic EL (Electro Luminescence) display. The output unit 102 displays the video of the moving image data being reproduced and the character string of the comment. The output unit 102 may be an interface for connecting the image display device to the terminal device 100a. In this case, the output unit 102 generates a video signal for displaying the video of the moving image data being reproduced and the character string of the comment, and outputs the video signal to the image display device connected to itself.

  The communication unit 103 communicates with other devices via the network 500. For example, the communication unit 103 receives an input candidate table from the input candidate server 200. For example, the communication unit 103 receives moving image data from the moving image distribution server 300 via the network 500. For example, the communication unit 103 receives a comment table from the comment server 400 via the network 500.

  The video playback unit 104 generates data (video request data) for requesting distribution of video data in accordance with a user instruction input via the input unit 101. The moving image reproduction unit 104 transmits the moving image request data to the moving image distribution server 300 via the communication unit 103. When the moving image distribution server 300 receives the moving image request data, the moving image distribution server 300 reads out the moving image data requested by the received moving image request data and transmits the read moving image data to the requesting terminal device 100a. The moving image reproduction unit 104 receives the moving image data transmitted from the moving image distribution server 300 according to the moving image request data via the communication unit 103. The moving image reproducing unit 104 reproduces moving image data and outputs video to the output unit 102.

  The comment control unit 105 generates data (comment request data) for requesting a comment table corresponding to the moving image data requested by the moving image reproducing unit 104. The comment request data includes, for example, identification information (moving image ID) for specifying moving image data. The comment control unit 105 transmits comment request data to the comment server 400 via the communication unit 103. Upon receiving the comment request data, the comment server 400 reads a comment table corresponding to the moving image ID included in the received comment request data, and transmits the read comment table to the requesting terminal device 100a. The comment control unit 105 receives a comment table transmitted from the comment server 400 according to the comment request data via the communication unit 103. The comment control unit 105 acquires a comment corresponding to the reproduction time (moving image reproduction time) of the moving image data from the comment table, and outputs the comment to the output unit 102.

  The storage unit 106 is configured using a storage device such as a magnetic hard disk device or a semiconductor storage device. The storage unit 106 stores an input candidate table. FIG. 3 is a diagram showing an outline of the input candidate table. The input candidate table has a plurality of records 61 in which character strings to be input candidates are associated with moving image playback times. The input candidate table is a table generated for each moving image (for each moving image ID). The character string recorded in the record 61 is a character string that is likely to be input by the user during the moving image playback time of the record 61. The character string included in the record 61 is provided to the user as an input candidate during the time indicated by the moving image playback time.

  For example, the character string “wwwwwwwww” is provided to the user as an input candidate during the video playback time from 0:00: 00 (0 hour 0 minute 0 second) to 0:52:00 (0 hour 52 minute 00 second) Is done. For example, the character string “888888” indicates that the video playback time is between 0:00:20 (0 hour, 0 minute, 20 seconds) and 0:26:20 (0 hour, 26 minutes, 20 seconds), and 0:45:00. It is provided to the user as an input candidate between (0 hours 45 minutes 0 seconds) and 0:50:00 (0 hours 50 minutes 0 seconds). For example, the character string “Sweet yeah yeah” is used as an input candidate when the video playback time is between 0:12:30 (0 hours 12 minutes 30 seconds) and 0:13:20 (0 hours 13 minutes 20 seconds). Provided to the user.

  The input candidate table may be generated by any process. The input candidate table is created, for example, by analyzing comments posted in the past on moving images. The input candidate table may be created by a human analyzing a comment posted in the past on a moving image, or may be automatically created by a device that analyzes a comment posted in the past on a moving image. Further, the input candidate table may be created based on the intuition of the operator by human work, regardless of comments posted in the past on the moving image.

  Returning to FIG. 2, the description of the configuration of the terminal device 100a will be continued. The selection unit 107 generates data (input candidate request data) for requesting an input candidate table corresponding to the moving image data requested by the moving image reproducing unit 104. The input candidate request data includes, for example, identification information (moving image ID) for specifying moving image data. The selection unit 107 transmits input candidate request data to the input candidate server 200 via the communication unit 103. When receiving the input candidate request data, the input candidate server 200 reads the input candidate table corresponding to the moving image ID included in the received input candidate request data, and transmits the read input candidate table to the requesting terminal device 100a. The selection unit 107 receives the input candidate table transmitted from the input candidate server 200 according to the input candidate request data via the communication unit 103 and records it in the storage unit 106.

  The selection unit 107 reads one or a plurality of character strings from the input candidate table of the storage unit 106 according to the reproduction time (moving image reproduction time) of the moving image data being reproduced by the moving image reproduction unit 104. Specifically, it is as follows. The selection unit 107 acquires the video playback time from the video playback unit 104. The selection unit 107 searches for a record 61 having a moving image playback time value corresponding to the acquired moving image playback time. The selection unit 107 selects a character string of the record 61 obtained by the search as an input candidate. The selection unit 107 displays the selected character string as an input candidate on the output unit 102.

FIG. 4 is a diagram illustrating a specific example of the screen 20 displayed by the output unit 102. The screen 20 has a moving image display area 210 and an input display area 220. Hereinafter, each region will be described.
In the moving image display area 210, a video of moving image data reproduced by the moving image reproducing unit 104 and a character string of a comment posted on the moving image are displayed. In FIG. 4, a character string 211 displayed in the moving image display area 210 is a comment. Character strings such as “blue”, “returned ...”, “Japan?”, “Beautiful” are also comment strings. For example, the comment character string 211 may appear so as to appear from the right end of the screen at the moving image playback time corresponding to the comment, cross the moving image display area 210 to the left, and disappear to the left end.

  In the input display area 220, a comment input field 22, a posting button 23, and a plurality of input candidate buttons 24 (24-1 to 24-10) are displayed. In the comment input field 22, a character string of a comment input by the user operating the input unit 101 is displayed. When the user performs an operation (click, tap, etc.) on the posting button 23 by operating the input unit 101, the comment control unit 105 displays the comment character string displayed in the comment input field 22 and the video at that time. The playback time and the video ID of the video being played back are acquired. The comment control unit 105 creates comment posting data by associating the acquired character string of the comment, the moving image playback time, and the moving image ID. Then, the comment control unit 105 transmits the created comment posting data to the comment server 400 via the communication unit 103. The comment server 400 registers the comment character string and the moving image playback time of the received comment posting data in the comment table corresponding to the moving image ID.

  Character strings selected as input candidates by the selection unit 107 are displayed on the input candidate buttons 24 (24-1 to 24-10). For example, the input candidate button 24-1 displays a character string “Sweet yeah yeah yeah”. When the user operates the input candidate button 24-1 by operating the input unit 101, the comment control unit 105 inputs the character string displayed on the input candidate button 24-1 into the comment input field 22. Thereafter, when an operation is performed on the posting button 23, comment posting data is created as described above and transmitted to the comment server 400.

  FIG. 5 is a flowchart showing a flow of operations of the terminal device 100a. First, the user operates the input unit 101 to select a moving image to be played (step S101). The moving image reproduction unit 104 generates moving image request data for requesting distribution of moving image data selected by the user. The video playback unit 104 transmits the generated video request data and receives video data from the video distribution server 300. In addition, the moving image reproduction unit 104 notifies the comment control unit 105 and the selection unit 107 of the ID (moving image ID) of the moving image selected by the user.

  The comment control unit 105 generates comment request data for requesting a comment table corresponding to the moving image selected by the user. The comment control unit 105 transmits the generated comment request data and receives a comment table from the comment server 400.

  The selection unit 107 generates input candidate request data for requesting an input candidate table corresponding to the moving image selected by the user. The selection unit 107 transmits the generated input candidate request data and receives an input candidate table from the input candidate server 200. The selection unit 107 writes the received input candidate table in the storage unit 106. In this manner, the terminal device 100a acquires an input candidate table corresponding to the moving image selected by the user (step S102).

  Next, the moving image reproduction unit 104 starts reproduction of the received moving image (step S103). Thereafter, the selection unit 107 selects an input candidate corresponding to the moving image reproduction time until the moving image reproduction ends, and displays the input candidate as the input candidate button 24 in the input display area 220 (steps S104 and S105—NO). When the reproduction of the moving image is finished (step S105—YES), the processing of the flowchart shown in FIG. 5 is finished.

  In the terminal device 100a of the first embodiment configured as described above, an input is made to the moving image reproduction time according to the elapsed time (moving image reproduction time) from the start of reproduction of the moving image data reproduced by the moving image reproducing unit 104. A character string having a high tendency is selected by the selection unit 107. Then, the character string selected by the selection unit 107 is provided to the user as an input candidate. Therefore, it is possible to present input candidates with higher accuracy.

<Modification>
In the moving image distribution system 1, some of the functions of the terminal device 100a may be implemented in another device. That is, the function of the terminal device 100a may be formed using a plurality of devices. The same applies to the input candidate server 200, the moving image distribution server 300, and the comment server 400. For example, the functions of the storage unit 106 and the selection unit 107 may be implemented in another device (for example, the input candidate server 200). In this case, the control unit of the terminal device 100a transmits the moving image reproduction time and the moving image ID of the moving image data to be reproduced to the other device, and sets the character strings that are input candidates according to the moving image reproduction time and the moving image ID to the other device. Receive from. Then, the control unit causes the output unit 102 to display the input candidate button 24 based on the received input candidate.

In the above description, the input candidate table is generated in advance for each moving image and stored in the input candidate server 200. However, the input candidate table may be generated in advance as a common table for a plurality of moving images. For example, an input candidate table common to each genre of moving images may be generated, or an input candidate table common to all moving images distributed by the moving image distribution server 300 may be generated.
The input candidate table is not transmitted from the input candidate server 200 to the terminal device 100 via the network 500, but may be stored in advance in the storage unit 106 of the terminal device 100.

  The value of the moving image playback time registered in each record 61 of the input candidate table may represent a ratio as a relative value instead of the time (hour, minute, second) as an absolute value as shown in FIG. For example, a value of 0% to 10% may be given as the moving image playback time. In this case, the moving image playback time represents the time from 0 second to 0.1 × T seconds with respect to the length T seconds of the entire moving image. With such a configuration, it is possible to cope with a case where the entire playback time of a moving image changes or a case where an input candidate table common to a plurality of moving images having different playback times is set.

A plurality of input candidate buttons 24 may not be displayed in the input display area 220, but only one input candidate button 24 may be displayed.
When the user inputs an operation on the input candidate button 24 by operating the input unit 101, the comment control unit 105 may execute processing up to the case where the operation on the posting button 23 is performed. Specifically, it is as follows. When the user operates the input candidate button 24 by operating the input unit 101, the comment control unit 105 displays the character string displayed on the input candidate button 24, the video playback time at that time, and the playback ID of the moving image (moving image ID) is acquired. The comment control unit 105 creates comment posting data by associating the acquired character strings of input candidates, the moving image playback time, and the moving image ID, and transmits the comment posting data to the comment server 400 via the communication unit 103. In this case, the posting button 23 may not be displayed.

  FIG. 6 is a diagram illustrating a screen 20 displayed by processing of the selection unit 107 configured as a modification. The selection unit 107 configured as a modification includes an input candidate button 24 for a character string related to the character string of the input candidate button 24 operated by the user (hereinafter referred to as “related character string”) as an input display area 220. To display. Hereinafter, a specific example of processing of the selection unit 107 configured as a modification will be described with reference to FIGS. 4 and 6.

  In FIG. 4, when the user of the terminal device 100 a operates the input candidate button 24-6 (the button corresponding to the character string “888888”), the selection unit 107 performs the input operated via the input unit 101. The character string “888888” corresponding to the candidate button 24-6 is acquired. Next, the selection unit 107 acquires a character string (related character string) related to the acquired character string “888888”. For example, the selection unit 107 stores in advance information indicating the category of each character string registered in the input candidate table (for example, a comment category table described later), and a character string in the same category as the character string “888888”. As a related character string. For example, the selection unit 107 stores in advance information (for example, a parameter table described later) indicating parameters indicating the strength of association with a plurality of categories for each character string registered in the input candidate table. A character string having a parameter value close to that of the character string “888888” may be acquired as a related character string.

  The selection unit 107 displays the related character string input candidate buttons 24 in the input display area 220. The selection unit 107 may display all the input candidate buttons 24 of the character strings selected from the input candidate table based on the moving image playback time and the input candidate buttons 24 of the related character strings. In this case, the input display area 220 may be larger than before, or the size of one input candidate button 24 may be smaller. Further, the selection unit 107 may display a part of the input candidate buttons 24 of the character strings selected from the input candidate table based on the moving image playback time and the input candidate buttons 24 of the related character strings. For example, the selection unit 107 preferentially displays the acquired related character string input candidate buttons 24, and displays as many character string input candidate buttons 24 selected from the input candidate table as can be displayed in the remaining portion. May be.

In addition, the selection unit 107 displays the operated input candidate button 24 at the upper left position of the input display area 220, and the related character string input candidate buttons in the vicinity thereof (for example, the uppermost row, the uppermost row, and the second row). 24, and the input candidate buttons 24 for the character strings selected from the input candidate table may be displayed in the remaining input display area 220. FIG. 6 shows a specific example of display by the selection unit 107 configured as described above. The above-described display method of the related character string input candidate buttons 24 is an example, and is not limited to the above method.
The input display area 220 need not always be displayed on the screen 20. For example, the input display area 220 may be displayed in response to an operation in which the user taps the input unit 101 configured using a touch panel.

[Second Embodiment]
FIG. 7 is a schematic block diagram illustrating a functional configuration of the terminal device 100 according to the second embodiment (terminal device 100b). The terminal device 100b is a terminal device according to the first embodiment in that it includes a coefficient information storage unit 106-1 and a parameter storage unit 106-2 instead of the storage unit 106, and includes a selection unit 107b instead of the selection unit 107. Unlike 100a, other configurations are the same as the terminal device 100a of the first embodiment. Hereinafter, the difference between the terminal device 100b of the second embodiment and the terminal device 100a of the first embodiment will be described.

  The coefficient information storage unit 106-1 is configured using a storage device such as a magnetic hard disk device or a semiconductor storage device. The coefficient information storage unit 106-1 stores coefficient information generated for each moving image. FIG. 8 is a diagram illustrating a specific example of coefficient information. In FIG. 8, the horizontal axis represents the moving image playback time, and the vertical axis represents the value of each coefficient. The coefficient information represents a time change of one or more kinds of coefficients. In other words, the coefficient information indicates the value of each coefficient for each moving image playback time. In the example of FIG. 8, the time change of the coefficient of three types of elements (laughter, excitement, sadness) is shown. Each coefficient represents the strength of each element in the content of the moving image being played back at that time for each moving image playback time. For example, at the time of the moving image playback time t1, the content of the moving image has a high degree of excitement and a low level of laughter and sadness. For example, at the time of moving image playback time t2, the content of the moving image has the highest degree of laughter, the second highest level of excitement, and the lower level of sadness. For example, at the time of moving image reproduction time t3, the content of the moving image has the highest degree of sadness and the degree of elements of laughter and excitement is low.

  The coefficient information may be generated by any process. The coefficient information is created, for example, by analyzing comments posted in the past on a moving image. The coefficient information may be created by a human analyzing a comment posted in the past on a moving image, or may be automatically created by a device that analyzes a comment posted in the past on a moving image. Also, the coefficient information may be created according to the impression of the worker by human work, regardless of comments posted in the past on the moving image.

  Returning to FIG. 7, the description of the configuration of the terminal device 100b will be continued. The parameter storage unit 106-2 is configured using a storage device such as a magnetic hard disk device or a semiconductor storage device. The parameter storage unit 106-2 stores a parameter table. FIG. 9 is a diagram showing a specific example of the parameter table. The parameter table includes a plurality of records 62 in which character strings that are input candidates and character string parameters are associated with each other. The character string recorded in the record 62 is provided to the user as an input candidate according to the parameter and the score calculated based on the coefficient in the moving image reproduction time of the reproduced moving image.

  Parameters C1 to C3 indicate the ratio of moving image elements when each character string is input. C1 indicates the value of the parameter regarding the laughing element, C2 indicates the value of the parameter regarding the excitement element, and C3 indicates the value of the parameter regarding sadness. A character string having a high C1 is a character string that has a strong tendency to be input when the moving image being played is a moving image having a high degree of laughter. A character string having a high C2 is a character string that has a strong tendency to be input when the moving image being played is a moving image having a high degree of excitement. A character string with a high C3 is a character string that has a strong tendency to be input when the moving image being played is a moving image with a high degree of sadness.

The parameter N is a parameter representing the frequency with which each character string is input regardless of the moving image playback time. It shows that the frequency with which the character string is input is so high that the value of N is large.
The parameter table may be generated by any process. The parameter table is created, for example, by analyzing comments posted in the past on moving images. The parameter table may be created by a human analyzing a comment posted in the past on a moving image, or may be automatically created by a device that analyzes a comment posted in the past on a moving image. Further, the parameter table may be created according to the impression of the worker by human work, regardless of comments posted in the past on the moving image.

  Returning to FIG. 7, the description of the configuration of the terminal device 100b will be continued. The selection unit 107b calculates a score corresponding to the video playback time of each character string based on the coefficient information and parameter table corresponding to the video being played back by the video playback unit 104, and inputs a character string having a high score as an input candidate Choose as. Formula 1 shown below is a specific example of a mathematical formula used when the selection unit 107b calculates a score.

  In Expression 1, x represents a character string, Score (x) represents the score of the character string x, Nx represents the value of the parameter N of the character string x, i represents an element, and M represents the number of element types. (M = 3 in the case of FIG. 8), Si represents the coefficient of the element i, and Cxi represents the character string x and the parameter of the element i (that is, Ci of the character string x).

  The selection unit 107b acquires the coefficient Si of each element according to the moving image playback time based on the coefficient information. The selection unit 107b acquires parameters (N, C1 to C3) of each character string. The selection unit 107b calculates the score of each character string at that time based on the coefficient Si and the parameter of each element. The selection unit 107b selects a predetermined number of character strings from the higher score as input candidates. The selection unit 107b causes the output unit 102 to display the input candidate button 24 of the selected input candidate.

The terminal device 100b according to the second embodiment operates according to the flowchart shown in FIG. In the process shown in step S104 of FIG. 5, the selection unit 107b calculates the score of each character string, and selects a character string having a high score as an input candidate.
In the terminal device 100b of the second embodiment configured as described above, an input is made in accordance with an element of the content of the moving image in the elapsed time (moving image reproduction time) from the start of reproduction of the moving image data reproduced by the moving image reproducing unit 104. A character string having a high tendency to be selected is selected by the selection unit 107b. Then, the character string selected by the selection unit 107b is provided to the user as an input candidate. Therefore, it is possible to present input candidates with higher accuracy.

<Modification>
The contents of the coefficients and parameters described above are merely examples. Accordingly, coefficients and parameters different from those described above may be used.
In the specific example shown in FIG. 9, the total value of the parameters C1 to C3 is 1. However, the total value of the parameters C1 to C3 may be different for each character string.
The parameter table may be generated for each moving image, may be generated for each moving image genre, or a table common to all moving images distributed by the moving image distribution server 300 may be generated.

The coefficient information may be provided from the input candidate server 200 to the terminal device 100b via the network 500, or may be stored in advance in the coefficient information storage unit 106-1 of the terminal device 100b.
The parameter table may be provided from the input candidate server 200 to the terminal device 100b via the network 500, or may be stored in advance by the parameter storage unit 106-2 of the terminal device 100b.
The terminal device 100b of the second embodiment may be modified and configured in the same manner as the terminal device 100a of the first embodiment.

[Coefficient information generator]
FIG. 10 is a schematic block diagram illustrating a functional configuration of a coefficient information generation apparatus 600 that generates coefficient information. The coefficient information generation device 600 includes a CPU, a memory, an auxiliary storage device, and the like connected by a bus, and executes a program. By executing the program, the coefficient information generation apparatus 600 functions as an apparatus including a comment category storage unit 601, a comment storage unit 602, a coefficient information generation unit 603, and a coefficient information storage unit 604. Note that all or part of the functions of the coefficient information generation apparatus 600 may be realized using hardware such as an ASIC, PLD, or FPGA. The program executed by the coefficient information generation apparatus 600 may be recorded on a computer-readable recording medium. The computer-readable recording medium is, for example, a portable medium such as a flexible disk, a magneto-optical disk, a ROM, a CD-ROM, or a storage device such as a hard disk built in the computer system. The program executed by the coefficient information generating apparatus 600 may be transmitted via a telecommunication line.

  The comment category storage unit 601 is configured using a storage device such as a magnetic hard disk device or a semiconductor storage device. The comment category storage unit 601 stores a comment category table. FIG. 11 is a diagram showing an outline of the comment category table. The comment category table includes a plurality of records 63 in which a character string of a comment is associated with a category to which the comment belongs (hereinafter referred to as “comment category”). The category to which each comment belongs may be determined based on the impression of the worker by human work, or may be automatically determined by a device that analyzes comments posted in the past on moving images.

  The comment storage unit 602 is configured using a storage device such as a magnetic hard disk device or a semiconductor storage device. The comment storage unit 602 stores a comment table for each moving image data. FIG. 12 is a diagram showing an outline of the comment table. The comment table includes a plurality of records 64 in which a comment posted to the moving image data is associated with a moving image playback time at the time when the comment is posted.

  The coefficient information generation unit 603 generates coefficient information for each moving image data based on the comment category table and the comment table. For example, the coefficient information generation unit 603 acquires comments posted during a predetermined time width (for example, 1 second, 5 seconds, 10 seconds, etc.) from the comment table. Next, the coefficient information generation unit 603 acquires the acquired category of each comment from the comment category table. Then, the coefficient information generation unit 603 determines the coefficient of the time zone (moving image playback time) according to the number of comments belonging to each category. For example, when nine comments belonging to laughter, one comment belonging to excitement, and zero comment belonging to sadness are posted in a certain time period, the coefficient information generation unit 603 sets the coefficient of laughter to 9 The coefficient of excitement is set to 1 and the coefficient of sadness is set to 0. The coefficient information generation unit 603 determines a coefficient over all moving image playback times and generates coefficient information.

  The coefficient information storage unit 604 is configured using a magnetic hard disk device or a semiconductor storage device. The coefficient information storage unit 604 stores the coefficient information generated by the coefficient information generation unit 603 for each moving image.

  According to the coefficient information generating apparatus 600 configured as described above, it is possible to automatically generate coefficient information based on the comment category table and the comment table. The coefficient information generated by the coefficient information generation apparatus 600 may be distributed to the terminal apparatus 100b by the input candidate server 200, or may be stored in advance by the terminal apparatus 100b.

[Comment classifier]
FIG. 13 is a schematic block diagram illustrating a functional configuration of the comment classification device 700. The comment classification device 700 includes a CPU, a memory, an auxiliary storage device, and the like connected by a bus, and executes a program. By executing the program, the comment classification device 700 includes a category definition storage unit 701, a comment storage unit 702, a moving image genre storage unit 703, a co-occurrence index calculation unit 704, a parameter storage unit 705, a classification unit 706, and a comment category storage unit 707. It functions as a device provided. All or some of the functions of the comment classification device 700 may be realized using hardware such as an ASIC, PLD, or FPGA. The program executed by the comment classification device 700 may be recorded on a computer-readable recording medium. The computer-readable recording medium is, for example, a portable medium such as a flexible disk, a magneto-optical disk, a ROM, a CD-ROM, or a storage device such as a hard disk built in the computer system. The program executed by the comment classification device 700 may be transmitted via a telecommunication line.

  The category definition storage unit 701 is configured using a storage device such as a magnetic hard disk device or a semiconductor storage device. The category definition storage unit 701 stores a category definition table. FIG. 14 is a diagram showing an outline of the category definition table. The category definition table includes a plurality of records 65 in which comment categories and typical comments (hereinafter referred to as “main comments”) among comments belonging to the comment category are associated with each other. A plurality of main comments may be associated with one comment category. The main comment corresponding to the comment category may be determined according to the impression of the worker by human work, or may be automatically determined by a device that analyzes comments posted in the past on the moving image.

  The comment storage unit 702 is configured using a storage device such as a magnetic hard disk device or a semiconductor storage device. The comment storage unit 702 stores a comment table for each moving image data. The comment table stored in the comment storage unit 702 has the same configuration as the comment table stored in the comment storage unit 602 of the coefficient information generation apparatus 600. That is, the comment table stored in the comment storage unit 702 is as shown in FIG. Therefore, detailed description of the comment table stored in the comment storage unit 702 is omitted.

  The moving picture genre storage unit 703 is configured using a storage device such as a magnetic hard disk device or a semiconductor storage device. The moving image genre storage unit 703 stores a moving image genre table. FIG. 15 is a diagram showing an outline of the moving image genre table. The moving image genre table includes a plurality of records 66 in which a moving image genre is associated with a moving image ID of moving image data belonging to the moving image genre. The moving image genre is a moving image classification based on moving image content, generation process, or the like. The distribution of the video ID to the video genre may be performed according to the impression of the worker by human work, or automatically by a device that analyzes additional information such as comments and tags posted in the past on the video. It may be done. One moving image ID may belong to a plurality of moving image genres.

  The co-occurrence index calculation unit 704 calculates a co-occurrence index with the main comment for each comment registered in the comment table based on the category definition table, the comment table, and the moving image genre table. The co-occurrence index is a value indicating the degree to which one comment and another comment appear together at the same timing in the same moving image. The co-occurrence index calculation unit 704 calculates a co-occurrence index with each main comment for each comment. Hereinafter, a specific example of the process of the co-occurrence index calculation unit 704 will be described.

  The co-occurrence index calculation unit 704 first selects one video genre based on the video genre table. Next, the co-occurrence index calculation unit 704 selects one moving image belonging to the selected moving image genre, and refers to the comment table of the selected moving image. Next, the co-occurrence index calculation unit 704 adds a comment posted within a predetermined time width (for example, 5 seconds, 10 seconds, 30 seconds, 1 minute) from the playback start time for the movie playback time for the referenced comment table. Each time, the co-occurrence index with the main comment of each comment category is calculated. The co-occurrence index calculation unit 704 acquires the calculated co-occurrence index as a co-occurrence index for the comment category. When a plurality of main comments are associated with one comment category, the co-occurrence index calculation unit 704 uses the total value of the co-occurrence indices with each main comment as the co-occurrence index for the comment category. get.

  FIG. 16 is a diagram illustrating the concept of the co-occurrence index. A circle shown in FIG. 16 indicates a comment category area. The center point 41 (41-1 to 41-3) indicates the center point of each comment category area. A point 42 represents the position of a certain comment. The distance l (11 to 13) between the point 42 and the center point 41 of each comment category region represents the reciprocal of the co-occurrence index between the comment at the point 42 and each comment category.

  The larger the value of the co-occurrence index, the smaller the distance l, indicating that the comment is likely to co-occur with the main comment of the comment category. That is, the larger the value of the co-occurrence index, the stronger the tendency that the comment is posted on a moving image having a higher degree of elements in the comment category. The smaller the value of the co-occurrence index, the greater the distance l, indicating that the comment is less likely to co-occur with the main comment in that comment category. That is, the smaller the value of the co-occurrence index, the weaker the tendency of comments to be posted to a moving image having a higher degree of elements in the comment category.

For example, the comment indicated by the point 42 shown in FIG. 16 has the strongest laughing element, the strongest excitement element, and the weak sadness element.
The co-occurrence index calculation unit 704 repeatedly executes the above processing from the beginning to the end of the video playback time of the selected video, and for each combination of comment and comment category, a statistical value of the co-occurrence index (for example, an average value or Mode and total value). The co-occurrence index calculation unit 704 performs the above-described processing on all the videos belonging to the selected video genre, and the co-occurrence index statistical values (for example, the average value, the mode value, (Total value) is calculated.

  In addition, the co-occurrence index calculation unit 704 counts the number of appearances of each comment for each video genre. Then, the co-occurrence index calculation unit 704 determines parameters for each comment category for each comment. For example, the co-occurrence index for each comment category is a parameter (for example, C1 to C3) indicating the ratio of moving image elements when the comment (character string) is input. Further, the number of appearances is a parameter (for example, N) indicating the frequency with which the comment (character string) is input.

  Through such processing, the co-occurrence index calculation unit 704 generates a parameter table for each comment for each moving image genre. For example, in the case of a comment indicated by a point 42 shown in FIG. 16, the parameter C1 is the highest, the parameter C2 is the second highest, and the parameter C3 is the smallest.

  The parameter storage unit 705 is configured using a storage device such as a magnetic hard disk device or a semiconductor storage device. The parameter storage unit 705 stores the parameter table generated by the co-occurrence index calculation unit 704.

  The classification unit 706 classifies comments into comment categories based on the parameter table stored in the parameter storage unit 705. The classification unit 706 generates a comment category table based on the classification result. For example, the classification unit 706 classifies comments into comment categories corresponding to the co-occurrence index with the highest comment. For example, in the case of FIG. 16, the comment indicated by the point 42 is classified into the comment category of laughter. The classification unit 706 records the generated comment category table in the comment category storage unit 707.

  The comment category storage unit 707 is configured using a storage device such as a magnetic hard disk device or a semiconductor storage device. The comment category storage unit 707 stores the comment category table generated by the classification unit 706.

  According to the comment classification device 700 configured as described above, it is possible to automatically classify comments into comment categories and generate a comment category table. The comment category table generated by the comment classification device 700 may be used by the terminal device 100 (100a, 100b) or may be used by the coefficient information generation device 600.

  Further, according to the comment classification device 700, it is possible to automatically determine a comment parameter and generate a parameter table. The parameter table generated by the comment classification device 700 may be distributed to the terminal device 100b by the input candidate server 200, or may be stored in advance by the terminal device 100b.

[Third embodiment]
Next, a third embodiment (terminal device 100c) of the terminal device 100 will be described. In the third embodiment, a predetermined number of input candidates are defined in advance every predetermined time.
FIG. 17 is a schematic block diagram illustrating a functional configuration of the third embodiment (terminal device 100 c) of the terminal device 100. The terminal device 100c is different from the terminal device 100a of the first embodiment in that it includes a storage unit 106c instead of the storage unit 106, and includes a selection unit 107c instead of the selection unit 107. It is the same as the terminal device 100a of the embodiment. Hereinafter, the difference between the terminal device 100c of the third embodiment and the terminal device 100a of the first embodiment will be described.

  The storage unit 106c is configured using a storage device such as a magnetic hard disk device or a semiconductor storage device. The storage unit 106c stores an input candidate table. FIG. 18 is a diagram showing an outline of an input candidate table used in the third embodiment. The input candidate table in the third embodiment includes a plurality of records 67 in which moving image playback times divided at predetermined intervals (for example, 5 seconds) are associated with character strings that are input candidates. In the specific example shown in FIG. 18, the record 67 has three input candidates (first candidate, second candidate, and third candidate). The input candidate table is a table generated for each moving image (for each moving image ID). The first to third candidates recorded in the record 67 are character strings that are likely to be input by the user during the video playback time of the record 67. The first to third candidate character strings included in the record 67 are provided to the user as input candidates during the time indicated by the moving image playback time.

  For example, between 0:00:00 (0 hour 0 minute 0 second) and 0:00:05 (0 hour 0 minute 5 second), “888888”, “blue”, “returned” Three character strings “” are provided to the user as input candidates. Between 0:00:05 (0 hour, 0 minute, 5 seconds) and 0:00:10 (0 hour, 0 minute, 10 seconds), the video playback time is “Otokare”, “Earth?”, “Barrage” Three character strings are provided to the user as input candidates. Between 0:00:10 (0 hours 0 minutes 10 seconds) and 0:00:15 (0 hours 0 minutes 15 seconds), “www”, “everything is nostalgic”, “42” Three character strings are provided to the user as input candidates.

  The input candidate table may be generated by any process. The input candidate table is created, for example, by analyzing comments posted in the past on moving images. The input candidate table may be created by a human analyzing a comment posted in the past on a moving image, or may be automatically created by a device that analyzes a comment posted in the past on a moving image. Further, the input candidate table may be created based on the intuition of the operator by human work, regardless of comments posted in the past on the moving image.

  Returning to FIG. 17, the description of the configuration of the terminal device 100c will be continued. The selection unit 107c generates data (input candidate request data) for requesting an input candidate table corresponding to the moving image data requested by the moving image reproducing unit 104. The input candidate request data includes, for example, identification information (moving image ID) for specifying moving image data. The selection unit 107 c transmits input candidate request data to the input candidate server 200 via the communication unit 103. When receiving the input candidate request data, the input candidate server 200 reads an input candidate table corresponding to the moving image ID included in the received input candidate request data, and transmits the read input candidate table to the requesting terminal device 100c. The selection unit 107c receives the input candidate table transmitted from the input candidate server 200 according to the input candidate request data via the communication unit 103, and records it in the storage unit 106c.

  The selection unit 107c reads one or more character strings from the input candidate table of the storage unit 106c in accordance with the reproduction time (moving image reproduction time) of the moving image data being reproduced by the moving image reproducing unit 104. Specifically, it is as follows. The selection unit 107 c acquires the moving image reproduction time from the moving image reproduction unit 104. The selection unit 107c searches for a record 67 having a moving image reproduction time value corresponding to the acquired moving image reproduction time. The selection unit 107c selects the first candidate, the second candidate, and the third candidate character string of the record 67 obtained by the search as input candidates. The selection unit 107c causes the output unit 102 to display the selected character string as an input candidate.

FIG. 19 is a diagram illustrating a specific example of the screen 20 displayed by the output unit 102 in the third embodiment. The screen 20 has a moving image display area 210 and an input display area 220. Hereinafter, each region will be described.
In the moving image display area 210, a video of moving image data reproduced by the moving image reproducing unit 104 and a character string of a comment posted on the moving image are displayed. In the input display area 220, a plurality of input candidate buttons 25 (25-1 to 25-9) are displayed. When the user performs an operation (click, tap, etc.) on the input candidate button 25 by operating the input unit 101, the comment control unit 105 displays the character string (comment character string) of the operated input candidate button 25, The video playback time at that time and the video ID of the video being played back are acquired. The comment control unit 105 creates comment posting data by associating the acquired character string of the comment, the moving image playback time, and the moving image ID. Then, the comment control unit 105 transmits the created comment posting data to the comment server 400 via the communication unit 103. The comment server 400 registers the comment character string and the moving image playback time of the received comment posting data in the comment table corresponding to the moving image ID.

  The input candidate buttons 25 (25-1 to 25-9) display a predetermined number of histories of character strings selected as input candidates by the selection unit 107c. For example, the three input candidate buttons 25-1 to 25-3 arranged vertically on the right side of the input display area 220 are used for inputting the first to third candidate character strings selected at the current timing. Button. The first candidate character string is displayed on the upper input candidate button 25-1, the second candidate character string is displayed on the middle input candidate button 25-2, and the lower input candidate button 25-3 is displayed on the lower input candidate button 25-3. The third candidate character string is displayed. Further, the three input candidate buttons 25-4 to 25-6 vertically arranged in the center of the input display area 220 are used to input the first to third candidate character strings selected at the previous timing. It is a button. The first candidate character string is displayed on the upper input candidate button 25-4, the second candidate character string is displayed on the middle input candidate button 25-5, and the lower input candidate button 25-6 is displayed on the lower input candidate button 25-6. The third candidate character string is displayed. The three input candidate buttons 25-7 to 25-9 arranged vertically on the left side of the input display area 220 are for inputting the character strings of the first candidate to the third candidate selected at the previous timing. It is a button. The first candidate character string is displayed on the upper input candidate button 25-7, the second candidate character string is displayed on the middle input candidate button 25-8, and the lower input candidate button 25-9 is displayed on the lower input candidate button 25-9. The third candidate character string is displayed.

The specific example shown in FIG. 19 shows a display example in which the moving image playback time is from 0:00:10 to 0:00:15. In this case, the first candidate “www”, the second candidate “everything is missing”, and the third candidate “42” searched based on the input candidate table at the current time point are input candidate buttons 25-1 to 25-, respectively. 3 is displayed. First candidate “Otsukare”, second candidate “Earth?”, Third, searched based on input candidate table at previous timing (between 0:00:05 and 0:00:10), third Candidate “barrages” are displayed on the input candidate buttons 25-4 to 25-6, respectively. The first candidate “888888”, the second candidate “blue”, and the third candidate “returned” searched based on the input candidate table at the previous timing (between 0:00:00 and 0:00:05) "..." is displayed on the input candidate buttons 25-7 to 25-9, respectively.
It should be noted that for character string input candidates of a predetermined number or more (for example, 7 characters or more), not all of the character strings may be displayed but only a part thereof may be displayed.

  In the terminal device 100c of the third embodiment configured as described above, the moving image playback time is input according to the elapsed time (moving image playback time) from the start of playback of the moving image data played back by the moving image playback unit 104. A character string having a high tendency is selected by the selection unit 107c. Then, the character string selected by the selection unit 107c is provided to the user as an input candidate. As described above, since a character string having a high tendency to be input every moving image playback time is provided as an input candidate, it is possible to present an input candidate with higher accuracy.

  In addition, in the terminal device 100c of the third embodiment, a button of a character string of input candidates selected at a past timing is displayed. Therefore, it is possible to present various variations of input candidates to the user.

<Modification>
In the specific example of the input candidate table shown in FIG. 18, each record 67 has three input candidates (first candidate to third candidate), but the number of input candidates is not limited to three. Each record 67 may have one input candidate, two input candidates, or four or more input candidates. Further, the number of input candidates may be different for each record 67.
In the input display area 220 shown in FIG. 19, selection histories from the present to the previous two timings are displayed as input candidate buttons 25. However, the displayed input candidate button 25 may be a selection history at the current and previous timing, or may be a selection history at a farther past timing.

[Input candidate table generator]
FIG. 20 is a schematic block diagram illustrating a functional configuration of an input candidate table generation device 800 that generates an input candidate table used in the terminal device 100c (third embodiment). The input candidate table generation device 800 includes a CPU, a memory, an auxiliary storage device, and the like connected by a bus, and executes a program. By executing the program, the input candidate table generation device 800 functions as a device including a comment storage unit 801, a cluster generation unit 802, a table generation unit 803, and a table storage unit 804. Note that all or part of the functions of the input candidate table generation apparatus 800 may be realized using hardware such as an ASIC, PLD, or FPGA. The program executed by the input candidate table generation device 800 may be recorded on a computer-readable recording medium. The computer-readable recording medium is, for example, a portable medium such as a flexible disk, a magneto-optical disk, a ROM, a CD-ROM, or a storage device such as a hard disk built in the computer system. The program executed by the input candidate table generation device 800 may be transmitted via a telecommunication line.

  The comment storage unit 801 is configured using a storage device such as a magnetic hard disk device or a semiconductor storage device. The comment storage unit 801 stores a comment table for each moving image data. The comment table stored in the comment storage unit 801 is as shown in FIG. The comment table includes a plurality of records 64 in which a comment posted to the moving image data is associated with a moving image playback time at the time when the comment is posted.

  The cluster generation unit 802 clusters comments for each moving image data based on the comment table. For example, the cluster generation unit 802 acquires comments posted during a predetermined time width (for example, 1 second, 5 seconds, 10 seconds, etc.) from the comment table. Next, the cluster generation unit 802 clusters the acquired comments based on the similarity of character strings, the similarity of meaning, and the like.

FIG. 21 is a diagram illustrating an example of the clustering process of the cluster generation unit 802. Hereinafter, an example of the clustering process performed by the cluster generation unit 802 will be described with reference to FIG. The clustering process performed by the cluster generation unit 802 is not limited to the following description, and may be performed in any manner as long as comments can be clustered for each moving image data.
Each character string shown in FIG. 21 is a comment posted during a predetermined time of certain moving image data. The cluster generation unit 802 generates a cluster with comments other than non-conforming comments among the posted comments. The non-conforming comment is a comment that does not satisfy the condition as a comment included in the cluster. Non-conforming comments are, for example, comments of a predetermined number of characters (for example, 2) or less, comments of a predetermined number of characters (for example, 20) or more, comments composed of only predetermined characters (for example, “w”), and the like. For example, “www” shown in FIG. 21 is a non-conforming comment.

The cluster generation unit 802 selects the comments as comments included in the same cluster when one character string is all included in the other. For example, “ahoge” and “hoge” are selected as the same cluster and connected by a link because four characters “hoge” are included in “ahoge”. “Foo” and “foobar” are selected as the same cluster and connected by a link because two characters “foo” are included in “foobar”. A group of comments connected by links in this way is a cluster 81 (81-1 to 81-4).
Since “fight” does not include the character string of another comment or is not included in the character string of another comment, “fight” is not selected for the same cluster as other comments. Therefore, the cluster 81-3 including “fight” has only “fight”. Similarly, the cluster 81-4 to which “huge” belongs has only “huge”. Note that the number shown with a colon “:” on the left of each comment indicates the number of the same comment. For example, the comment “hoge” is posted twice during a predetermined time width of the moving image data to be processed.

  Returning to FIG. 20, the description of the input candidate table generation device 800 will be continued. The table generation unit 803 generates an input candidate table based on the result of clustering performed by the cluster generation unit 802. For example, the table generation unit 803 selects a predetermined number (for example, three) of clusters in descending order of the number of comments (hereinafter referred to as “cluster elements”) included in the cluster. The table generation unit 803 selects a comment of the main element of each selected cluster as an input candidate. The main element is a cluster element having the largest number. For example, in the specific example of the cluster shown in FIG. 21, “hogeww” is posted most frequently among the cluster elements of the cluster 81-1. Therefore, in the cluster 81-1, “hogeww” is the main element. In cluster 81-2, “foobar” is the main element. Note that the number shown with a colon “:” on the left of the character string of the main element shown at a position off the circle represents the total number of comments included in the cluster 81 (the total number of cluster elements).

The table generation unit 803 generates a record 67 of the input candidate table by associating the main elements of the selected predetermined number of clusters with the time width of the moving image playback time to be processed. The table generation unit 803 generates an input candidate table of moving image data to be processed by generating the record 67 over the entire moving image playback time of the moving image data to be processed. The table generation unit 803 records the generated input candidate table in the table storage unit 804.
The table storage unit 804 is configured using a storage device such as a magnetic hard disk device or a semiconductor storage device. The table storage unit 804 stores the input candidate table generated by the table generation unit 803.

  In the input candidate table generation device 800 configured as described above, a comment with a large number of posted times is selected as an input candidate for each time duration of the video playback time based on comments posted in the past on the video data. Is possible. Therefore, it is possible to generate an input candidate table that can provide input candidates with high accuracy in the terminal device 100c.

  In addition, a comment of the main element is selected as an input candidate from each cluster, and an input candidate table is generated. For this reason, it is possible to suppress a plurality of similar comments from being included in the record 67 and to increase the diversity of comments of input candidates provided for each time width.

  In addition, by removing the removed comments and generating a cluster, it is possible to prevent comments that are actually significantly different from being included in the same cluster. Therefore, since different comments are divided into different clusters, it is possible to maintain the possibility that each different comment is selected as an input candidate, and to increase the variety of input candidate comments provided for each time span. Become.

<Modification>
The table generation unit 803 selects all comments in the cluster selected in the record 67 within a predetermined time (for example, within 10 seconds) from the time width of the moving image playback time to be processed. It may be configured not to select an input candidate in the moving image playback time record 67. For example, when “888888” is selected as an input candidate in the record 67 of the time from the moving image playback time “0: 00: 00: 00” to “0:00:05”, “0: In the record 67 of the time from “00:05” to “0:00:10” and the record 67 of the time from “0:00:10” to “0:00:15”, the table generating unit 803 displays “ All comments included in the cluster to which “888888” belongs are not selected as input candidates. In this case, the table generation unit 803 selects a main element of another cluster as an input candidate instead of the duplicate comment. For example, the table generation unit 803 may select a main element of a cluster having the fourth largest number of cluster elements as an input candidate.
With this configuration, it is possible to further increase the diversity of input candidate comments provided for each time width.

  When the same comment is posted a plurality of times by the same user in the duration of the video playback time to be processed, the table generation unit 803 may count as one regardless of the number posted by the user. . With this configuration, it is possible to prevent a comment posted many times by the same user from being selected as an input candidate. Therefore, it is possible to prevent selection of comments that are actually posted by only a small number of users as input candidates, and to provide highly accurate input candidates to the user.

  The embodiment of the present invention has been described in detail with reference to the drawings. However, the specific configuration is not limited to this embodiment, and includes designs and the like that do not depart from the gist of the present invention.

DESCRIPTION OF SYMBOLS 1 ... Movie delivery system, 100, 100a, 100b, 100c ... Terminal device, 200 ... Input candidate server, 300 ... Movie delivery server, 400 ... Comment server, 500 ... Network, 101 ... Input part, 102 ... Output part, 103 ... Communication unit 104 ... Movie playback unit 105 ... Comment control unit 106 106c Storage unit 106-1 Coefficient information storage unit 106-2 Parameter storage unit 107 107b 107c Selection unit 20 Screen, 210 ... Video display area, 220 ... Display area for input, 22 ... Comment input field, 23 ... Post button, 24-1 to 24-10 ... Input candidate button, 211 ... Comment character string, 600 ... Coefficient information generation Device, 601 ... comment category storage unit, 602 ... comment storage unit, 603 ... coefficient Information generation unit, 604 ... Coefficient information storage unit, 700 ... Comment classification device, 701 ... Category definition storage unit, 702 ... Comment storage unit, 703 ... Movie genre storage unit, 704 ... Co-occurrence index calculation unit, 705 ... Parameter storage unit , 706 ... Classification unit, 707 ... Comment category storage unit, 800 ... Input candidate table generation device, 801 ... Comment storage unit, 802 ... Cluster generation unit, 803 ... Table generation unit, 804 ... Table storage unit, 81-1, 81 -2, 81-3, 81-4 ... Cluster

Claims (5)

A candidate selection device comprising a selection unit that selects a character string as a user input candidate in accordance with an elapsed time from the start of reproduction of moving image data reproduced on an output device. A coefficient acquisition unit that acquires a coefficient that changes in accordance with an elapsed time from the start of reproduction of the moving image data reproduced on the output device;
A score calculation unit that calculates a score of a character string based on the coefficient acquired by the coefficient acquisition unit;
The candidate selection device according to claim 1, wherein the selection unit selects the character string having a high score as a character string of a user input candidate.   The candidate according to claim 1, wherein the selection unit causes the output device to display an input candidate selected in the past and an input candidate selected at the present time, and does not select an input candidate already selected in a certain past. Selection device. A candidate selection method comprising a selection step of selecting a character string as a user input candidate in accordance with an elapsed time from the start of reproduction of moving image data reproduced on an output device. A computer program for causing a computer to execute a selection step of selecting a character string as a user input candidate in accordance with an elapsed time from the start of reproduction of moving image data reproduced on an output device.

Images