JP2009060567A - Information processing apparatus, method, and program - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To efficiently extract information of names of performers of a program within information included in an electronic program guide (EPG). <P>SOLUTION: An EPG text data extraction part 13 extracts text data of an electronic program guide; a morphological analyzing part 15 performs morphological analysis on text information of the electronic program guide; a pattern comparison part 42 compares morphological analysis results of the morphological analyzing part 15 with a plurality of list patterns of predetermined performer names; and when there is a list pattern of predetermined performer names having matched at least one part or more out of the morphological analysis results on the basis of the comparison results, a performer name extraction means 43 extracts a performer name with the list pattern of the matched predetermined performer name. This invention can be applied to a content management system. <P>COPYRIGHT: (C)2009,JPO&INPIT

Description

  The present invention relates to an information processing apparatus, method, and program, and in particular, an information processing apparatus, method, and program that can efficiently extract information on the names of performers of content from information included in content metadata. About.

  A technique for selecting a program that is a content by using an electronic program guide composed of content metadata called EPG (Electric Program Guide) and for reserving a program selected on the EPG is becoming widespread.

  A technique has been proposed that enables more appropriate information to be extracted reliably and easily using keywords used for automatic recording as information (see Patent Document 1).

  Further, a technique has been proposed for reliably searching for a desired program even when the program name included in the EPG is omitted as the program time elapses (see Patent Document 2).

JP 2006-339947 A JP 2004-134858 A

  However, in the past, when trying to extract information on the name of a performer of a program that is content from content metadata such as EPG, it is possible to find a person name by morphological analysis, but whether it is a person name as a role name, Alternatively, since it is impossible to identify whether the name is a person name as a performer name, when a performer name is simply extracted, a role name and other person names may also be extracted.

  The present invention has been made in view of such a situation, and in particular, among the information included in content metadata such as an electronic program guide (EPG), information on the names of performers of programs that are content can be efficiently obtained. It is to be able to extract.

  An information processing apparatus according to an aspect of the present invention includes: an acquisition unit that acquires content metadata; a morpheme analysis unit that performs morphological analysis on text information included in the content metadata; and a morpheme analysis result of the morpheme analysis unit; A comparison means for comparing a plurality of predetermined performer name enumeration patterns, and a list of predetermined performer names matching at least one of the morphological analysis results based on the comparison result of the comparison means Extracting means for extracting a performer name in a sequence of the predetermined predetermined performer names when there is a pattern.

  From the morpheme analysis result of the morpheme analysis unit, a layout recognition unit for recognizing a layout for each content described can be further included, and the comparison unit recognizes the layout recognition unit. In the layout of the morpheme analysis result of the morpheme analysis means, the information outside the performer name field can be compared with the enumeration pattern of a plurality of predetermined performer names.

  Of the layout of the morpheme analysis result of the morpheme analysis unit recognized by the layout recognition unit and the layout recognition unit that recognizes the layout for each content described from the morpheme analysis result of the morpheme analysis unit, performer name Similarity calculation means for calculating the distance of similarity between the information in the column and the enumeration pattern of a plurality of predetermined performer names, and the morphological analysis based on the similarity distance calculation result of the similarity distance calculation means Among the results, a second extraction means for extracting a performer name with a predetermined enumeration pattern of performer names having a minimum similarity distance can be further included.

  In the enumeration pattern of the predetermined performer name, performer name, symbol, performer name, symbol ..., performer name, symbol, role name, performer name ..., or role name, symbol, performer name, It is possible to include an enumeration pattern of role names ... or performer names, performer names ....

  The content may include a television program, and the metadata may include information related to the television program.

  An information processing method according to one aspect of the present invention includes an acquisition step of acquiring content metadata, a morphological analysis step of analyzing text information included in the content metadata, and a morpheme in the processing of the morpheme analysis step. Comparison result in morpheme analysis step for comparing morphological analysis of text information of comparison step table for comparing analysis result and enumeration pattern of plural predetermined performer names, processing in said morpheme analysis step, and processing in said comparison step Based on the morpheme analysis result, if there is a list of predetermined performer names that match at least one place, the extraction is performed to extract the performer names with the predetermined pattern of the predetermined performer names Steps.

  The program according to one aspect of the present invention includes an acquisition step of acquiring content metadata, a morpheme analysis step of analyzing morpheme of text information included in the content metadata, and a morpheme analysis result in the processing of the morpheme analysis step And a comparison step for comparing a plurality of predetermined performer name enumeration patterns, and a predetermined appearance matching at least one or more of the morphological analysis results based on the comparison result in the processing of the comparison step If there is an enumeration pattern of performer names, the computer is caused to execute a process including an extraction step of extracting performer names with the enumerated pattern of the predetermined performer names.

  The program storage medium of the present invention stores the program according to claim 4.

  In the information processing apparatus, method, and program according to an aspect of the present invention, content metadata is acquired, text information included in the content metadata is analyzed, a morphological analysis result, and a plurality of predetermined appearances If a list pattern of a given performer name that matches at least one of the morphological analysis results is present based on the comparison result, the list of matched performers is displayed. The names of the performers are extracted with the enumeration pattern of names.

  The information processing apparatus of the present invention may be an independent apparatus or a block that performs information processing.

  According to one aspect of the present invention, it is possible to efficiently extract information on the names of performers of content from information included in the metadata of the content.

  Embodiments of the present invention will be described below. The correspondence relationship between the invention described in this specification and the embodiments of the invention is exemplified as follows. This description is intended to confirm that the embodiments supporting the invention described in this specification are described in this specification. Therefore, although there is an embodiment which is described in the embodiment of the invention but is not described here as corresponding to the invention, it means that the embodiment is not It does not mean that it does not correspond to the invention. Conversely, even if an embodiment is described herein as corresponding to an invention, that means that the embodiment does not correspond to an invention other than the invention. Absent.

  Further, this description does not mean all the inventions described in this specification. In other words, this description is for the invention described in the present specification, which is not claimed in this application, that is, for the invention that will be applied for in the future or that will appear and be added by amendment. It does not deny existence.

  That is, an information processing apparatus according to an aspect of the present invention includes an acquisition unit (for example, the EPG acquisition unit 12 or the iEPG acquisition unit 14 in FIG. 1) that acquires content metadata, and text information included in the content metadata. Morphological analysis means (for example, morpheme analysis unit 15 in FIG. 1), and comparison means (for example, FIG. 1) for comparing the morpheme analysis results of the morpheme analysis means and the enumeration patterns of a plurality of predetermined performer names. 1 pattern comparison unit 42) and, based on the comparison result of the comparison means, if there is an enumeration pattern of a given performer name that matches at least one place among the morphological analysis results, the matching predetermined Extraction means (for example, a performer name extraction unit 43 in FIG. 1) for extracting performer names in an enumerated pattern of performer names.

  From the morpheme analysis result of the morpheme analysis unit, a layout recognition unit (for example, the layout recognition unit 20 in FIG. 1) for recognizing a layout for each content described can be further included, and the comparison unit ( For example, the pattern comparison unit 42) in FIG. 1 includes information outside the performer name column and a plurality of predetermined performer names in the layout of the morpheme analysis result of the morpheme analysis unit recognized by the layout recognition unit. It can be made to compare with the enumeration pattern.

  From the morpheme analysis result of the morpheme analysis unit, a layout recognition unit (for example, the layout recognition unit 20 in FIG. 1) that recognizes the layout for each content described, and the morpheme analysis unit recognized by the layout recognition unit Of the layout of the morpheme analysis results, similarity calculation means for calculating the similarity distance between the information in the performer name column and the enumeration pattern of a plurality of predetermined performer names (for example, similarity distance calculation in FIG. 1) Part 33) and based on the similarity distance calculation result of the similarity distance calculation means, performer names are extracted from the morpheme analysis result in a list pattern of predetermined performer names having the minimum similarity distance. Second extraction means (for example, performer name extraction unit 35 in FIG. 1) can be further included.

  An information processing method according to an aspect of the present invention includes an EPG acquisition step (for example, step S2 in FIG. 3) for acquiring content metadata, and a morpheme analysis step (for performing morphological analysis on text information included in the content metadata). For example, step S4) in FIG. 3 and a comparison step (for example, step S53 in FIG. 10) for comparing the morpheme analysis result in the processing of the morpheme analysis step and the enumeration pattern of a plurality of predetermined performer names, Based on the comparison result in the process of the comparison step, if there is an enumeration pattern of a predetermined performer name that matches at least one place among the morphological analysis results, an enumeration of the matched predetermined performer names And an extraction step (for example, step S55 in FIG. 10) for extracting performer names by pattern.

  FIG. 1 is an information processing apparatus showing the configuration of an embodiment to which the present invention is applied.

  The information processing apparatus 1 acquires an EPG (electronic program guide) composed of metadata represented by a network represented by the Internet or the like, or broadcast waves, and the like (programs included in the electronic program guide ( Performer name is extracted as a keyword from the information of the content), and the program corresponding to the performer name selected by the operation unit 5 such as a remote controller including an operation button or a keyboard among the extracted performer names is extracted. Is displayed.

  The receiving unit 11 receives a broadcast wave via the antenna 2 and supplies it to the EPG acquisition unit 12 and the tuner 26. The EPG acquisition unit 12 acquires EPG (electronic program guide) information from the signals supplied from the reception unit 11, and supplies the EPG text data extraction unit 13, the layout recognition unit 20, and the program search unit 25.

  The iEPG acquisition unit 14 accesses the EPG distribution server 4 specified by a predetermined URL (Uniform Resource Locator) or the like via the network 3 represented by the Internet, acquires EPG information, and acquires the EPG text data extraction unit 13. To the layout recognition unit 20 and the program search unit 25.

  The EPG text data extraction unit 13 extracts text data from each of the EPG information supplied from the EPG acquisition unit 12 or the EPG information supplied from the iEPG acquisition unit 14 and supplies the text data to the morpheme analysis unit 15.

  The morpheme analysis unit 15 divides the text data of the EPG information into minimum units of words (hereinafter referred to as words), and information registered in the dictionary storage unit 16 for each word. The morpheme analysis process is executed by specifying the part of speech and the result is stored in the morpheme analysis result buffer 17.

  The layout recognition unit 20 recognizes the layout for each information displayed as EPG based on the EPG information supplied from the EPG acquisition unit 12 or the iEPG acquisition unit 14, and supplies the recognized layout information to the classification extraction unit 21. To do.

  The classification extraction unit 21 recognizes the arrangement of the performer column in which the performer name is described, and based on the layout information supplied from the layout recognition unit 20, the information in the performer column is obtained from the morphological analysis result buffer 17. The information is read out and supplied to the performer column determination unit 24, and information outside the performer column is read from the morphological analysis result buffer 17 and supplied to the performer column determination unit 18. Details of the performer column will be described later.

  The performer margin determination unit 18 extracts a performer name based on a morphological analysis result included in an area other than the layout of the performer column in the information displayed as the EPG, and stores the performer name extraction result storage. Store in the unit 22.

  The pattern extraction unit 41 of the performer margin determination unit 18 sequentially reads one of a plurality of attribute enumeration patterns stored in the pattern storage unit 19 and stores them in the morphological analysis result buffer 17 as the enumeration pattern. The words existing in the area outside the performer column and the corresponding attribute information are extracted and supplied to the pattern comparison unit 42.

  Here, the attribute pattern is an enumeration pattern in which the attribute is a performer name, a role name, a foreigner performer, a Japanese voice, a foreign character name, a reading name, and a group name, for example, as shown in FIG. , First pattern to eighth pattern.

  The first pattern is, for example, “performer name, performer name”, “performer name, performer name”, “performer name / performer name”, “performer name performer name”, “performer name”. / Performer name "and" performer name (new line) performer name ", there is some symbol (including space and new line) between the performer name and performer name, and the performer name is continuous It is a pattern that is enumerated.

  The second pattern is, for example, “performer name (title)” and “performer name → role name”, where the role name is arranged next to the performer name, and some symbol (including space and line feed) is present. There is a pattern in which something is continuously listed.

  Furthermore, the third pattern includes, for example, “title: performer name”, “role name ... performer name”, “role name____performer name”, “role name_performer name”, and “role name / performer”. The name of the performer is arranged next to the role name, and is inserted between some symbols (including spaces and line feeds), and these are continuously enumerated.

  The fourth pattern is, for example, a “performer name (group name)”, and a group name to which the performer belongs is arranged next to the performer name, and these are continuously enumerated. is there.

  Furthermore, the fifth pattern is, for example, “foreign performer ... Japanese voice” and “foreign performer (Japanese voice)”, and the dubbed Japanese name is placed next to the foreign performer name. , It is a pattern in which they are sandwiched between some symbols and they are enumerated continuously. The foreign performer name here is composed of a person name written in katakana or alphabet.

  In addition, the sixth pattern is, for example, “foreign character name = foreign performer (Japanese voice)”, a symbol is arranged next to the foreign character name, and then the foreign character name is arranged. It is a pattern in which dubbed Japanese names are placed in parentheses and are enumerated continuously.

  Furthermore, the seventh pattern is, for example, “foreign performer reading Kana”, in which a reading kana is arranged next to a foreign performer name, and these are continuously enumerated.

  The eighth pattern is, for example, “foreign character name (Yomikana)... Foreign performer (Yomikana), and after the foreign character name, the Yomi name is placed in parentheses, and then some symbol is In this pattern, the names of foreign performers are arranged, followed by reading names in parentheses, and these are continuously enumerated.

  The names of performers in the first to eighth patterns include not only the name of a person as part of speech but also an attribute for identifying a famous person such as an actress name, an actor name, and a singer name. Further, the title includes not only words indicating titles such as “moderator” and “producer” as attributes, but also names of characters appearing in the story.

  The pattern comparison unit 42 extracts the pattern extraction unit 41 on the assumption that the pattern extraction unit 41 is one of the above-described first to eighth patterns stored in the pattern storage unit 19 from the morphological analysis result buffer 17. The enumeration pattern of the attribute that has been performed is compared with the assumed enumeration pattern to determine whether or not they match.

  The performer name extraction unit 43 extracts the information of the performer name with the matched enumeration pattern based on the comparison result of the pattern comparison unit 42 and stores it in the performer name extraction result storage unit 22.

  The performer column determination unit 24 extracts the performer name based on the morphological analysis result included in the region in the performer column from the information displayed as the EPG, and the performer name extraction result storage unit 22 is stored.

  The attribute determination unit 31 determines an attribute for each word supplied from the classification extraction unit 21 and supplies the attribute to the pattern extraction unit 32. The pattern extraction unit 32 extracts an attribute pattern based on the attribute determination result supplied from the attribute determination unit 31 and supplies the attribute pattern to the similarity distance calculation unit 33. The similarity distance calculation unit 33 calculates a similarity distance indicating the similarity between the pattern supplied from the pattern extraction unit 32 and the pattern stored in the pattern storage unit 19, and sequentially supplies the similarity distance to the pattern determination unit 34. To do. The pattern determination unit 34 recognizes the pattern having the smallest similarity distance as the pattern extracted by the pattern extraction unit 32 based on the similarity distance information supplied from the similarity distance calculation unit 33. The extracted pattern is determined, and the determined pattern is supplied to the director name extraction unit 35. The director name extraction unit 35 extracts only the performer name from the words supplied from the classification extraction unit 21 based on the pattern supplied from the pattern determination unit 34 and stores it in the performer name extraction result storage unit 22. Remember me.

  The output unit 23 outputs the performer name stored in the performer name extraction result storage unit 22.

  Next, the performer name extraction process will be described with reference to the flowchart of FIG.

  In step S1, the EPG acquisition unit 12 or the iEPG acquisition unit 14 determines whether or not the operation unit 5 has been operated to instruct the display of the performer name, and the same processing is repeated until it is determined that the operation has been instructed. . For example, when the option tab 101 as shown in FIG. 4 is displayed and the button 116 described as “person name” instructing to display the performer name as a person name is operated, the display of the performer name is instructed. The process proceeds to step S2.

  FIG. 4 shows an example of an image displayed on the display unit 6. A display field 102 for a normal broadcast program selected by the tuner 24 is provided on the left side of the option tab 101. On the tab 101, buttons 111 to 117 are displayed with “HDD information”, “DVD information”, “picture and sound quality setting”, “program recording”, “program description”, “person name”, and “keyword” from the top. It is displayed. The button 111 is operated to display program information recorded in an HDD (Hard Disk Drive) (not shown). The button 112 is operated when displaying program information recorded on a DVD inserted in a DVD (Digital Versatile Disk) drive (not shown). The button 113 is operated when executing image quality setting. The button 114 is operated when program recording is executed. The button 115 is operated when displaying the program description of the program displayed in the display field 102 included in the EPG. The button 116 is operated to display the performer name of the program displayed in the display field 102 included in the EPG as a person name. The button 117 is operated when displaying the keyword of the program displayed in the display field 102 included in the EPG.

  In step S <b> 2, the EPG acquisition unit 12 acquires EPG information of a predetermined program included in the broadcast wave received by the antenna 2 via the reception unit 11 and supplies the EPG information to the EPG text data extraction unit 13 and the layout recognition unit 20. To do. Alternatively, the iEPG acquisition unit 14 accesses the EPG distribution server 4 on the network 3 specified by a predetermined URL, acquires EPG information of a predetermined program, and sends it to the EPG text data extraction unit 13 and the layout recognition unit 20. Supply.

  In step S <b> 3, the EPG text data extraction unit 13 extracts text data from the supplied EPG information and supplies the text data to the morpheme analysis unit 15.

  In step S4, the morpheme analysis unit 15 divides the text data of the supplied EPG information into words based on the information stored in the dictionary storage unit 16, specifies the part of speech of each word, and determines the morpheme. The result is stored in the analysis result buffer 17. In the morphological analysis of the morphological analysis unit 15 using the dictionary storage unit 16, it is possible to specify a person name as a part of speech if it is a person name among the nouns as part of speech. As for the famous actor name, famous actress name, famous singer name, etc., it is also possible to use attributes that specify that the person name is the actor name, actress name, and singer name, respectively. Therefore, the morphological analysis unit 15 not only specifies the part of speech in the grammar for each word, but also classifies whether it is a person name, a product name, or a place name when it is a noun. Furthermore, when it is a person name, it is classified as an attribute including whether it is an actor name, an actress name, or a singer name.

  In step S <b> 5, the layout recognition unit 20 recognizes the layout based on the EPG display information supplied from the EPG acquisition unit 12 or iEPG acquisition unit 14, and supplies the recognition result to the classification extraction unit 21. For example, when the EPG information is displayed as shown in FIG. 5, the layout recognition unit 20 recognizes the layout as follows.

  In the case of FIG. 5, the layout recognizing unit 20 recognizes the area Z1 displayed at the top as “I fall in love with the octopus's eyes-You are destined ...” as the title display column. In addition, the layout recognition unit 20 said in the middle “The angelic woman whom the main character Osamu Yamada (Inaro Goro) met was an octopus ... My colleague Takeuchi Takeuchi (Nanori Obayashi) ...” Is recognized as a story explanation column. Furthermore, the layout recognition unit 20 is the bottom stage "Performer Goro Inada (Osamu Yamada) Tomoko Murashita (Miyuki Oguchi) Yuri Sugawara (Yumi Sugawara) MEGMU (Shoko Yoyogi) Screenplay Mother Director Yoshishige Mikami Yuta Hashimoto and others Music Mitsubishiki People Theme Song: “Koi no Hana” Mirai Hamada (Rhythm Zombie) Program Description A couple of couples were born three years ago. The area Z3 displayed as “The work can be done”, which is the evaluation standard of the man who has not changed from the past, has been greatly cleared. Since the EPG display screen includes regions having substantially the same structure even if the arrangement order of the regions as indicated by the regions Z1 to Z3 described above is different, the layout recognition unit 20 includes these regions. Recognize (estimate) from attributes.

  Furthermore, in this area Z3, the layout recognizing unit 20 particularly describes a description column of “Performer Goro Inada (Osamu Yamada) Tomoko Murashita (Miyuki Oguchi) Yuri Sugawara (Yumi Sugawara) MEGUMU (Shoko Yoyogi) (described later). Region Z3 ′) is recognized as the performer column. That is, in the case of FIG. 5, the layout recognizing unit 20 recognizes an area that is a lump of words including a description “performer” as a performer column.

  In step S <b> 6, the classification extraction unit 21 extracts words in the performer column from the morpheme analysis result buffer 17 based on the layout information and supplies the extracted words to the performer column determination unit 24.

  In step S <b> 7, the classification extraction unit 21 extracts words outside the performer column from the morphological analysis result buffer 17 based on the layout information, and supplies the extracted word to the performer margin determination unit 18.

  In step S <b> 8, the performer column determination unit 24 executes a performer column determination process, extracts the performer's words from the words in the performer column, and stores them in the performer name extraction result storage unit 22.

  Here, with reference to the flowchart of FIG.

  In step S31, the attribute determination unit 31 determines whether or not all the words supplied from the separation / extraction unit 21 are words registered in attributes such as actors and actresses, and sets the determination result as a pattern. This is supplied to the extraction unit 32. That is, the attribute determination unit 31 uses the name of the attribute registered as the name of an actual actor or actress, not the name of an attribute indicating a nonexistent person, such as a role name, for the word in the performer column that has been supplied. It is determined whether or not there is.

  In step S <b> 32, the pattern extraction unit 32 generates a determination pattern from a pattern indicating the presence / absence of registration as a personal name based on the determination result supplied from the attribute determination unit 31. That is, for example, as shown in the upper part of FIG. 7, when the performer column is indicated by the region Z3 ′, the words “Inada Goro”, “Murashita Tomoko”, “MEGUMU” are actors, actresses, etc. If it is recognized as a non-existing person name, other than "Osamu Yamada", "Miyuki Oguchi", "Yuri Sugawara", "Yumi Sugawara", "Syoyo Yoyogi" As shown in the lower part of FIG. 7, the patterns shown are “Yes”, “No”, “Yes”, “No”, “No”, “No”, “Yes”, “No”. From this determination result, since the performer name is a real person name, the determination result of “Yes” indicates that it is the name of the performer, and the determination result of “No” indicates that it is a role name. Therefore, the pattern extraction unit 32 generates “out”, “comb”, “out”, “comb”, “comb”, “comb”, “out”, and “comb” as determination patterns, and is similar. To the sex distance calculator 33. Here, “Out” indicates the name of the performer, and “Combination” indicates the name of the role.

  In step S33, the similarity distance calculation unit 33 initializes a counter i (not shown) for identifying a pattern to 1.

  In step S <b> 34, the similarity distance calculation unit 33 compares the determination pattern with the i-th pattern stored in the pattern storage unit 19 and counts the number of correct / incorrect. That is, for example, as shown in the lower part of FIG. 7, the determination pattern is “out”, “comb”, “out”, “comb”, “comb”, “comb”, “out”, “comb”. In this case, when the counter i = 1, as shown in the upper part of FIG. 8, the pattern P1, which is the first pattern, is “out”, “out”, “out”, “out”, “out”, “out”. ”,“ Out ”,“ Out ”, and“ Out ”in the determination pattern are all counted as positive, and“ Combination ”is counted as“ No ”. In this case, positive is 5 and negative is 3. .

  Further, when the counter i = 2, as shown in the middle part of FIG. 8, the pattern P2, which is the second pattern, is “out”, “combination”, “out”, “combination”, “out”, “combination”. , “Out” and “role”, positive is 7 and negative is 1.

  Further, when the counter i = 3, as shown in the lower part of FIG. 8, the pattern P3 which is the third pattern is “combination”, “out”, “combination”, “out”, “combination”, “out”. , “Position” and “Out”, so positive is 1 and negative is 7.

  In step S <b> 35, the similarity distance calculation unit 33 calculates the similarity distance between the determination pattern and the i-th pattern based on the correct / incorrect count result, and supplies the similarity distance to the pattern determination unit 34. More specifically, for example, the similarity distance calculation unit 33, for example, when the counter i is 1, the determination pattern includes eight elements, and among them, there are three incorrect elements. The calculation unit 33 calculates the similarity distance as 37.5% (= 3/8 × 100). The similarity distance is closer to 0% as the similarity is more similar. Similarly, when the counter i is 2, the similarity distance calculation unit 33 calculates the similarity distance as 12.5% (= 1/8 × 100). Further, when the counter i is 3, the similarity distance calculation unit 33 calculates the similarity distance as 87.5% (= 7/8 × 100). Since the fourth to eighth patterns are the same process, the description thereof is omitted.

  The similarity distance is not limited to the above-described definition, and any other distance may be used as long as a quantitative value of the similarity can be shown. For example, an edit graph algorithm may be used.

  Here, the edit graph algorithm is, for example, as shown in FIG. That is, in FIG. 9, the counter i is 1, 2, and 3 in order from the left, and the i-th pattern and the determination pattern are arranged at intervals of 1 on the X axis and the Y axis, respectively. Then, the patterns of the portions where the X coordinate and the Y coordinate coincide with each other from the origin are compared, and if they match, a diagonal line is drawn, and if they do not match, a line is drawn in the horizontal direction one by one in the X direction and the Y direction. With the same processing, the total distance when the number of diagonal lines is 0 and the number of horizontal and vertical lines is 1 is the similarity distance according to the edit graph algorithm.

  Therefore, when the counter i is 1, as shown in the left part of FIG. 9, the numbers corresponding to the “combination” in the discrimination pattern are respectively rejected, so in (0, 0) to (1, 1) Are diagonal lines in (1, 1) to (2, 2), one straight line in the X direction and one in the Y direction, and diagonal lines in (2, 2) to (3, 3), (3 , 3) to (6, 6), there are two straight lines in the X direction and Y direction, respectively, (6, 6) to (7, 7) are diagonal lines, and (7, 7) to (8 , 8), there are two straight lines, one in the X direction and one in the Y direction. As a result, when the counter i is 1, the similarity distance using the edit graph algorithm is 10. Similarly, when the counter i is 2, as shown in the central part of FIG. 9, in (4, 4) to (5, 5), all are diagonal lines except for two straight lines. The sex distance is 2. Further, similarly, when the counter i is 3, as shown in the right part of FIG. 9, all of them are two straight lines except for the diagonal lines in (4, 4) to (5, 5). The similarity distance is 14.

  In step S36, the similarity distance calculation unit 33 determines whether or not the similarity distance between all patterns stored in the pattern storage unit 19 and the determination pattern has been calculated. If the distance has not been calculated, the process proceeds to step S37, and after the counter i is incremented by 1, the process returns to step S34. That is, the processes in steps S34 to S37 are repeated until similarity distances with all patterns are calculated. In step S37, when the similarity distances to all patterns are obtained, in step S38, the pattern determination unit 34 determines that the minimum value of the calculated similarity distances is greater than a predetermined threshold value. It is determined whether or not it is small. That is, it is determined whether or not the minimum similarity distance is a reliable value. If it is determined in step S38 that the minimum similarity distance is smaller than the predetermined threshold value, that is, it is a reliable value, the process proceeds to step S39.

  In step S <b> 39, the pattern determination unit 34 determines the pattern having the minimum similarity distance as a pattern used for the performer name extraction, and supplies the pattern information to the performer name extraction unit 35. The performer name extraction unit 35 extracts a performer name from the words supplied from the classification extraction unit 21 based on the pattern supplied from the pattern determination unit 34. That is, for example, when the area Z3 ′ in FIG. 7 has been supplied as the area in the performer column, since the second pattern has the minimum similarity distance in the case of the first pattern to the third pattern, the performer The name extraction unit 35 sequentially extracts “Inada Goro”, “Murashita Tomoko”, “Yuri Sugawara”, and “MEGUMU” as performer names, and stores them in the performer name extraction result storage unit 22 in step S40.

  On the other hand, if it is determined in step S38 that the minimum similarity distance is greater than the predetermined value and is not a reliable value, in step S41, the pattern determination unit 34 displays all the names according to the first pattern. It is extracted as a performer, and is stored in the performer name extraction result storage unit 22 in step S40.

  By performing the above process, the performer field is identified from the EPG display screen, the performer name pattern is determined, and the performer name is extracted. A pattern in which performer names are arranged in an area that is likely to be regularly arranged can be determined using the similarity distance. As a result, it becomes possible to improve the recognition accuracy of the arrangement pattern of performer names. When the similarity distance is low in reliability, it is possible to suppress omission of performer name extraction by extracting all person names in the performer column as performer names.

  Now, the description returns to the flowchart of FIG.

  In step S8, when the in-performer column determination process ends, in step S9, the performer out-of-line determination unit 18 performs a performer out-of-line determination process, and extracts performer words from words other than the performer column. It is stored in the performer name extraction result storage unit 22.

  Here, with reference to the flowchart of FIG. 10, the performer margin determination process will be described.

  In step S51, the pattern extraction unit 41 initializes a counter i (not shown) for identifying a pattern to 1.

  In step S52, assuming that the pattern extraction unit 41 is an enumeration pattern of attributes corresponding to the i-th pattern, the pattern extraction unit 41 sequentially extracts from the morpheme analysis results other than the performer column supplied from the classification extraction unit 21, This is supplied to the pattern comparison unit 42. At this time, the pattern extraction unit 41 notifies the pattern comparison unit 42 that extraction is performed with the i-th pattern.

  In step S53, the pattern comparison unit 42 sequentially extracts the attribute of the word extracted by the pattern extraction unit 41 and the attribute of the i-th pattern from the morpheme analysis result other than the performer column supplied from the classification extraction unit 21. Compare with the enumeration pattern.

  That is, for example, when text data as shown in FIG. 11 is extracted from the EPG text data extraction unit 13, the following comparison is made. In Fig. 11, “This time, newly-married actor Satoshi Tomomi and veteran actress Yumiko Fujikawa challenged Mr. Millionaire. Attention is the first time for Mr. Mirinea who finally won the center seat. A face-to-face confrontation is realized, and the dream of 10 million yen is “Give a gorgeous bicycle to a friend to go cycling.” The challenge of Tori begins.

  However, the quiz is also over there, and the topic is to the secret story of the shooting of “Third year C-Gumi Kin-kun teacher”. “Teacher” Tetsuya Takekawa went to the 14th question. The lifeline was used at a good timing, and I managed to come up to the 14th question. It is a sports problem that attacks Tomi. Can Torimi really exceed the teacher? ? In addition, Fujikawa ’s 10 million yen dream, “I came to stare at you,” was “Fund for the Tokyo Theater Performance”. Performer Moderator: Minato also guest challenger: Toriko Saito Yumiko Fujikawa et al. "

  Among these, for example, “Newlywed acting group actor Tori Tomomi, Liquor veteran actress Yumiko Fujikawa” was analyzed by morphological analysis as “newly married”, “acting”, “group”, “actor”, “ It is broken down into “辰 吾”, “,” “sake-go”, “no”, “veteran”, “actress” “·”, and “Yuko Fujikawa”. In the case of i = 1, that is, in the case of the first pattern, the pattern extraction unit 41 assumes “performer name”, “symbol”, and “performer name”, and sequentially sets “newly-married”, Extract "act", "sect", then extract "act", "sect", "actor", then extract "sect", "actor", "." Are extracted as a pattern and supplied to the pattern comparison unit 42.

  The pattern comparison unit 42 compares the attribute enumeration pattern for the three words supplied from the pattern extraction unit 41 with the attribute enumeration pattern in the first pattern.

  In step S54, the pattern comparison unit 42 determines whether the enumeration patterns match. That is, for example, in the case of “the newlywed acting actor Tori Tomomi, the liquor veteran actress Yumi Fujikawa”, the names are only “Tomi Tomi” and “Yuko Fujikawa”, “Tomi Tomi” and “Yuko Fujikawa” However, even if the names of actors and actresses are recognized, the patterns of “performer name”, “symbol”, and “performer name” are not established, so it is determined that they do not match, and the process proceeds to step S55. move on.

  In step S55, the pattern comparison unit 42 determines whether or not all the patterns stored in the pattern storage unit 19 have been tried. If all the patterns have not been tried, in step S56, the pattern extraction unit 41 determines whether the counter i Is incremented by 1, and the process returns to step S52.

  On the other hand, for example, in the latter part of the text data, “Performer”, “Moderator”, “:”, “Mitanmo” for the part of “Performer Moderator: Mintanmo Guest Challenger: Toriko Saito Yumiko etc.” ”,“ Guest ”,“ Challenge ”,“: ”,“ Torimi Satoshi ”,“ Fujikawa Yumiko ”, and“ Other ”are extracted, but when the counter i = 2, the pattern extraction unit 41 Assuming “title”, “symbol”, and “performer name”, first “performer”, “moderator”, “:” are extracted, then “moderator”, “:”, “Mitan” ”,“: ”,“ Mintanmo ”, and“ Guest ”are extracted, and three words are sequentially extracted and supplied to the pattern comparison unit 42.

  Then, in this case, the extracted attributes of “Moderator”, “:”, “Mitanmo”, “Moderator” is the title, “:” is a symbol, and “Mitanmo” is a celebrity. If registered, in step S54, the pattern comparison unit 42 considers that the pattern matches the third pattern, and the process proceeds to step S55.

  In step S55, the pattern comparison unit 42 instructs the performer name extraction unit 43 to extract performer names with the matched pattern. Therefore, the performer name extraction unit 43 extracts the performer name based on the third pattern of “title”, “symbol”, and “performer name”, and performs the performer name extraction result storage unit 22. Remember me. Then, the process proceeds to step S56.

  That is, in the latter part of the text data in FIG. 11, the word placed after “symbol” is the name of the performer. Therefore, the attribute is arranged from the arrangement of “moderator”, “:”, “Mitanmo”. Since the arrangement pattern of “Role name”, “Symbol”, and “Performer name” is “Mitanmo”, “Torimi” is the same as “Challenge”, “:”, and “Tomi”. Are extracted as performers and stored in the performer name extraction result storage unit 22.

  If it is determined in step S56 that all patterns have been tried, that is, in this case, the counter i indicating the number of enumerated patterns is up to 8, so if the counter i is greater than 8, the pattern comparison is performed in step S58. The unit 42 determines whether or not there is a matching pattern in all patterns. In this case, since the third pattern matches, the process of step S59 is skipped.

  On the other hand, if the pattern does not match any pattern in step S58, the pattern comparison unit 42 instructs the performer name extraction unit 43 to extract the performer name in the first pattern in step S59. In other words, if it does not match any of the patterns, the person name will not be extracted as a performer. Read out.

  For example, as shown in FIG. 12, when text data is extracted from EPG data, “Suzuko Kuroishi: Kyoko Tanigawa Haruki Kuzuyama: Saori Oda Saori Nakajima: Mao Obayashi Tamio Ogai: Masahiko Higashimura Michiru : Sachiko Makimoto Maki: Eiko Oike Takanobu Ogauchi: Masanori Iwai Toru Kuroishi: Hiroki Amano Michizo Kuzuyama: Isao Nakazume The first pattern “Performer Name”, “Symbol”, “Performer Name” Or the second pattern “title”, “symbol”, or “performer name”. That is, in the case of a drama or the like, the name of the person can be recognized as both “title” and “performer name”. For example, if it is actually the third pattern, at least “Haruki Kuzuyama” will be recognized if “Oda Keisuke” is recognized as the actor name in the morphological analysis process. , “:”, And “Kesuke Oda” are recognized as “title”, “symbol”, and “performer name”, so that at least one is regarded as a match in the third pattern. So, “Kyoko Tanigawa”, “Kesuke Oda”, “Mao Obayashi”, “Mashiko Higashimura”, “Sachiko”, “Eiko Oike”, “Massunori Iwai”, “Hiroki Amano”, and “Isao Nakazume” performers It will be extracted as a name.

  In the latter part of FIG. 12, if all names can only be recognized as person names, no pattern is recognized in step S58. Therefore, in step S59, all person names are “ Suzuko Kuroishi, Kyoko Tanigawa, Haruki Kuzuyama, Keisuke Oda, Saori Shimanaka, Mao Obayashi, Tamio Oga, Masahiko Higashimura, Michiru Fukakura, Sachiko, "Maki Enomoto", "Eiko Oike", "Takanobu Ogai", "Masunori Iwai", "Toru Kuroishi", "Hiroki Amano", "Michizo Kazuyama", "Isao Nakazume" are selected as performers It will be. In this case, the performer names may include an error, but at least all performer names are displayed.

  Through the processing as described above, an enumeration pattern of attributes in which the performers are displayed is set in advance, and the morphological analysis result is compared with the enumeration pattern of the set attributes, based on the matching attribute enumeration pattern. By extracting performers, performers can be extracted efficiently.

  Now, the description returns to the flowchart of FIG.

  In step S9, when the performer margin determination process is completed, in step S10, the output unit 23 reads the name of the performer stored in the performer name extraction result storage unit 22 and displays it on the display unit 6.

  By this processing, the display unit 6 displays the name of the performer as the name of the person on a screen as shown in FIG. 13, for example. In FIG. 13, a personal name display field 121 is provided on the right side of the normal broadcast program display field 102, and buttons 131 to 133 operated when selecting the extracted personal name correspond to the extracted performer names. Is provided. In FIG. 13, a button 131 is provided for the performer name of “President Manager”, a button 132 is provided for the performer name of “Beki Man”, and the appearance of “Emi Kanina” A button 133 is provided for the name of the person.

  In step S <b> 11, the program search unit 25 determines whether or not the operation unit 5 is operated and any of the buttons 131 to 133 is operated to select a performer name. For example, in FIG. 13, when the button 131 is operated by the operation unit 5 and the keyword “president's manager” is selected, the program search unit 25 supplies from the EPG acquisition unit 12 or the iEPG acquisition unit 14 in step S12. Based on the received EPG information, a program is searched with the keyword “President Manager” (search for a program including the keyword “President Director” in the program information of EPG information). As shown in FIG. 14, it is displayed on the display unit 6. If no selection is made in step S11, the processes in steps S12 and S13 are skipped.

  In FIG. 14, a selected keyword tab 151 is provided to show the selected keyword. In FIG. 14, “president president”, which is the selected keyword, is shown. Below that, a search result display field 152 is provided, and a program searched with the selected keyword is displayed. In FIG. 14, “Tomorrow 1:05 AM movie theater“ Beyond the table ”” is displayed at the top, and “2:30 AM Howbiz Extra # 201” is displayed at the second, In the third row, “9:30 PM Thursday Western Theater“ Indian Game ”is displayed. In the fourth row,“ 0:00 AM Indie Movie Festival ~ Independent Production ”is displayed. Displays “050 AM movie theater“ My Home ””, the sixth row displays “2:30 AM Billy talks about himself”, and the seventh row displays “11: 00 PM The movie “Tomb and Marriage” (free broadcast) ”is displayed, and the program name and broadcast time of each is displayed. For example, a recording reservation may be made by selecting the display column of these programs. Below the search result display field, a button 153 with “return” displayed on the right side is provided. The button 153 is operated to end the display by the selected keyword tab 151 and return it to the original state. Further, on the left side of the button 153, a button 154 displayed as “option” is provided. The button 154 is operated when an option operation is executed.

  In step S14, the program search unit 25 determines whether the operation unit 5 has been operated to instruct word registration. For example, the button 154 is operated by the operation unit 5 and the option operation dialog box 171 is displayed as shown in FIG. 15, and the button 181 displayed as “Register word” on the option operation dialog box 171 is pressed. When word registration is instructed, in step S15, the program search unit 25 causes the performer name extraction result storage unit 22 to store the currently selected word “president president” in the performer name extraction result storage unit 22. By this processing, the name of the performer instructed to register the phrase is always displayed in the personal name display column 121 even if it is not included in the EPG data.

  The option operation dialog box 171 shown in FIG. 15 is provided with a button 181 that is operated when “register word” is instructed and a button 182 that is operated when canceling the option operation.

  On the other hand, when the word registration is not instructed in step S14, the process of step S15 is skipped.

  In step S16, it is determined whether or not termination has been instructed. If termination has not been instructed, the process returns to step S11. If termination has been instructed, the process ends.

  According to the above processing, based on the information included in the electronic program guide (EPG), the area of the performer column is specified from the layout information, and the performer name is regularly specified for the information in the performer column. Since there is a high possibility of being placed, the pattern is analyzed by placing information that does not include symbols such as performer name and role name, and the performer name is extracted based on the analyzed pattern, so with higher accuracy It is possible to extract performer names.

  Also, for information outside the performer column, the performer name may not be regularly arranged than in the performer column, so in addition to the performer name and role name, based on the arrangement of symbols, Since the pattern is analyzed and the performer name is extracted based on the analyzed pattern, the performer name can be extracted with higher accuracy.

  As a result, it is possible to extract the performer name with high accuracy and efficiency by switching the method of extracting the performer name by distinguishing between the performer field and the performer field.

  In the above description, the example in which the content metadata is EPG has been described. However, any metadata other than EPG may be used as long as it is metadata as content additional information. For example, ECG (Electronic Contents Guide) Or the like.

  Furthermore, the example in which the content is a television program has been described above. However, the content may be other than a television program as long as the content includes metadata, for example, a video downloaded via a network. It may be image content or music content, or may be moving image content or music content stored in a data storage medium such as a DVD (Digital Versatile Disc) or BD (Blu-Ray Disc).

  According to one aspect of the present invention, it is possible to efficiently extract information on the names of performers of content from information included in the metadata of the content.

  Incidentally, the series of text processing described above can be executed by hardware, but can also be executed by software. When a series of processing is executed by software, a program constituting the software may execute various functions by installing a computer incorporated in dedicated hardware or various programs. For example, it is installed from a recording medium in a general-purpose personal computer or the like.

  FIG. 16 shows a configuration example of a general-purpose personal computer. This personal computer incorporates a CPU (Central Processing Unit) 1001. An input / output interface 1005 is connected to the CPU 1001 via the bus 1004. A ROM (Read Only Memory) 1002 and a RAM (Random Access Memory) 1003 are connected to the bus 1004.

  An input / output interface 1005 includes an input unit 1006 including an input device such as a keyboard and a mouse for a user to input an operation command, an output unit 1007 for outputting a processing operation screen and an image of a processing result to a display device, a program and various data. A storage unit 1008 including a hard disk drive to be stored, a LAN (Local Area Network) adapter, and the like, and a communication unit 1009 for performing communication processing via a network represented by the Internet are connected. Also, a magnetic disk (including a flexible disk), an optical disk (including a CD-ROM (compact disc-read only memory), a DVD (digital versatile disc)), a magneto-optical disk (including an MD (mini disc)), or a semiconductor A drive 1010 for reading / writing data from / to a removable medium 1011 such as a memory is connected.

  The CPU 1001 is read from a program stored in the ROM 1002 or a removable medium 1011 such as a magnetic disk, an optical disk, a magneto-optical disk, or a semiconductor memory, installed in the storage unit 1008, and loaded from the storage unit 1008 to the RAM 1003. Various processes are executed according to the program. The RAM 1003 also appropriately stores data necessary for the CPU 1001 to execute various processes.

  In this specification, the step of describing the program recorded on the recording medium is not limited to the processing performed in time series in the order described, but of course, it is not necessarily performed in time series. Or the process performed separately is included.

It is a block diagram which shows the structural example of the information processing apparatus to which this invention is applied. It is a figure explaining the enumeration pattern of a performer name. It is a flowchart explaining a performer name extraction process. It is a figure which shows the example of the display image at the time of performing a performer name extraction process. It is a figure explaining a performer name extraction process. It is a flowchart explaining the determination process in a performer column. It is a figure explaining the determination process in a performer column. It is a figure explaining the calculation method of similarity distance. It is a figure explaining the calculation method of the similarity distance using an edit graph algorithm. It is a flowchart explaining a performer margin determination process. It is a figure explaining a performer margin determination process. It is a figure explaining a performer margin determination process. It is a figure explaining the example of a display of the display screen of a person name. It is a figure explaining the example of a display of a display screen when a person name is selected. It is a figure explaining the example of a display of a display screen when instruct | indicating word registration. And FIG. 11 is a diagram illustrating a configuration example of a personal computer.

Explanation of symbols

  DESCRIPTION OF SYMBOLS 1 Information processing apparatus, 13 EPG text data extraction part, 15 Morphological analysis part, 16 Dictionary storage part, 17 Morphological analysis result buffer, 18 Performer extracurricular judgment part, 19 Pattern storage part, 20 Layout recognition part, 21 Sorting extraction part, 22 Performer name extraction storage unit, 23 output unit, 24 performer in-column determination unit, 25 program search unit, 26 tuner, 31 attribute determination unit, 32 pattern extraction unit, 33 similarity distance calculation unit, 34 pattern determination unit, 35 performer name extraction unit, 41 pattern extraction unit, 42 pattern comparison unit, 43 performer name extraction unit

Claims (8)

An acquisition means for acquiring content metadata;
Morphological analysis means for morphological analysis of text information included in the metadata of the content;
Comparison means for comparing the morphological analysis results of the morpheme analysis means and the enumeration patterns of a plurality of predetermined performer names;
Based on the comparison result of the comparison means, if there is an enumeration pattern of a predetermined performer name that matches in at least one of the morphological analysis results, an appearance is made with the enumeration pattern of the corresponding predetermined performer name An information processing apparatus comprising: first extraction means for extracting a person's name. From the morpheme analysis result of the morpheme analysis means, further comprising a layout recognition means for recognizing the layout for each content described,
The comparison unit compares information outside the performer name column with a plurality of predetermined performer name enumeration patterns in the layout of the morpheme analysis result of the morpheme analysis unit recognized by the layout recognition unit. The information processing apparatus according to 1. From the morpheme analysis result of the morpheme analysis means, layout recognition means for recognizing the layout for each content described;
Of the layouts of the morpheme analysis results of the morpheme analysis means recognized by the layout recognition means, the similarity for calculating the distance of the similarity between the information in the performer name field and the enumeration pattern of a plurality of predetermined performer names Degree calculation means,
Based on the similarity distance calculation result of the similarity distance calculation means, a second extraction means for extracting a performer name from the morpheme analysis result with an enumeration pattern of predetermined performer names having a minimum similarity distance The information processing apparatus according to claim 1. The enumeration pattern of the predetermined performer name is performer name, symbol, performer name, symbol, .., performer name, symbol, role name, performer name ..., or role name, symbol, performer name, role name. The information processing apparatus according to claim 1, further comprising: an enumeration pattern of performer names, performer names, and so on. The information processing apparatus according to claim 1, wherein the content includes a television program, and the metadata includes information related to the television program. An acquisition step of acquiring content metadata;
A morpheme analysis step for morphological analysis of text information included in the metadata of the content;
A comparison step for comparing a morphological analysis result in the processing of the morpheme analysis step and an enumeration pattern of a plurality of predetermined performer names;
Based on the comparison result in the process of the comparison step, if there is an enumeration pattern of a predetermined performer name that matches at least one place among the morphological analysis results, an enumeration of the matched predetermined performer names A first extraction step of extracting a performer name by a pattern. An acquisition step of acquiring content metadata;
A morpheme analysis step for morphological analysis of text information included in the metadata of the content;
A comparison step for comparing a morphological analysis result in the processing of the morpheme analysis step and an enumeration pattern of a plurality of predetermined performer names;
Based on the comparison result in the process of the comparison step, if there is an enumeration pattern of a predetermined performer name that matches at least one place among the morphological analysis results, an enumeration of the matched predetermined performer names A program for causing a computer to execute a process including a first extraction step of extracting a performer name by a pattern.   A program storage medium in which the program according to claim 7 is stored.

Images