JP2009043189A - Information processor, information processing method, and program - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To allow a user to confirm description of content from a word selected based on a value calculated with consideration given to deflection of an appearance frequency and an appearance position of the word and to designate a reproduction position. <P>SOLUTION: A movement window is set for every area of a predetermined number of words constituting a string of words obtained by performing morphological analysis on subtitle data. The whole of the words included in each of the movement windows is considered as one document, and as to each word, a TF-IDF (Term Frequency-Inverted Document Frequency) value is calculated. For example, TF-IDF values of the respective words are compared mutually, and the words are selected sequentially as characteristic words in the descending order of the TF-IDF value. When a predetermined characteristic word is selected from the words displayed on a screen of program detailed information by the user, reproduction of a recorded program is started from the position wherein a sentence including the selected characteristic word is displayed as subtitles. This Invention is applicable to a digital recording apparatus. <P>COPYRIGHT: (C)2009,JPO&INPIT

Description

  The present invention relates to an information processing apparatus, an information processing method, and a program, and in particular, a user confirms the contents of content from a word selected based on a value calculated in consideration of the number of appearances of words and the deviation of appearance positions. The present invention relates to an information processing apparatus, an information processing method, and a program that can specify a reproduction position.

  Various methods have been proposed as a method for designating the playback position of content such as a recorded program. For example, Patent Literature 1 discloses a technique for searching for the same keyword as the keyword input by the user from the subtitles of the content, and outputting a scene in which a sentence including the same keyword is displayed as the subtitle as a scene that the user pays attention to. Has been.

JP 2006-129122 A JP 2003-16091 A

  The above-described method is effective when it is clear what kind of scene the user wants to see, but cannot be used when it is not clear.

  Also, since the user does not know how important the entered keyword is in the content, a scene that is displayed as subtitles but is not closely related to the keyword is played back. There is also such a thing.

  Therefore, the device can select a word representing the feature of the content by analyzing subtitles, present the selected word, and the user can select the keyword to input the keyword. There are some, but it is not appropriate to simply select words with high frequency of appearance. Even words that often appear in subtitles in any scene and are not considered to represent the characteristics of a specific scene are presented to the user and used for scene selection.

  Therefore, from the viewpoint of use in scene selection, it is preferable to present words that do not appear so much in the subtitles of other scenes but concentrate on a specific scene.

  The present invention has been made in view of such a situation, and the user confirms the content of the content from the word selected based on the value calculated in consideration of the appearance frequency and the bias of the appearance position, and plays back the content. The position can be specified.

  An information processing apparatus according to an aspect of the present invention analyzes a subtitle of a content, acquires a word constituting the subtitle, and considers an appearance frequency and a bias of an appearance position for each of the words acquired by the analyzing unit. Calculating means for calculating the weighting factor, selection means for selecting a predetermined number of feature words representing the features of the content based on the weighting factor calculated by the calculating means, and the feature selected by the selecting means Replaying means for displaying the word and reproducing the content from the position where the sentence including the feature word selected from the displayed ones is displayed as a subtitle.

  The calculation unit generates a column in which the words acquired by the analysis unit are arranged in the display order, sets a detection window for each range of a predetermined number of words constituting the column, and is included in each detection window. TF is the number of occurrences of the word of interest within the detection window containing the word, DF is the number of detection windows in which the word of interest appears among all the detection windows, and the weighting factor is TF It is possible to provide a calculation means for calculating the weighting coefficient so that the weighting coefficient increases in accordance with the increase in DF and decreases in accordance with the increase in DF.

  The selection unit compares the weighting factor calculated by the calculation unit, and a predetermined number of words are obtained in order from the word with the largest maximum value obtained in the same content, or in the same content. A word having a maximum value greater than a threshold value can be selected as the feature word.

  The reproduction means can display the feature word selected by the selection means together with information on the content acquired from the EPG data.

  An information processing method or program according to one aspect of the present invention analyzes subtitles of content, acquires words constituting the subtitles, and calculates a weighting factor for each of the acquired words in consideration of appearance frequency and appearance position bias. Then, based on the calculated weighting factor, a predetermined number of feature words representing the feature of the content are selected, the selected feature word is displayed, and the sentence including the feature word selected from the displayed ones is subtitled The content is reproduced from the position displayed as.

  In one aspect of the present invention, subtitles of content are analyzed, words constituting the subtitles are acquired, and a weighting factor is calculated for each of the acquired words in consideration of the appearance frequency and the bias of the appearance position. Further, a predetermined number of feature words representing the features of the content are selected based on the calculated weighting factor, the selected feature words are displayed, and the selected feature words are included from the displayed ones. The content is reproduced from the position where the text is displayed as a subtitle.

  According to one aspect of the present invention, the user confirms the content of the content from the word selected based on the value calculated in consideration of the number of appearances of the word and the bias of the appearance position, and designates the reproduction position. be able to.

  Embodiments of the present invention will be described below. Correspondences between the constituent elements of the present invention and the embodiments described in the specification or the drawings are exemplified as follows. This description is intended to confirm that the embodiments supporting the present invention are described in the specification or the drawings. Therefore, even if there is an embodiment which is described in the specification or the drawings but is not described here as an embodiment corresponding to the constituent elements of the present invention, that is not the case. It does not mean that the form does not correspond to the constituent requirements. On the contrary, even if an embodiment is described herein as corresponding to the invention, this does not mean that the embodiment does not correspond to other than the configuration requirements. .

  The information processing apparatus according to one aspect of the present invention (for example, the information processing apparatus 1 in FIG. 1) analyzes content subtitles and acquires analysis words (for example, the morphological analysis unit 33 in FIG. 3) that acquires words constituting the subtitles. And calculating means (for example, TF-IDF value calculating unit 34 in FIG. 3) for calculating the weighting factor in consideration of the appearance frequency and the bias of the appearance position for each of the words acquired by the analyzing means, and the calculating means Based on the calculated weighting factor, a selection unit (for example, the feature word selection unit 35 in FIG. 3) that selects a predetermined number of feature words representing the feature of the content, and the feature word selected by the selection unit Reproduction means (for example, the content reproduction unit 37 in FIG. 3) reproduces the content from a position where the sentence including the feature word selected from the displayed ones is displayed as a caption.

  An information processing method or program according to one aspect of the present invention analyzes subtitles of content, acquires words constituting the subtitles, and calculates a weighting factor for each of the acquired words in consideration of appearance frequency and appearance position bias. Then, based on the calculated weighting factor, a predetermined number of feature words representing the feature of the content are selected, the selected feature word is displayed, and the sentence including the feature word selected from the displayed ones is subtitled Including the step of reproducing the content from the position displayed as (for example, step S16 in FIG. 12).

  Hereinafter, embodiments of the present invention will be described with reference to the drawings.

  FIG. 1 is a diagram showing an information processing apparatus 1 according to an embodiment of the present invention.

  The information processing apparatus 1 is a digital recording device having a built-in hard disk. A television receiver 2 is connected to the information processing apparatus 1 via a cable.

  The information processing device 1 is a television receiver 2 that receives video (audio) of a program (content) provided by BS (Broadcasting Satellite) / CS (Communications Satellite) digital broadcasting, terrestrial digital broadcasting, or broadcasting performed via the Internet. Or output program data, record program data on the hard disk, and record. That is, the information processing apparatus 1 is supplied with a signal from an antenna (not shown). The information processing apparatus 1 reproduces a recorded program in response to an instruction from the user, and causes the television receiver 2 to output the video and audio of the program.

  Further, the information processing apparatus 1 acquires and manages EPG data distributed by a broadcasting station or the like via broadcast waves or the Internet. The EPG data includes information on the program such as the program title, broadcast date and time, performers, and overview for each program.

  The information processing apparatus 1 has a function of extracting words that are considered to represent the characteristics of each program by analyzing subtitle data such as closed caption data included in recorded program data. As a word that is considered to represent the characteristics of the program, there is a bias in the appearance position in the caption data of a specific program, not a general word that is also included in the caption data of other programs, and Words with high appearance frequency are extracted as described later.

  The information processing apparatus 1 displays the extracted word on a title list screen on which recorded program information is displayed in a list or a program detailed information screen displayed when a predetermined program is selected from the title list. It is displayed with information such as the title and presented to the user. When a predetermined word is selected by the user from the presented items and playback is instructed, the information processing apparatus 1 starts the recorded program from the position where the sentence including the word selected by the user is displayed as a subtitle. Start playback.

  By presenting these words for each recorded program, the user can identify the topics covered in each program from words that are thought to better represent the characteristics of each program. You can check the summary. In the caption data of a specific program, words whose appearance position is biased and whose appearance frequency is high are characterized by the program characteristics compared to general words that often appear in the caption data of other programs. It can be said that it is a more appropriate word.

  In addition, the user can select a word related to the scene he is interested in from such words, and can start playing the program from a position where a sentence including the selected word is displayed as a subtitle. .

  A process of the information processing apparatus 1 that extracts a characteristic word and presents it to the user and controls the reproduction of the recorded program will be described later with reference to a flowchart.

  FIG. 2 is a block diagram illustrating a hardware configuration example of the information processing apparatus 1.

  A CPU (Central Processing Unit) 11 executes various processes according to a program recorded in a ROM (Read Only Memory) 12 or a recording unit 19. A RAM (Random Access Memory) 13 appropriately stores programs executed by the CPU 11 and data. The CPU 11, ROM 12, and RAM 13 are connected to each other by a bus 14.

  An input / output interface 15 is also connected to the CPU 11 via the bus 14. The input / output interface 15 is connected to a receiving unit 16, an input unit 17, an output unit 18, a recording unit 19, a communication unit 20, and a drive 21.

  The receiving unit 16 receives and demodulates the broadcast wave signal from the antenna 16A, and acquires MPEG-TS (Moving Picture Experts Group-Transport Stream). The receiving unit 16 acquires program data and EPG data to be recorded from the MPEG-TS, and outputs the acquired data to the recording unit 19 via the input / output interface 15.

  The input unit 17 receives a signal from the remote controller, and outputs information representing the contents of the user's operation to the CPU 11 via the input / output interface 15 and the bus 14.

  The output unit 18 decodes the data of the program instructed to be reproduced, and displays the video of the program on the television receiver 2 based on the obtained video signal.

  The recording unit 19 includes, for example, a hard disk, and records various data such as a program executed by the CPU 11, program data supplied from the receiving unit 16 via the input / output interface 15, and EPG data. Recorded program data is associated with words extracted from subtitle data and information about the program acquired from EPG data.

  The communication unit 20 communicates with the server, and acquires program data distributed by broadcasting performed via the Internet and EPG data distributed by the server. The communication unit 20 outputs the acquired data to the recording unit 19 via the input / output interface 15 and records it.

  When a removable medium 22 such as a magnetic disk, an optical disk, a magneto-optical disk, or a semiconductor memory is mounted, the drive 21 drives them to acquire programs and data recorded therein. The acquired program and data are transferred to the recording unit 19 and recorded as necessary.

  FIG. 3 is a block diagram illustrating a functional configuration example of the information processing apparatus 1. At least a part of the functional units shown in FIG. 3 is realized by executing a predetermined program by the CPU 11 of FIG.

  As shown in FIG. 3, in the information processing apparatus 1, a content recording unit 31, a caption data acquisition unit 32, a morpheme analysis unit 33, a TF-IDF (Term Frequency-Inverted Document Frequency) value calculation unit 34, a feature word selection The unit 35, the content information display unit 36, and the content reproduction unit 37 are realized.

  The content recording unit 31 is realized in the recording unit 19 of FIG. 2 and manages program data in association with information related to the program acquired from EPG data. Also, the content recording unit 31 extracts words representing program features from the caption data, and manages the extracted words in association with program data when the extracted words are supplied from the feature word selection unit 35. To do.

  The caption data acquisition unit 32 pays attention to each program whose data is recorded in the content recording unit 31 and reads out the caption data included in the data of the program of interest from the content recording unit 31. The caption data acquisition unit 32 outputs the caption data read from the content recording unit 31 to the morpheme analysis unit 33.

  The morpheme analysis unit 33 performs a morphological analysis on the caption data supplied from the caption data acquisition unit 32 to divide sentences displayed as captions into words, and a sequence of words in which the words are arranged in display order. Is generated. The morpheme analyzer 33 outputs the generated word string to the TF-IDF value calculator 34.

  The TF-IDF value calculation unit 34 calculates the TF-IDF value of each word constituting the sequence supplied from the morpheme analysis unit 33. A large TF-IDF value is calculated for a word that is not a general word but has a biased appearance position and a high appearance frequency.

  Here, calculation of the TF-IDF value will be described.

  FIG. 4 is a diagram illustrating an example of setting a moving window by the TF-IDF value calculation unit 34.

  The wide horizontal dotted line shown in the upper part of FIG. 4 represents each sentence displayed as subtitles, and the horizontal dotted line shown in the lower part is generated by the morphological analysis unit 33 by performing morphological analysis. Represents a sequence of words. While the recorded program is being played back, the television receiver 2 displays the subtitles and words on the left side.

  In the TF-IDF value calculation unit 34, a moving window (detection window) is set for each range of a predetermined number of words such as 50, 100, etc. constituting the word string, and each included in each moving window. The TF-IDF value is calculated for this word. That is, the entire word included in the moving window is regarded as one document, and the TF-IDF value is calculated.

In the example of FIG. 4 are set moving window W ₁ is a sequence of words that are generated by the morphological analysis is performed, then, is moved window W ₂ is shifted in the order of displaying the word to only a range of a predetermined number of Is set.

  For example, when one moving window is set to a range of 100 words, and each moving window is set by shifting the range by 10 words, 1 of a sequence of words arranged in display order. The first moving window is set in the range from the 10th word to the 100th word, and then the second moving window is set in the range from the 10th word to the 110th word. Then, the third moving window is set in the range from the 20th word to the 120th word.

In the TF-IDF value calculation unit 34, paying attention to each word included in each moving window, the TF-IDF value is calculated by the following equation (1).

  In equation (1), TF is the number of appearances of the word of interest within the moving window containing that word, and N is the number of all moving windows. DF is the number of moving windows in which a noticed word appears in all moving windows.

  In other words, the TF-IDF value is a TF that shows how often the word you are interested in appears in a certain moving window, and the number of words that you are interested in only in a few moving windows. This is a value obtained by multiplying IDF, which expresses whether or not there is a numerical value, and is a numerical index that takes into account the bias of the appearance position and the appearance frequency.

  Information representing the appearance position of each word, such as the TF-IDF value obtained by the TF-IDF value calculation unit 34 and the display time based on the beginning of the program, is output to the feature word selection unit 35.

  The feature word selection unit 35 selects a feature word that is a word considered to represent the feature of the program of interest based on the TF-IDF value supplied from the TF-IDF value calculation unit 34. For example, the feature word selection unit 35 compares the TF-IDF values of the respective words, and selects a predetermined number of words such as 10 as feature words in order from the largest TF-IDF value.

  5 to 7 are diagrams illustrating examples of TF-IDF values obtained when attention is paid to a certain program.

  FIG. 5 is a diagram illustrating an example of the TF-IDF value calculated for “Soba” that is a word included in the caption of the program of interest.

In the example of FIG. 5, 40 TF-IDF values are calculated for “Soba” appearing at time t ₁ , and 35 TF-IDF values are calculated for “Soba” appearing at time t _2. For “soba” appearing at t ₃ , 55 TF-IDF values are calculated. Similarly, 40 TF-IDF values are calculated for “Soba” appearing at time t _4, and 20 TF-IDF values are calculated for “Soba” appearing at time t ₅ . The maximum TF-IDF value is 55 calculated for “soba” appearing at time t ₃ .

Depending on how the moving window is set, a plurality of moving windows including the same “soba” appearing at, for example, time t ₁ are set, and different TF-IDF values are calculated for the “soba”. For convenience of explanation, here, only one representative value is shown as the TF-IDF value of the word appearing at each time. The same applies to FIGS. 6 and 7.

  FIG. 6 is a diagram illustrating an example of the TF-IDF value calculated for “service”.

In the example of FIG. 6, 10 TF-IDF values are calculated for “service” that appears at time t ₁₁ , and 15 TF-IDF values are calculated for “service” that appears at time t _12. Twenty TF-IDF values are calculated for the “service” appearing at t ₁₃ . Similarly, 15 TF-IDF values are calculated for “service” appearing at time t _14, and 10 TF-IDF values are calculated for “service” appearing at time t ₁₅ . The maximum TF-IDF value is 20 calculated for the “service” that appears at time t ₁₃ .

  FIG. 7 is a diagram illustrating an example of TF-IDF values calculated for “pigs”.

In the example of FIG. 7, 65 TF-IDF values are calculated for “pigs” that appear at time t ₂₁ , and 60 TF-IDF values are calculated for “pigs” that appear at time t _22. TF-IDF value of 40 in the "pig" that appear in the t ₂₃ is calculated. Similarly, 55 TF-IDF values are calculated for “pigs” appearing at time t _24, and 50 TF-IDF values are calculated for “pigs” appearing at time t ₂₅ . The maximum TF-IDF value is 65 calculated for the “pig” appearing at time t ₂₁ .

  In the feature word selection unit 35, 55 is the maximum TF-IDF value for "Soba", 20 is the maximum TF-IDF value for "Service", and the maximum TF-IDF for "Pig" The maximum TF-IDF values for each word including the value 65 are compared, and a predetermined number of feature words are selected in descending order of the maximum TF-IDF value.

  The feature word selected by the feature word selection unit 35 and the information on the appearance position where the maximum TF-IDF value of each feature word is calculated are output to the content recording unit 31 and recorded in association with the program data. The It should be noted that a predetermined number of words are not selected in order from the largest TF-IDF value, but a word having a maximum TF-IDF value larger than a threshold value is selected as a feature word. Also good.

  When the user is instructed to display the title list and information indicating that is supplied from the input unit 17, the content information display unit 36 reads information about the recorded program from the content recording unit 31, and outputs it. 18 is controlled to display the title list screen on the television receiver 2. The information read from the content recording unit 31 includes information such as the title of a recorded program, a representative image obtained by capturing a video at a predetermined position, and a feature word.

  FIG. 8 is a diagram illustrating an example of a title list screen.

  In the example of FIG. 8, the titles of three recorded programs “Today's menu“ Tomato is amazing ””, “Travel, good mood“ Ibaraki, Fukushima Oishi ... ”,“ Sake Brewery Visit-Tohoku ” Is displayed.

The image P ₁ displayed on the left side of the title of “Today is Tomato is amazing” is a representative picture of this program. “Negi”, “Bacon”, “Wings” displayed below the title “First” and “Simmer” are feature words selected from the subtitles of the “Today is Tomato is Great” program.

The image P ₂ displayed on the left side of the title “Travel, Good Feeling“ Ibaraki, Fukushima Oishi ... ”is a representative picture of this program, and“ Soba ”,“ “Pig”, “Esthetic”, and “Collagen” are feature words selected from the subtitles of the program “Journey, Good Feeling“ Ibaraki, Oishi Fukushima ... ”.

  FIGS. 5 to 7 show examples of TF-IDF values calculated when words included in the subtitles of the program “Journey, Good mood“ Ibaraki, Fukushima Oishi ... ”are shown. Based on the TF-IDF values shown in FIGS. 5 to 7, words such as “soba” and “pig” are selected as feature words.

The image P ₃ displayed on the left side of the title of “Sake Brewery -Tohoku-” is a representative picture of this program, and “rice”, “water”, and “sake brewery” displayed below the title are: It is a feature word selected from the subtitles of the program “Sake Brewery -Tohoku-”.

  Since the feature words selected from the subtitles are displayed in this way, the user can check the outline of each program from the feature words and select the program to be viewed. Only feature words for which the maximum TF-IDF value exceeds the threshold value may be displayed on the title list screen.

  FIG. 9 is a diagram illustrating an example of a program detail information screen.

  For example, when the title “Travel, good mood“ Ibaraki, Fukushima Oishi ... ”is selected from the screen shown in FIG. 8 and a predetermined operation is performed, the screen shown in FIG. Is displayed.

In the example of FIG. 9, the same image as the image P ₂ is displayed at the upper center of the screen, and below that, the title “Travel, good mood“ Ibaraki, Fukushima delicious food tour ”,“ Soba ”,“ Pig ” In addition to those displayed on the title list screen of “esthetic” and “collagen”, characteristic words of “vegetable” and “angko” are displayed. The characteristic words "Soba", "Pig", "Esthetic", and "Collagen" have a maximum TF-IDF value that exceeds the threshold value, and are particularly words that are considered to represent the characteristics of the program. It is displayed in a conspicuous color compared to other characteristic words of “vegetable” and “angkou”. A feature word whose maximum TF-IDF value exceeds the threshold value may be highlighted by another method, for example, displayed as a large character.

  A cursor C is displayed under “Soba”. The user moves the cursor C by operating a button provided on the remote controller and performs a determination operation. From the appearance position of the characteristic word to which the cursor C is assigned at that time, “travel, good mood” Ibaraki, You can start playing the “Fukushima Delicious Food Tour” program.

  In addition to the title, various information acquired from the EPG data such as information on the performers of the program and an outline may be further displayed on the program detailed information screen.

  Returning to the description of FIG. 3, the content playback unit 37 selects a predetermined feature word from the program detailed information screen as shown in FIG. 9, and starts playback of the recorded program from the appearance position of the feature word. Is read from the content recording unit 31 and the screen of the program is displayed on the television receiver 2 by controlling the output unit 18 based on the read data. Let

  FIG. 10 is a diagram illustrating an example of a program screen displayed by the content reproduction unit 37 when the determination operation is performed in the state of FIG. 9 in which the feature word “soba” is selected.

When the determination operation is performed in the state shown in FIG. 9 in which the feature word “soba” is selected, as shown in FIG. 10, from the position where the sentence including “soba” is displayed as subtitles, “travel, good mood” The program “Ibaraki, Fukushima Delicious Food Tour” starts playing. In the lower part of the screen of FIG. 10, a subtitle of “Soba of new soba, is this soba set ...” is displayed. The screen in FIG. 10 is a screen at a position specified at time t ₃ (FIG. 5), which is the appearance position of “Soba” for which 55 is calculated as the maximum TF-IDF value.

  In this way, the user can start playing the program from the appearance position of a word that is considered to appropriately represent the characteristics of the program.

  Next, processing of the information processing apparatus 1 having the above configuration will be described.

  First, the processing of the information processing apparatus 1 for selecting feature words will be described with reference to the flowchart of FIG. This process is performed for each recorded program.

  In step S <b> 1, the subtitle data acquisition unit 32 reads out and acquires subtitle data included in the data of the program of interest from the content recording unit 31. The caption data acquisition unit 32 outputs the acquired caption data to the morphological analysis unit 33.

  In step S <b> 2, the morpheme analysis unit 33 performs morpheme analysis on the caption data supplied from the caption data acquisition unit 32, and generates a word string in which the respective words constituting the caption are arranged in the display order. The morpheme analyzer 33 outputs the generated word string to the TF-IDF value calculator 34.

  In step S3, the TF-IDF value calculation unit 34 calculates the TF-IDF value of each word constituting the sequence supplied from the morpheme analysis unit 33 as described above. The TF-IDF value calculation unit 34 outputs the calculated TF-IDF value and information indicating the appearance position of each word to the feature word selection unit 35.

  In step S4, the feature word selection unit 35 compares the TF-IDF values of the respective words, and selects a predetermined number of words as the feature words in order from the largest TF-IDF value. The feature word selection unit 35 outputs the selected feature word and the information on the appearance position where the maximum TF-IDF value of each feature word is calculated to the content recording unit 31.

  In step S5, the content recording unit 31 records the feature word and the appearance position information supplied from the feature word selection unit 35 in association with the program data, and ends the process.

  Next, processing of the information processing apparatus 1 that reproduces a recorded program will be described with reference to the flowchart of FIG. This process is started, for example, when the user instructs to display the title list.

  In step S <b> 11, the content information display unit 36 reads out feature words from the content recording unit 31 together with information related to programs such as the titles of recorded programs and representative images.

  In step S <b> 12, the content information display unit 36 causes the television receiver 2 to display a title list screen based on the information read from the content recording unit 31. On the title list screen, as shown in FIG. 8, representative images, titles, feature words, and the like are displayed in a list for each recorded program.

  In step S13, the content information display unit 36 determines whether or not a predetermined program is selected from the programs displayed in the title list and an instruction to display the detailed information screen of the program is given. Wait until it is determined that it has been done.

  If it is determined in step S13 that an instruction to display a detailed information screen has been given, in step S14, the content information display unit 36 displays a program detailed information screen including a list of feature words as shown in FIG. Display.

  In step S15, the content reproduction unit 37 determines whether one feature word is selected from the program detail information screen, and whether or not the user has instructed to start playing the program from the appearance position of the feature word, Wait until it is determined that it is instructed.

  When one feature word is selected in step S15 and it is determined that the user has instructed to start playing the program from the appearance position of the feature word, in step S16, the content playback unit 37 selects the selected feature word. Is read from the content recording unit 31 and reproduced, from the appearance position after the appearance position where the maximum TF-IDF value of the selected feature word is calculated. On the television receiver 2, a program screen as shown in FIG. 10 is displayed. When the reproduction of the recorded program ends, the process ends.

  In the above, the case where the recorded program information is displayed and reproduced has been described. However, not only the recorded program but also other contents including subtitle data such as a broadcast program and a movie recorded in a package. Also when the above information is displayed, the feature word selected based on the TF-IDF value as described above may be presented to the user.

  In the above, a moving window is set for the caption of one content, and the TF-IDF value is calculated by regarding the entire word included in the moving window as one document. The TF-IDF value may be calculated by regarding the entire caption as one document.

  When the TF-IDF value is calculated in this way, a large value is shown as a TF-IDF value for a word that appears biased when looking at a plurality of contents as a whole and has a high appearance frequency. Will be calculated. Similar to the case where the TF-IDF value is calculated by setting the moving window, the TF-IDF value of each word is compared, and a predetermined number of words are selected as feature words in order from the largest TF-IDF value. And may be presented to the user.

  FIG. 13 is a diagram illustrating an example of another title list screen.

  In the example of FIG. 13, the "continuous drama" Hello "Episode 1", "continuous drama" Hello "Episode 2", "continuous drama" Hello "Episode 3" of three of the recorded program title display Has been.

"Drama series" Hello "Episode 1" image P ₁₁ that is displayed on the left side of the title is a typical image of this program, it is displayed on the lower side of the title "horse", "Hokkaido", "Ramen "is, the calculation of" continuous drama "Hello" TF-IDF values the entire subtitle of the program is regarded as one of the documents of the first episode "is performed, the feature words that have been selected on the basis of the calculated TF-IDF value It is.

"Drama series" Hello "Episode 2" image P ₁₂ that is displayed on the left side of the title is a typical image of this program, it is displayed on the lower side of the title "America", "airplane", "Horse Racing "is, the calculation of" continuous drama "Hello" the entire subtitle of the program the is regarded as one of the documents TF-IDF value of the second story "is performed, the feature words that have been selected on the basis of the calculated TF-IDF value It is.

"Drama series" Hello "Episode 3" image P ₁₃ that is displayed on the left side of the title is a typical image of this program, are displayed on the lower side of the title "Japan", "Tokyo", "return home "is, the calculation of" continuous drama "Hello" the entire subtitle of the program the is regarded as one of the documents TF-IDF value of episode 3 "is performed, the feature words that have been selected on the basis of the calculated TF-IDF value It is.

  When a predetermined program is selected from the screen as shown in FIG. 13, a program detail information screen including feature words and the like is displayed. When a feature word is selected from those displayed on the program detailed information screen and playback is instructed, the playback of the content is started from the appearance position of the feature word as described above.

  The feature word selected based on the TF-IDF value is presented, so that the user can confirm the content of each content. If a word with a high frequency of appearance is simply selected as a feature word, the name of the performer is often selected as a feature word when subtitles for each broadcast episode of the series are targeted. Thus, when a feature word is selected based on the TF-IDF value, a word that is included only in the subtitles of each broadcast time, not a word that is included in all the subtitles of each broadcast time Will be presented.

  The series of processes described above can be executed by hardware or can be executed by software. When a series of processing is executed by software, a program constituting the software executes various functions by installing a computer incorporated in dedicated hardware or various programs. It can be installed on a general-purpose personal computer or the like.

  The installed program to be executed is, for example, a magnetic disk (including a flexible disk), an optical disk (CD-ROM (Compact Disc-Read Only Memory), DVD (Digital Versatile Disc), etc.), a magneto-optical disk, or a semiconductor memory. It is recorded on the removable medium 22 shown in FIG. 2 which is a package medium comprising the above, or is provided via a wired or wireless transmission medium such as a local area network, the Internet, or digital satellite broadcasting. The program can be installed in advance in the ROM 12 or the recording unit 19.

  The program executed by the computer may be a program that is processed in time series in the order described in this specification, or in parallel or at a necessary timing such as when a call is made. It may be a program for processing.

  The embodiments of the present invention are not limited to the above-described embodiments, and various modifications can be made without departing from the scope of the present invention.

It is a figure which shows the example of the information processing apparatus which concerns on one Embodiment of this invention. It is a block diagram which shows the hardware structural example of information processing apparatus. It is a block diagram which shows the function structural example of information processing apparatus. It is a figure which shows the example of the setting of a movement window. It is a figure which shows the example of a TF-IDF value. It is a figure which shows the other example of TF-IDF value. It is a figure which shows the further another example of TF-IDF value. It is a figure which shows the example of the screen of a title list. It is a figure which shows the example of the screen of program detailed information. It is a figure which shows the example of the screen of a program. It is a flowchart explaining the feature word selection process of information processing apparatus. It is a flowchart explaining the program reproduction | regeneration processing of information processing apparatus. It is a figure which shows the example of the other screen of a title list.

Explanation of symbols

  DESCRIPTION OF SYMBOLS 1 Information processing apparatus, 2 Television receivers, 31 Content recording part, 32 Subtitle data acquisition part, 33 Morphological analysis part, 34 TF-IDF value calculation part, 35 Feature word selection part, 36 Content information display part, 37 Content reproduction | regeneration Part

Claims (6)

Analysis means for analyzing the subtitles of the content and obtaining words constituting the subtitles;
For each word acquired by the analysis means, a calculation means for calculating a weighting factor considering the number of appearances and the bias of the appearance position;
Selection means for selecting a predetermined number of feature words representing features of the content based on the weighting factor calculated by the calculation means;
An information processing apparatus comprising: a reproduction unit that displays the feature word selected by the selection unit, and reproduces the content from a position where a sentence including the feature word selected from the displayed ones is displayed as a caption. The calculation unit generates a column in which the words acquired by the analysis unit are arranged in the display order, sets a detection window for each range of a predetermined number of words constituting the column, and is included in each detection window. TF is the number of occurrences of the word of interest within the detection window containing the word, DF is the number of detection windows in which the word of interest appears among all the detection windows, and the weighting factor is TF The information processing apparatus according to claim 1, wherein the information processing apparatus calculates a weighting factor so as to increase as the DF increases and decrease as the DF increases. The selection means compares the weighting factors calculated by the calculation means, and obtains a predetermined number of words in order from the word with the largest maximum value obtained in the same content or in the same content. The information processing apparatus according to claim 1, wherein a word whose maximum value is larger than a threshold value is selected as the feature word. The information processing apparatus according to claim 1, wherein the reproduction unit displays the feature word selected by the selection unit together with information on content acquired from EPG data. Analyze the subtitles in the content, get the words that make up the subtitles,
For each acquired word, calculate a weighting factor that takes into account the number of appearances and the bias of the appearance position,
Based on the calculated weighting factor, select a predetermined number of feature words representing the features of the content,
An information processing method comprising: displaying the selected feature word, and reproducing the content from a position where a sentence including the feature word selected from the displayed ones is displayed as a caption. Analyze the subtitles in the content, get the words that make up the subtitles,
For each acquired word, calculate a weighting factor that takes into account the number of appearances and the bias of the appearance position,
Based on the calculated weighting factor, select a predetermined number of feature words representing the features of the content,
A program for displaying a selected feature word and causing a computer to execute a process including a step of reproducing the content from a position where a sentence including the feature word selected from the displayed ones is displayed as a caption.

Images