JP2006277091A - Index data generation device, data retrieval device, and program - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a means for acquiring a retrieval result whose quality is much higher when retrieving a specific part in a document. <P>SOLUTION: A specifying part 1012 specifies a section in a document attached with notes. A first retrieving part 1013 extracts a keyword included in the document attached with notes. An index data generating part 1014 generates index data showing a position where the extracted keyword is included in the document. At that time, when notes are attached to a section including a character string matched with the keyword, the index data generating part 1014 makes the data specifying the notes correspond to the index data. A second retrieval part 1015 retrieves the section in the document including the keyword and the notes attached to the section by using the keyword inputted by the user as a retrieval key based on the generated index data, and makes a display part 103 display the retrieval result. <P>COPYRIGHT: (C)2007,JPO&INPIT

Description

  The present invention relates to a document retrieval technique.

Documents may be annotated by the document reader. The places marked with notes are generally more important in the document compared to other places. Therefore, if a point with a note is picked up and displayed together with the note, it is convenient for a reader of the document to easily grasp the important point. For example, Patent Document 1 discloses a technique that enables such document display.
Japanese Patent Laid-Open No. 11-219245

In addition, when a note is added to a document by a reader, if the location in the same document or other document related to the location where the note was added is automatically displayed, the reader with the note will be interested in it. Conveniently, you can easily get relevant information about what you have. For example, Patent Document 2 discloses a technique that enables such document display.
JP 2000-10981 A

  In general, keywords that are frequently used among keywords included in a sentence are extracted, and a search index is generated based on the extracted keywords. However, in many cases, the entire document includes general terms that are not important in the document but appear frequently. As a result, a search result using an index generated by a statistical method is not necessarily high quality.

  By the way, as described above, the portion of the document that is annotated by the reader is generally more important than the other portions. Therefore, it is conceivable to pay attention to the part with the note in the search. The techniques disclosed in Patent Documents 1 and 2 above are techniques that focus on the part to which a note is attached, but are not techniques that increase the efficiency of search.

  In view of the above-described circumstances, the present invention provides a means for enabling acquisition of a higher-quality search result when searching for a specific part in a document by paying attention to a part with a note in the document. The purpose is to do.

  In order to solve the above-described problem, the present invention stores document data indicating one document and one or more annotation data indicating each of notes associated with any partial document constituting the one document. Index data indicating a position where a phrase that matches one phrase included in the partial document associated with the note indicated by any one of the storage means and the one or more annotation data is included in the one document is generated. An index data generation device comprising index data generation means is provided.

  According to the index data generating apparatus having such a configuration, index data that enables efficient search is generated based on the keywords included in the part of the document that is annotated.

  In a preferred aspect, the index data generation device comprises a specifying means for specifying a partial document with one note from the one document, and the storage means stores the note data indicating the one note as the note data. You may comprise so that it may memorize | store as note data matched with the partial document specified by the specific means.

  In this case, the specifying unit may perform the specifying based on a spatial positional relationship when the document indicated by the document data and the note indicated by the note data are displayed on the same plane.

  According to the index data generation apparatus having such a configuration, index data that enables efficient search is generated even when document data and annotation data in which the correspondence between the annotation and the partial document with the annotation is unknown are used. Is done.

  In another preferred embodiment, the index data generation device comprises conversion means for converting a phrase expressed in one language into a translated word expressed in another language having the same meaning as the phrase, and generating the index data The means may be configured to associate a translation obtained from the one word / phrase by the conversion means with the search data.

  According to the index data generating apparatus having such a configuration, index data that can be used even when a keyword in a language different from the language of the document is used and that enables efficient search is generated.

  In another preferable aspect, the storage unit of the index data generation device stores keyword data indicating a keyword in advance, and the index data generation unit is configured to use the keyword as the one phrase. Also good.

  According to the index data generating apparatus having such a configuration, index data is generated without going through a process such as statistically extracting keywords.

  In another preferred aspect, the index data generation means of the index data generation device has a phrase that matches one phrase included in a note indicated by any one of the one or more annotation data, as the one or more annotations. Annotation index data may be configured to indicate positions included in the annotations indicated by any of the data.

  According to the index data generating apparatus having such a configuration, index data that enables efficient search is generated based on keywords included in a note attached to a document.

  In another preferred embodiment, the index data generation device includes display means for displaying a document indicated by the document data, and handwriting data generation means for generating handwriting data indicating handwriting according to a user's writing operation. The storage unit may be configured to store the handwriting data generated by the handwriting data generation unit as the annotation data.

  In another preferred embodiment, the index data generation device includes a display unit that displays a document indicated by the document data, a handwriting data generation unit that generates handwriting data indicating handwriting according to a user's writing operation, Character recognition processing means for performing character recognition processing on the handwriting data generated by the handwriting data generation means, and generating text data indicating the recognized character string, and the storage means is generated by the character recognition processing means. The text data may be stored as the annotation data.

  According to the index data generating apparatus having such a configuration, even when a user writes a note using a pointing device or the like, index data that enables efficient search is generated based on the note.

  In another preferred embodiment, the index data generation device optically reads an image from a document described in type and one or more notes written on the document by hand, and indicates the read image. Image reading means for generating image data, image data indicating a document described in the type from the image data generated by the image reading means, and one or more images each indicating one or more notes written in handwriting Image data separating means for generating data, and character recognition processing for performing text recognition processing on the image data indicating the document generated by the image data separating means and generating text data indicating the contents of the one document And the storage means stores text data indicating the content of the one document generated by the character recognition processing means. And one or more pieces of image data each indicating one or more notes written by handwriting generated by the image data separation means are stored as the one or more pieces of note data. Good.

  In another preferred embodiment, the index data generation device optically reads an image from a document described in type and one or more notes written on the document by hand, and indicates the read image. Image reading means for generating image data, image data indicating a document described in the type from the image data generated by the image reading means, and one or more images each indicating one or more notes written in handwriting Image data separation means for generating data, and character recognition processing is performed on the image data indicating the document generated by the image data separation means to generate text data indicating the contents of the one document, and the image Character recognition processing is performed on each of the one or more image data each indicating the one or more notes generated by the data separation means. Character recognition processing means for generating one or more text data indicating the contents of the note above, and the storage means stores the text data indicating the contents of the one document generated by the character recognition processing means in the document In addition to storing as data, one or more text data indicating the contents of the one or more annotations generated by the character recognition processing means may be stored as the one or more annotation data.

  According to the index data generating apparatus having such a configuration, even when the user adds a handwritten note to the printed matter, the index data that enables efficient search is generated based on the note.

  In addition, the present invention relates to document data indicating one document, one or more note data each indicating a note associated with one of the partial documents constituting the one document, and one or more keywords. Based on the index data, the storage unit stores index data indicating a position where the phrase matching the keyword is included in the one document, the input unit receiving keyword data indicating the one keyword, and the one document. A search means for searching for one or more partial documents including the one keyword from among a plurality of partial documents constituting the document, and note data indicating a note associated with the partial documents searched by the search means And a data search device characterized by comprising an output means.

  According to the data search device having such a configuration, it is possible to obtain an output of a note attached to a partial document including the keyword by inputting the keyword.

  Further, in a preferred aspect, the data search device includes a conversion unit that converts a phrase expressed in one language into a translation expressed in another language having the same meaning as the phrase, and the search unit includes: Instead of the one keyword, the search may be performed by using a translation obtained from the one keyword by the conversion means.

  According to the data search device having such a configuration, it is possible to obtain an output of a note attached to a partial document in a language different from the language of the keyword by inputting the keyword.

  Further, the present invention provides at least one first document data representing a first document expressed in one language and one or more notes each associated with any partial document constituting the first document. First note data, index data indicating a position in the first document that includes a phrase that matches the keyword with respect to each of the one or more keywords, and another having the same meaning as the first document Second document data indicating a second document expressed in a language, and one or more second note data each indicating a note associated with any partial document constituting the second document Storage means for storing; input means for receiving keyword data indicating one keyword; and one or more including the one keyword among a plurality of partial documents constituting the first document based on the index data Partial document A search means for searching, note data indicating a note associated with the partial document searched by the search means, and a partial document constituting the second document and searched for by the search means There is provided a data search device comprising output means for outputting note data indicating a note associated with a corresponding partial document.

  According to the data search device having such a configuration, by inputting a keyword, in addition to a note attached to a partial document including the keyword, an output of a note attached to another partial document corresponding to the partial document can be obtained. it can.

  The present invention also provides a program that causes a computer to execute the same processing as the processing performed by the index data generation device or the data search device. According to such a program, the above-described index data generation device or data search device can be realized using a general-purpose computer.

[1. First Embodiment]
Hereinafter, as a first embodiment of the present invention, after a user who is a reader attaches a note to a text document prepared in advance using a word processor function, index data relating to the document with the note is generated. An apparatus for displaying a portion including a keyword in a document using generated index data when a keyword is input by a user will be described.

  FIG. 1 is a block diagram illustrating a configuration of a search device 10 according to the first embodiment. The search device 10 includes a control unit 101 that controls components of the search device 10, a storage unit 102 that stores programs and various data for instructing various types of processing by the control unit 101, and is used as a work area for the control unit 101 and the like. A display unit 103 that displays characters and graphics to the user, and an operation unit 104 that generates a predetermined signal in response to a user operation and delivers the signal to the control unit 101 are provided. The operation unit 104 in the present embodiment is, for example, a keyboard or a mouse. The search device 10 may be realized by causing a general-purpose computer to perform processing according to an application program, or may be realized by dedicated hardware.

  The storage unit 102 stores, in advance, a keyword DB 1023 (hereinafter referred to as “DB”) that stores document data 1021 in text format and keyword data indicating keyword candidates that may be included in the document indicated by the document data 1021. Called).

  The user can display the contents of the document data 1021 on the display unit 103 and input a note at a specific position of the document by an operation using the operation unit 104. The control unit 101 associates text data indicating a note input by the user with a portion in a document with a note (hereinafter referred to as “partial document”), and stores the note data in the storage unit 102 as note data. An adding unit 1011 is provided. In accordance with an instruction from the annotation adding unit 1011, the storage unit 102 stores the annotation data in the annotation DB 1022.

  FIG. 2 is a diagram illustrating a state in which a document indicated by the document data 1021 and a note input by the user are displayed on the display unit 103. In the display shown in FIG. 2, the note input by the user is shown as a footnote in the lower margin. In addition, (* 1) and (* 2) shown in the document are markers that indicate the positions where the user has added their notes.

  FIG. 3 is a diagram illustrating the contents of the note DB 1022. The note DB 1022 stores a plurality of note data including “number” for identifying the note, “position” in the document in which the note is inserted, and “content” of the note. In the “position” of the annotation data, for example, data in a format such as “third paragraph, first sentence, 24th character” is stored. This data indicates that a note has been inserted after the 24th character of the first sentence in the third paragraph counting from the beginning of the document.

  The control unit 101 specifies which partial document in the document the note indicated by the “content” of the note data is attached, and data indicating the specified partial document (hereinafter “partial document data”). A specifying unit 1012 having a function of cutting out from the document data 1021. First, the specifying unit 1012 specifies the position in the document to which the note is attached based on the data stored in the “position” for each piece of note data included in the note DB 1022. For example, the identification unit 1012 has inserted a note “... in Japan” at the position of the marker (* 1) shown in FIG. 2 based on the note data of the number “* 1” illustrated in FIG. Is identified.

  Subsequently, the specifying unit 1012 specifies the target partial document to which a note is attached. Specifically, the specifying unit 1012 searches for a character string indicating the latest punctuation mark or subject that appears before the insertion position of the note, and determines the character immediately after the searched punctuation mark or the first character of the character string indicating the searched subject. A character string in a range having a start position and a character immediately before the insertion position of the note as an end position is specified as a target partial document to which a note is attached. For example, in the case of a note with a number (* 1) (hereinafter referred to as “note (* 1)”), “application of the provision of Article 7 (2)” is specified as the target partial document.

  However, the method by which the specifying unit 1012 specifies the target partial document to which the annotation is attached is not limited to the above, and for example, the entire sentence including the insertion position of the note or the entire paragraph may be specified as the partial document. If an underline is added before and after the insertion position of the note, the underlined part may be specified as a partial document. In addition, the user may be able to select the unit of the partial document.

  Subsequently, the specifying unit 1012 delivers the partial document data indicating the specified partial document to the first search unit 1013. In addition, the specifying unit 1012 generates range data indicating the range of the partial document in the entire document, and delivers the range data to the index data generation unit 1014. The range data is, for example, data in a format such as “(* 1) 3rd paragraph 1st sentence 13th character to 3rd paragraph 1st sentence 24th character”. The head (* 1) of the range data is the number of the corresponding note data.

  The control unit 101 includes a first search unit 1013 that searches a keyword indicated by the keyword DB 1023 from the partial documents. FIG. 4 is a diagram illustrating the contents of the keyword DB 1023. The keyword DB 1023 stores a plurality of keyword data indicating keywords.

  A so-called wild card can be used as a keyword included in the keyword DB 1023. For example, the keyword “Article * (*)” may include any character or character string in the “*” part. Therefore, the search result using the keyword “Article * (*)” as a search key includes, for example, “Article 1 (1)”, “Article 9 (12)”, and the like.

  The first search unit 1013 searches the partial document indicated by the partial document data received from the specifying unit 1012 for a character string that matches the keyword indicated by the keyword data for each of the keyword data included in the keyword DB 1023. To do. For example, the keyword “Article 7 (2)” is obtained as a search result from the partial document “Application of Article 7 (2)”. The first search unit 1013 controls the data indicating the search results, that is, the keyword data included in the keyword DB 1023, extracted from the keyword indicating the keyword included in the partial document with the note, as extracted keyword data. The data is delivered to the index data generation unit 1014 included in the unit 101.

  For each keyword extracted by the first search unit 1013, the index data generation unit 1014 generates index data indicating the position where the keyword appears in the document. As described above, the index data generation unit 1014 receives the range data indicating the range of the partial document in the document from the specifying unit 1012 and extracts the extracted keyword data indicating the keyword appearing in any one of the notes from the first search unit 1013. receive. The index data generation unit 1014 searches the entire document for a character string that matches each of the keywords indicated by the extracted keyword data, and generates position data indicating the position of the searched character string in the document. Subsequently, the index data generation unit 1014 determines whether or not the position is included in the range in the document indicated by the range data. The index data generation unit 1014 generates index data based on the results of these processes. The index data generated by the index data generation unit 1014 is stored in the index DB 1024 of the storage unit 102.

  FIG. 5 is a diagram illustrating the contents of the index DB 1024. The index DB 1024 stores index data having items of “keyword”, “position”, and “corresponding note number”. The “keyword” of the index data is any of the keyword data included in the extracted keyword data. The “position” of the index data is data indicating the position of the first character of the keyword appearing in the document. When the same keyword appears multiple times in the document, the “position” of the index data includes a plurality of data.

  The “corresponding note number” of the index data is data indicating the number of the note attached to the partial document including the keyword appearing in the document. Therefore, when a keyword is included in a partial document to which no note is attached, the “corresponding note number” corresponding to the “position” of the keyword is blank (“−”).

  For example, the data in the first line in FIG. 5 indicates that the keyword “international application” appears at the position of the 10th character of the first sentence in the second paragraph, and that the partial document containing the keyword is not annotated. Yes. On the other hand, according to the data on the 8th line ("*" for simplification of each figure is counted as 1 line), the keyword "Article 7 (2)" appears at the position of the 13th character in the 3rd paragraph, 1st sentence. The partial document including the keyword is indicated with a note (* 1).

  As described above, in a state where the index DB 1024 is stored in the storage unit 102, the user can input an arbitrary keyword by operating the operation unit 104 and display a partial document including the keyword. At that time, if a note is attached to the partial document, the note is also displayed at the same time. FIG. 6 is a diagram illustrating a screen (hereinafter referred to as “search screen”) displayed on the display unit 103 when the user searches for a partial document including a keyword.

  On the search screen, a box 1031 for the user to directly input a keyword or select from a pull-down menu, a button 1032 for the user to issue a search execution instruction, and a list 1033 for displaying the search result are displayed. In the pull-down menu of the box 1031, all keywords included in the index DB 1024 are displayed as options.

  The control unit 101 includes a second search unit 1015 that searches for a partial document and a note attached thereto using the input keyword as a search key. When the user inputs a keyword in the box 1031 and presses the button 1032, the second search unit 1015 searches the index data including the keyword input from the index DB 1024 in the “keyword” column, and The partial document including the position indicated by the data in the “position” column and its front and rear portions are cut out from the document data 1021 and displayed in the left column of the list 1033. At that time, the second search unit 1015 emphasizes the keyword included in the display in the left column of the list 1033 by surrounding it with a frame or the like.

  If the “corresponding note number” field of the searched index data is not blank, the second search unit 1015 searches the note DB 1022 for note data including the number included in that field in the “number” field, The contents of the “content” column are displayed in the right column of the list 1033.

  In the above description, the document data 1021 and the annotation DB 1022 are described as separated data. However, these data may be integrated into one piece.

  Further, in the above description, the search device 10 is described as generating the note DB 1022 using the note input by the user. However, the note DB 1022 generated by another device is stored via a storage medium, a communication line, or the like. May be input to the search device 10.

  In the above description, it has been described that keyword data indicating keywords as search key candidates is stored in the keyword DB 1023 in advance. However, for example, the control unit 101 applies to an entire document or a partial document with a note. The keyword DB 1023 corresponding to the document may be generated, for example, by selecting a word having a high appearance frequency among the included words as a keyword.

  As described above, according to the search device 10, index data is automatically generated using a keyword included in a partial document with a note. As a result, compared with index data generated using keywords extracted from the full text of the document, high-quality index data including keywords with a higher probability can be easily obtained, which is convenient.

  Further, according to the search device 10, when a user performs a search using the index data generated as described above, when a note is attached to a partial document including a keyword input by the user as a search key, Since the note is also displayed, the user can easily specify the partial document that includes the specific keyword and is annotated, and can read the searched partial document while referring to the note. It is possible and convenient.

[2. Second Embodiment]
Hereinafter, as a second embodiment of the present invention, index data is generated based on a note added by hand to a text document prepared in advance by a user using a tablet PC (Personal Computer), and the index data is stored in a server. A system that enables efficient search in a document by sharing with other users via the apparatus will be described. In the system of the second embodiment, index data and search results are translated, so that the user can use a keyword in a language different from the language of the document as a search key. Hereinafter, only points different from the first embodiment will be described in order to avoid duplication of explanation. In the drawings referred to below, the same reference numerals are given to the same or corresponding components as those in the first embodiment.

  FIG. 7 is a block diagram showing the configuration of the search system 2 according to the second embodiment. The search system 2 is used by each of a plurality of users to attach a note to a document and used for performing a search in the document. Tablet PCs 20-1 to k (where k is an arbitrary natural number), each tablet PC 20 Is composed of a server device 21 that stores the index data generated in FIG. 4 and provides search results in response to a request from the tablet PC 20, and a network 22 that interconnects each tablet PC 20 and the server device 21. In the following description, when it is not necessary to distinguish each of the tablet PCs 20-1 to 20-k, they are simply referred to as a tablet PC 20.

  The display unit 103 is a liquid crystal display, for example. The operation unit 104 measures the position where the writing pressure is applied and the magnitude of the writing pressure according to the writing operation of the user using the pen-type stylus, and data indicating the position and the magnitude of the measured writing pressure (hereinafter referred to as the writing pressure). , Referred to as “handwriting data”), and is integrated with the display unit 103. The handwriting data is displayed on the display unit 103 together with the document. Therefore, the user can add a handwritten note to the document displayed on the display unit 103. FIG. 8 is a diagram illustrating a state in which a document and a note are displayed on the display unit 103.

  The control unit 101 includes a character recognition processing unit 2011 that performs character recognition processing on handwriting data indicating a note handwritten by a user and generates text data indicating the content of the note. The character recognition processing unit 2011 generates note data by associating the generated text data of the note with position data indicating the position on the screen where the note is written, and stores the note data in the note DB 1022. As the position where the note is written, for example, the position of the upper left corner of the rectangle surrounding the handwriting data indicating the note is used as the representative point.

  FIG. 9 is a diagram illustrating the contents of the annotation DB 1022 stored in the storage unit 102. In the second embodiment, an array including three elements is stored in the “position” column of the note data, the first element is a page number, the second element is a distance from the left end of the page (the unit is millimeters), the third Each element indicates a distance from the upper end of the page (unit: millimeters, for example).

  The specifying unit 1012 specifies the target partial document to which the note is attached based on the positional relationship between the document displayed on the screen and the note, and delivers the partial document data indicating the specified partial document to the first search unit 1013. At the same time, range data indicating the range of the partial document is delivered to the index data generation unit 1014.

  The first search unit 1013 extracts keywords included in the partial document data, and generates extracted keyword data indicating the extracted keywords. Incidentally, the control unit 101 includes a conversion unit 2012 that translates a word or document in one language into a word or document in another language. The first search unit 1013 delivers the generated extracted keyword data to the conversion unit 2012. Upon receiving extracted keyword data in a language other than the reference language, the conversion unit 2012 translates each of the keywords indicated by the received extracted keyword data into the reference language, and generates reference language extracted keyword data indicating the result. In the following description, as an example, it is assumed that the language used by the user is Japanese and the reference language is English. The conversion unit 2012 delivers the generated reference language extraction keyword data to the index data generation unit 1014.

  The index data generation unit 1014 generates index data based on the range data received from the specifying unit 1012 and the reference language extraction keyword data received from the conversion unit 2012 and stores it in the index DB 1024 of the storage unit 102. FIG. 10 is a diagram illustrating the contents of the index DB 1024, and is different from the index DB 1024 in the first embodiment in that English keywords are stored in the “keyword” column.

  In the second embodiment, the index data generation unit 1014 generates index data related to the keywords included in the notes, in addition to generating index data related to the keywords included in the partial documents annotated as described above.

  Specifically, the index data generation unit 1014 searches each of the keywords indicated by the keyword DB 1023 from the notes indicated by the “content” column of the note DB 1022. If the search is successful, the index data generation unit 1014 passes the keyword data indicating the keyword used as the search key to the conversion unit 2012, and receives reference language keyword data indicating the English keyword corresponding to the keyword from the conversion unit 2012.

  Subsequently, the index data generation unit 1014 indicates a “keyword” for storing the reference language keyword data, a “note number” indicating the number of a note including the keyword, and a position of the first character of the partial document with the note in the document. Index data including each item of “corresponding position” is generated and stored in the note index DB 2021 of the storage unit 102. FIG. 11 is a diagram illustrating the contents of the note index DB 2021.

  The tablet PC 20 includes a communication unit 205 that transmits and receives data to and from the server device 21 via the network 22. The control unit 101 transmits the document data 1021, the annotation DB 1022, the index DB 1024, and the annotation index DB 2021 to the server device 21 using the communication unit 205 together with the name data indicating the name of the document.

  The server device 21 includes a control unit 211 that controls the components of the server device 21, a program that instructs various processes by the control unit 211 and various data, and a storage unit 212 that is used as a work area for the control unit 211, and the like. A communication unit 213 that transmits and receives data to and from each tablet PC 20 via the network 22 is provided.

  The control unit 211 includes a registration unit 2111 that stores data received from the tablet PC 20 in the storage unit 212. The registration unit 2111 adds the document data 1021 received from the tablet PC 20 to the document data group 2121. The document data group 2121 is a collection of document data received by the server device 21 from each of the tablet PCs 20, and each document data is identified from each other by, for example, name data.

  Also, the registration unit 2111 adds the data included in the note DB 1022, the index DB 1024, and the note index DB 2021 received from the tablet PC 20 to the note DB 2122, the index DB 2123, and the note index DB 2124, respectively. FIGS. 12, 13, and 14 are diagrams illustrating the contents of the note DB 2122, the index DB 2123, and the note index DB 2124, respectively. The annotation DB 2122, the index DB 2123, and the annotation index DB 2124 each include a “document name” column for storing name data in addition to the items included in the annotation DB 1022, the index DB 1024, and the annotation index DB 2021.

  As described above, the document data group 2121 is stored in the storage unit 212, and the user adds any data transmitted from the tablet PC 20 to the note DB 2122, the index DB 2123, and the note index DB 2124. A keyword can be input and a partial document including the keyword can be displayed together with a note attached to the partial document. At that time, partial documents and notes included in a document in a language different from the language of the input keyword are also displayed.

  FIG. 15 is a diagram illustrating a search screen displayed on the display unit 103. The search screen in the second embodiment includes an option button 2031 for selecting either a document or a note as a search target in addition to the objects included in the search screen in the first embodiment.

  When the user selects the option “document” with the option button 2031 and then performs an operation of pressing the button 1032, the conversion unit 2012 generates reference language keyword data indicating the input keyword translated into English, and the user Together with the selection data indicating the option “document” selected in step S3, it is transmitted to the server device 21.

  The control unit 211 of the server device 21 includes a search unit 2112 that performs in-document search using the index DB 2123 or the note index DB 2124 using the reference language keyword received from the tablet PC 20 as a search key. When receiving the selection data indicating the option “document” from the tablet PC 20, the search unit 2112 transmits the search result using the index DB 2123 to the tablet PC 20.

  Specifically, the search unit 2112 first searches the index data included in the index DB 2123 for data that matches the reference language keyword data received from the tablet PC 20 in the “keyword” column, and searches the index data. The partial document data indicating the partial document specified by the “position” of the retrieved index data is acquired from the document data specified by the “document name”. Further, when the “corresponding note number” field of the retrieved index data is not blank, the search unit 2112 matches the data of “document name” and “corresponding note number” of the retrieved index data from the note data included in the note DB 2122. The note data including the data to be searched is searched, and the data in the “content” column of the searched note data is acquired.

  As described above, the search unit 2112 acquires partial document data indicating a partial document including a keyword, and data indicating the note when the partial document is annotated, and uses the data as search result data. It transmits to tablet PC20.

  When the search result data received from the server device 21 includes partial documents or notes indicating a language other than Japanese, the tablet PC 20 translates them into Japanese by the conversion unit 2012. The tablet PC 20 displays the list 1033 of the search screen (see FIG. 15) using the received search result data or data obtained by translating the search result data.

  FIG. 16 is a diagram illustrating a search screen when the user selects the option “note” with the option button 2031. The operation of the tablet PC 20 and the server device 21 when the user selects the option “note” and presses the button 1032 is compared with the operation when the option “document” is selected. Since the search unit 2112 is the same except that the note index DB 2124 is used instead of the index DB 2123, the description thereof is omitted.

  In the above description, the conversion unit 2012 is described as being arranged in each tablet PC 20. However, the conversion unit 2012 is arranged in the server device 21 so that the server device 21 translates keywords and search results. May be.

  Further, in the above description, in the generation of index data, the search unit 2112 has been described as generating the reference language keyword data by translating keywords in units of words. However, the search unit 2112 translates a sentence or paragraph including the keywords. Then, reference language keyword data indicating a keyword portion included in the translation obtained as a result may be generated. In that case, compared with the case where the keyword is translated in units of words, as a result of selecting a translation suitable for the context, index data with higher accuracy is generated.

  Further, in the above description, the index data that can be used between different languages is generated by translating the keywords used for the index data into the reference language. For example, when a Japanese keyword is input by a user, search processing may be performed using the Japanese index data.

  In the above description, data indicating a plurality of languages is mixed in the data indicating the notes included in the search result data transmitted from the server device 21 to the tablet PC 20, and these languages are translated by the conversion unit 2012. Explained as a thing. However, it is possible to arbitrarily change at which stage the data indicating the annotation is translated. For example, the conversion unit 2012 of the tablet PC 20 may transmit the text included in the document data 1021, the note DB 1022, the index DB 1024, and the note index DB 2021 to the server device 21 after being translated into a plurality of languages. In this case, since the server device 21 can store in advance the document data group 2121, the note DB 2122, the index DB 2123, and the note index DB 2124 corresponding to each language in the storage unit 212, the server device 21 uses the user who made the search request. Using the document data group 2121 or the like corresponding to the language to be used, data indicating a note described only in the language can be transmitted to the tablet PC 20 as search result data.

  In the above description, the user uses the tablet PC 20 as the terminal device. However, the terminal device used by the user is, for example, a general PC and a pointing device such as a pen tablet independent of a mouse and a display unit. May be connected. In that case, the user adds notes to the document by a writing operation using those pointing devices.

  In the above description, the text data generated by the character recognition processing unit 2011 is used as the data indicating the content of the note. However, the handwriting data indicating the note is used instead of or in addition to the text data. You may do it. In such a case, for example, handwritten characters that are difficult to recognize in character recognition processing or notes such as graphics that cannot be recognized can be used to be displayed in search results.

  As described above, according to the search system 2, it is convenient to efficiently generate index data by using a note added by handwriting to a document by a user. Further, according to the search system 2, as a result of translating the keywords included in the index data, the documents included in the search results, or the language of the notes into other languages as necessary, the documents in various languages are searched. Can be convenient. Furthermore, according to the search system 2, since index data based on the keyword included in the note is generated, it is possible to search for a note including a specific keyword. At that time, the target partial document to which the note is attached is also displayed together with the note, which is convenient when examining the content of the note.

[3. Third Embodiment]
For example, in international conferences, documents showing the same content in a plurality of languages are often distributed to participants as printed matter. In that case, participants write notes on the distributed printed matter by hand in a language convenient to them. Hereinafter, as a third embodiment of the present invention, a description will be given of a search device that enables efficient search by associating handwritten notes with printed documents of documents having different language versions. Hereinafter, only points different from the first embodiment and the second embodiment will be described in order to avoid duplication of explanation. In the drawings referred to below, the same reference numerals are given to the same or corresponding components as those in the first embodiment or the second embodiment.

  FIG. 17 is a block diagram showing the configuration of the search system 3 according to the third embodiment. The search system 3 includes a search device 30 used by a user for generating and searching index data, and a scanner 31 connected to the search device 30. The scanner 31 is a device that optically reads characters, diagrams, images, and the like written on a paper surface, and generates and outputs image data indicating them.

  The storage unit 102 stores a document data group 3021 including document data indicating versions of documents distributed in advance to participants in international conferences. The document data included in the document data group 3021 is distinguished from each other by language data indicating which language the version is.

  In addition, the storage unit 102 stores an inter-document correspondence DB 3022 that indicates the correspondence between partial documents included in documents in different languages. FIG. 18 is a diagram illustrating the contents of the inter-document correspondence DB 3022. The inter-document correspondence DB 3022 stores inter-document correspondence data corresponding to each partial document constituting the document. Inter-document correspondence data has items corresponding to each language such as “English”, “Japanese”, etc., and each item stores position data indicating the position of the first character of the partial document in the corresponding language. Yes. If word-by-sentence translation is performed on a sentence-by-sentence basis in any of the language versions, the correspondence between partial documents in different languages can be specified by the order of sentence separation. However, there are cases where one sentence in one language is translated into a plurality of sentences in another language. Even in such a case, the retrieval apparatus 30 can specify the correspondence relationship of the partial documents by using the inter-document correspondence DB 3022.

  The user of the search system 3, for example, the organizer of the international conference, collects the distributed printed matter from the participants and reads the description content of the printed matter with the scanner 31. Image data indicating the description content of the printed matter generated by the scanner 31 is delivered to the search device 30. The control unit 101 includes an image data separation unit 3011 that separates and generates image data indicating each character from image data in which printed characters and handwritten characters are mixed. The image data separation unit 3011 performs, for example, a filtering process on the image data received from the scanner 31 to display image data indicating a document portion that has been printed in advance on the printed matter, and an image indicating a note portion handwritten by a participant. Generate data.

  As filters for separating image data of printed characters and handwritten characters, those using various existing techniques can be used. For example, a filter may be used that extracts characters whose ratio between the linear component and the non-linear component included in each diagram is equal to or greater than a predetermined threshold value as a printed character, and extracts characters that are less than the predetermined threshold value as handwritten characters. Alternatively, a filter may be used that extracts characters whose density variance is less than a predetermined threshold as printed characters and extracts characters that are greater than or equal to the predetermined threshold as handwritten characters.

  When the image data separation unit 3011 generates image data indicating the document portion and the note portion as described above, the image data separation unit 3011 delivers the image data to the character recognition processing unit 2011. The character recognition processing unit 2011 recognizes the character indicated by each of the image data, and generates text data indicating the recognition result. At that time, the character recognition processing unit 2011 also recognizes which language the character indicated by the image data belongs to. The character recognition processing unit 2011 associates the generated text data with the position data indicating the position on the screen of the image corresponding to the text data, and delivers it to the specifying unit 1012 together with the language data indicating the recognized language.

  The specifying unit 1012 specifies a part of the text data of the document part corresponding to the text data of the note part received based on the vertical positional relationship between the document part and the note part on the screen, for example. A part of the text data identified as such is partial document data indicating a partial document to which a note is attached. Subsequently, the specifying unit 1012 searches the document data identified by the received language data for a portion that matches the specified partial document data. The specifying unit 1012 associates the position data indicating the position of the first character of the partial document obtained by the search with the text data of the annotation part, and stores it in the annotation DB of the storage unit 102 as annotation data.

  The storage unit 102 stores a note DB, a keyword DB, and an index DB corresponding to each language. In FIG. 17, a collection of note DBs is shown as a note DB group 3023, a collection of keyword DBs is shown as a keyword DB group 3024, and a collection of index DBs is shown as an index DB group 3025. For example, when a document read by the scanner 31 is in Japanese, the specifying unit 1012 stores note data in an index DB corresponding to Japanese. The configuration of the note DB stored in the storage unit 102 of the search device 30 is the same as the note DB 1022 (see FIG. 3) in the first embodiment.

  As described above, when note data is stored in the note DB by the specifying unit 1012, the keyword extraction processing by the first search unit 1013 is performed by the same method as in the first embodiment, and the index data generation unit The index data is generated by 1014 and stored in the index DB. However, in the processing of the first search unit 1013 and the index data generation unit 1014 in the third embodiment, document data, note DB, keyword DB, and index DB corresponding to the language recognized by the character recognition processing unit 2011 are selected. Used. The configuration of the index DB stored in the storage unit 102 of the search device 30 is the same as the index DB 1024 (see FIG. 5) in the first embodiment.

  As described above, for each of a plurality of languages, a note DB that stores note data relating to notes added by a plurality of readers, and an index that stores index data relating to keywords included in the partial documents to which those notes are attached. When the DB is stored in the storage unit 102, the user can input a keyword and display a partial document including the keyword together with a note attached to the partial document. FIG. 19 is a diagram illustrating a search screen displayed on the display unit 103 of the search device 30.

  In addition to the objects included in the search screen (see FIG. 6) in the first embodiment, the search screen includes a box 3031 for the user to select a language used for keyword input and result display. Hereinafter, it is assumed that the user inputs “Japanese” in the box 3031. When “Japanese” is input in the box 3031, data in the “keyword” column of the index DB corresponding to Japanese is displayed in the pull-down menu of the box 1031.

  Subsequently, when the user inputs a keyword in the box 1031 and performs an operation of pressing the button 1032, the second search unit 1015 searches the index DB corresponding to Japanese for index data including the input keyword.

  Subsequently, the second search unit 1015 searches the inter-document correspondence DB 3022 for inter-document correspondence data in which data matching the data in the “position” column of the retrieved index data is stored in the “Japanese” column. The second search unit 1015 retrieves position data corresponding to each language included in the retrieved inter-document correspondence data, and retrieves note data including data matching the retrieved position data from the note DB corresponding to each language.

  The second search unit 1015 delivers the text data in the “content” column of the searched note data to the conversion unit 2012, and converts the note indicated by the delivered text data into Japanese text data indicating the result of the translation before the note. It is received from the converter 2012 together with language data indicating the language. The second search unit 1015 searches the list 1033 in FIG. 19 using the partial document data specified by the “position” column of the Japanese note data and the translated text data and language data received from the conversion unit 2012. Display the results.

  By the way, in the above description, based on text data obtained as a result of performing character recognition processing on image data obtained by scanning a printed matter, the language of the document described in the printed matter is recognized, and the printed matter is a document. However, for example, a symbol or page number indicating a language is printed in advance on a printed matter, and the character recognition processing unit 2011 recognizes the symbol or number. Thus, the language and page may be specified.

  In the above description, the image data received by the search device 30 from the scanner 31 is filtered to separate the image data indicating the document portion and the image data indicating the note portion. The method for separating the document portion and the note portion is not limited to using a filter. For example, since the search device 30 stores document data, matching is performed between the image data received from the scanner 31 and the display data of each page of the document, and the received image data corresponds to any document in any document. You can specify whether it corresponds to a page. Therefore, the search device 30 uses the display data of the specified page as image data indicating the document portion, and uses the difference between the image data received from the scanner 31 and the display data of the specified page as image data indicating the note portion. Also good. In this case, the search device 30 does not need to perform character recognition processing on the image data indicating the document portion, and can directly use the corresponding page portion of the document data.

  In the above description, it has been described that in the display of the search result, notes written by different participants on the document in the same language are not distinguished. The ID and name are printed, and after the ID and name are recognized by the character recognition processing unit 2011, the contents are displayed in the search result, thereby identifying the participant to whom the user wrote each note. It may be possible.

  As described above, according to the search system 3, the user is accompanied by display of notes written in a plurality of languages without repeating the search operation for each language when there are versions related to different languages for the same document. Search results can be obtained, which is convenient.

[4. Modified example]
The present invention is not limited to the above-described embodiment, and various modifications can be made. For example, an apparatus or a system that provides the same result may be constructed by arbitrarily combining separable technical elements included in the first to third embodiments. For example, the index data may be shared by interconnecting a plurality of search devices according to the first embodiment via a network as in the second embodiment.

  Moreover, arrangement | positioning of the structure part of each apparatus shown in the above-mentioned embodiment can be changed arbitrarily. For example, the search system according to the third embodiment can be realized by incorporating a scanner in the search device, or a search device can be searched by connecting a translation device having a similar function to the search device instead of including a conversion unit. A system may be realized.

It is the block diagram which showed the structure of the search device concerning 1st Embodiment. It is the figure which illustrated the mode that the document concerning 1st Embodiment and the note were displayed. It is the figure which illustrated the contents of note DB concerning a 1st embodiment. It is the figure which illustrated the contents of keyword DB concerning a 1st embodiment. It is the figure which illustrated the contents of index DB concerning a 1st embodiment. It is the figure which illustrated the search screen concerning 1st Embodiment. It is the block diagram which showed the structure of the search system concerning 2nd Embodiment. It is the figure which illustrated the mode that the document concerning 2nd Embodiment and the note were displayed. It is the figure which illustrated the contents of note DB concerning a 2nd embodiment. It is the figure which illustrated the contents of index DB concerning a 2nd embodiment. It is the figure which illustrated the contents of note index DB concerning a 2nd embodiment. It is the figure which illustrated the contents of note DB concerning a 2nd embodiment. It is the figure which illustrated the contents of index DB concerning a 2nd embodiment. It is the figure which illustrated the contents of note index DB concerning a 2nd embodiment. It is the figure which illustrated the search screen concerning 2nd Embodiment. It is the figure which illustrated the search screen concerning 2nd Embodiment. It is the block diagram which showed the structure of the search system concerning 3rd Embodiment. It is the figure which illustrated the contents of DB corresponding to documents concerning a 3rd embodiment. It is the figure which illustrated the search screen concerning 3rd Embodiment.

Explanation of symbols

2.3 ... Search system, 10.30 ... Search device, 20 ... Tablet PC, 21 ... Server device, 22 ... Network, 31 ... Scanner, 101, 211 ... Control unit, 102/212 ... Storage unit, 103 ... Display unit , 104 ... operation unit, 205 213 ... communication unit, 1011 ... note addition unit, 1012 ... identification unit, 1013 ... first search unit, 1014 ... index data generation unit, 1015 ... second search unit, 1021 ... document data, 1022 · 2122 ... Note DB, 1023 ... Keyword DB, 1024 · 2123 ... Index DB, 2011 ... Character Recognition Processing Unit, 2012 ... Conversion Unit, 2021 · 2124 ... Note Index DB, 2111 ... Registration Unit, 2112 ... Search Unit, 2121 3021 ... Document data group, 3011 ... Image data separation unit, 3022 ... Inter-document correspondence DB, 3023 ... Note DB group 3024 ... keyword DB group, 3025 ... index DB group.

Claims (15)

Storage means for storing document data indicating one document, and one or more note data each indicating a note associated with any partial document constituting the one document;
Index data generating means for generating index data indicating a position where a phrase that matches one phrase included in the partial document associated with the note indicated by any one of the one or more annotation data is included in the one document An index data generating device comprising: A specifying unit for specifying a partial document with one note from the one document;
The index data generation apparatus according to claim 1, wherein the storage unit stores note data indicating the one note as note data associated with the partial document specified by the specifying unit. The said specifying means performs the said specification based on the spatial positional relationship when the document shown by the said document data and the note shown by the said note data are displayed on the same plane. Index data generator. Conversion means for converting a phrase expressed in one language into a translation expressed in another language having the same meaning as the phrase;
The index data generation device according to claim 1, wherein the index data generation unit associates a translated word obtained from the one word / phrase by the conversion unit with the search data. The storage means stores in advance keyword data indicating a keyword,
The index data generation device according to claim 1, wherein the index data generation unit uses the keyword as the one phrase. The index data generating means indicates a position where a phrase that matches one phrase included in the note indicated by any one of the one or more note data is included in the note indicated by any of the one or more note data. The index data generation device according to claim 1, wherein the index data is generated. Display means for displaying a document indicated by the document data;
Handwriting data generating means for generating handwriting data indicating handwriting according to a user's writing operation,
The index data generation device according to claim 1, wherein the storage unit stores the handwriting data generated by the handwriting data generation unit as the annotation data. Display means for displaying a document indicated by the document data;
Handwriting data generating means for generating handwriting data indicating handwriting according to a user's writing operation;
Character recognition processing means for performing character recognition processing on the handwriting data generated by the handwriting data generation means, and generating text data indicating the recognized character string, and
The index data generation apparatus according to claim 1, wherein the storage unit stores the text data generated by the character recognition processing unit as the annotation data. An image reading means for optically reading an image from a document including a document described in type and one or more notes written by hand on the document, and generating image data indicating the read image;
Image data separation means for generating image data indicating the document written in the type and one or more image data each indicating one or more notes written by handwriting from the image data generated by the image reading means; ,
Character recognition processing means for performing character recognition processing on the image data indicating the document generated by the image data separation means, and generating text data indicating the content of the one document, and
The storage means stores text data indicating the content of the one document generated by the character recognition processing means as the document data, and at least one of the handwritten information generated by the image data separation means The index data generation device according to claim 1, wherein one or more pieces of image data indicating each of the notes are stored as the one or more pieces of annotation data. An image reading means for optically reading an image from a document including a document described in type and one or more notes written by hand on the document, and generating image data indicating the read image;
Image data separation means for generating image data indicating the document written in the type and one or more image data each indicating one or more notes written by handwriting from the image data generated by the image reading means; ,
The image data indicating the document generated by the image data separating unit is subjected to character recognition processing to generate text data indicating the contents of the one document, and the one or more generated by the image data separating unit Character recognition processing means for performing character recognition processing on each of the one or more image data each indicating a note and generating one or more text data indicating the contents of the one or more notes;
The storage means stores text data indicating the contents of the one document generated by the character recognition processing means as the document data, and stores the contents of the one or more notes generated by the character recognition processing means. The index data generation device according to claim 1, wherein one or more pieces of text data indicated are stored as the one or more annotation data. Document data indicating one document, one or more note data each indicating a note associated with any partial document constituting the one document, and a phrase that matches the keyword with respect to each of the one or more keywords Storage means for storing index data indicating a position included in the one document;
An input means for receiving keyword data indicating one keyword;
Search means for searching for one or more partial documents including the one keyword from a plurality of partial documents constituting the one document based on the index data;
A data search apparatus comprising: output means for outputting note data indicating a note associated with the partial document searched by the search means. Conversion means for converting a phrase expressed in one language into a translation expressed in another language having the same meaning as the phrase;
The data search apparatus according to claim 11, wherein the search means performs the search using a translation obtained from the one keyword by the conversion means instead of the one keyword. First document data representing a first document expressed in one language, and one or more first annotation data each representing a note associated with one of the partial documents constituting the first document And index data indicating a position in the first document that includes a phrase that matches the keyword for each of the one or more keywords, and a second language expressed in another language having the same meaning as the first document. Storage means for storing second document data indicating two documents and one or more second note data each indicating a note associated with any partial document constituting the second document;
An input means for receiving keyword data indicating one keyword;
Search means for searching one or more partial documents including the one keyword from a plurality of partial documents constituting the first document based on the index data;
Corresponding to annotation data indicating notes associated with the partial document searched by the searching means, and partial documents constituting the second document and corresponding to the partial documents searched by the searching means An output means for outputting note data indicating the attached note. Processing for storing document data indicating one document and one or more note data each indicating a note associated with any partial document constituting the one document;
A process of generating index data indicating a position where a phrase that matches one phrase included in the partial document associated with the note indicated by any one of the one or more annotation data is included in the one document; A program characterized by being executed. Document data indicating one document, one or more note data each indicating a note associated with any partial document constituting the one document, and a phrase that matches the keyword with respect to each of the one or more keywords Storing index data indicating a position included in the one document;
Processing to receive keyword data indicating one keyword,
A process of searching for one or more partial documents including the one keyword from a plurality of partial documents constituting the one document based on the index data;
A program for causing a computer to execute processing for outputting note data indicating a note associated with a partial document searched in the search.

Images