JP2007305046A - Information processor for generating kanji reading, information processing method, program for attaining information processing and recording medium with the program recorded thereon - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To efficiently generate reading of a full name. <P>SOLUTION: This program includes the steps of: acquiring individual identification information from mail address information (S200); dividing the acquired individual identification information (S202); calculating the matching degree among the divided individual identification information, family name candidate and given name candidate (S300); calculating the matching degree between the divided individual identification information and given name candidate (S300); and determining the reading of a full name on the basis of the calculated matching degrees (S400). <P>COPYRIGHT: (C)2008,JPO&INPIT

Description

  The present invention relates to an information processing apparatus that generates kanji readings, and more particularly to an apparatus that generates readings of surnames expressed in kanji.

  Currently, optical character recognition (OCR) technology is used in various character readers such as document readers and form readers. Among them, a device for recognizing character information from image information obtained by imaging the paper surface of a business card and classifying the recognized character information (character string) into items such as first name, last name, address, company name, etc. There is. There is a technique for generating readings of character strings classified in this way. As a technique for generating such a character string reading, for example, there is a technique disclosed in the following publications.

  Japanese Patent Application Laid-Open No. 5-20300 (Patent Document 1) discloses a technique for improving the recognition accuracy of character recognition and generating a recognized character reading accurately. The document processing device disclosed in Patent Document 1 includes a reading unit for reading a front image and a back image of a business card, and a front character and a back character of a business card based on the front image and the back image read by the reading unit. Recognizing means for recognizing, a determining means for determining whether or not the back side character recognized by the recognizing means is in English notation, and when the determining means determines that it is in English notation, And the back side characters are divided into each address book data such as address, first name, phone number, etc., and the front side character and back side character address book data divided by the split means are respectively compared and the front side character notation Based on the notation of the reverse character selected by the selection means and the reverse character selected by the selection means, the reading corresponding to the notation is generated. Comprising a generation unit of order, the readings generated by the generating means, and storage means for storing in association with the representation of the surface character corresponding to the read.

According to the document processing apparatus disclosed in this publication, the front image and the back image of the business card read by the reading unit are recognized as the front and back characters by the recognition unit. It is determined by the determining means whether or not the back side character is in English. When the determination result is written in English, the dividing means divides the front character and the reverse character into address book data such as an address, first name, and phone number. The address book data of the divided front character and back character are respectively compared by the selecting means, and the notation of the back character that matches the notation of the front character is selected. The generation means generates a reading corresponding to the notation based on the notation of the selected back side character. The generated reading is stored in the storage means in association with the notation of the surface character corresponding to the reading. Therefore, when there is an English notation of the front character on the back side of the business card, the front character is recognized by referring to the English notation, so that the recognition rate is improved. In addition, since readings are generated from the English notation on the back side, accurate readings can be generated.
JP-A-5-20300

  However, the document processing apparatus disclosed in Patent Document 1 has a problem in that it takes time to read the front and back images of a business card in order to generate a reading corresponding to the notation of the front character. It was.

  The present invention has been made to solve the above-described problems, and an object of the present invention is to provide an information processing apparatus, an information processing method, and a program for realizing the information processing capable of efficiently generating readings of first and last names And a recording medium on which the program is recorded.

  An information processing apparatus according to a first aspect of the present invention represents means for preliminarily storing first information in which kanji and phonetic characters (for example, romaji and kana) representing kanji reading are associated with each other, and represents an individual's first and last name Acquisition means for acquiring kanji and second information included in the personal mail address; generation means for generating candidates for reading kanji representing first and last names based on the first information; Determining means for determining the reading of the kanji representing the first and last names based on the result of collating the information with the candidates.

  According to the first aspect of the invention, the kanji representing the personal first name and the second information included in the personal e-mail address are acquired by the acquiring means. For example, in the case of a business card, an individual's full name and e-mail address are often written on the same surface, either the front surface or the back surface. Therefore, the kanji representing the first and last name and the second information can be acquired from the image information obtained by imaging only one of the business cards. A kanji reading candidate representing the acquired first and last name of the individual is generated based on first information stored in advance as a phonetic character. The result (for example, the number of matching characters) of the generated candidate and the second information (for example, information for identifying an individual (hereinafter also referred to as personal identification information)) included in the acquired email address And kanji readings representing first and last names are determined. Email addresses often include romanized readings of all or part of an individual's full name. Therefore, the reading of the full name can be determined from the generated candidates. As a result, it is possible to provide an information processing apparatus that can efficiently determine the reading of first and last names.

  In the information processing apparatus according to the second invention, in addition to the configuration of the first invention, the determining means is means for determining reading based on the number of characters matching the second information and the candidate including.

  According to the second invention, reading is determined based on the number of characters that match the second information and the candidate. Therefore, it is possible to determine the reading of an individual's full name based on the number of characters that match the reading of all or part of the full name and the candidate. Therefore, the reading of first and last names can be determined with high accuracy.

  In the information processing apparatus according to the third invention, in addition to the configuration of the second invention, the determining means includes means for determining a candidate having the largest number of matching characters as a reading.

  According to the third invention, the candidate having the largest number of matching characters between the second information and the candidate is determined as the reading. Therefore, a candidate that more closely matches the reading of a person's full name than other candidates can be determined as a reading of the full name.

  In the information processing apparatus according to the fourth invention, in addition to the configuration of the first invention, the determining means determines the reading based on the position of the character that matches the second information and the candidate. including.

  According to the fourth aspect, reading is determined based on the position of the character that matches the second information and the candidate. For example, reading is determined with priority given to candidates whose first or last character matches the first or last character of the second information. Candidates whose leading or trailing characters match the leading or trailing characters of the second information are often more accurate readings than non-matching candidates. Therefore, the reading of first and last names can be determined with high accuracy.

  In the information processing apparatus according to the fifth aspect of the invention, in addition to the configuration of the fourth aspect of the invention, the determination means prioritizes a candidate whose leading character matches the leading character of the second information and determines reading Means for doing so.

  According to the fifth invention, the candidate whose leading character matches the leading character of the second information is prioritized and reading is determined. Candidates whose first character matches the first character of the second information often read more accurately than candidates that do not match. Therefore, the reading of first and last names can be determined with high accuracy.

  In the information processing device according to the sixth aspect of the invention, in addition to the configuration of the fourth aspect of the invention, the deciding means prioritizes candidates whose last character matches the last character of the second information and decides reading Means for doing so.

  According to the sixth aspect of the present invention, the candidate whose last character matches the last character of the second information is given priority and reading is determined. Candidates whose last character matches the last character of the second information often read more accurately than candidates that do not match. Therefore, the reading of first and last names can be determined with high accuracy.

  In the information processing apparatus according to the seventh invention, in addition to the configuration of any one of the fourth to sixth inventions, the generation means reads the surname candidate and the kanji character representing the first name as the kanji character representing the surname. Means for generating a name candidate. The determining means determines the reading so that the position of the character matching the surname candidate and the second information is different from the position of the character matching the first name candidate and the second information in the second information. Including means.

  According to the seventh invention, the surname candidate and the first name candidate are generated separately. Reading is determined so that the position of the character that matches the surname candidate and the second information is different from the position of the character that matches the surname candidate and the second information in the second information. Thereby, since it is suppressed that the character which matches the last name candidate and the first name candidate in 2nd information is duplicated, the correct reading of a last name and a first name can be determined.

  In the information processing apparatus according to the eighth invention, in addition to the configuration of any one of the first to seventh inventions, the determining means includes the second information and the second information for any number of candidate characters and When the ratio of the number of characters matching the candidate is smaller than a predetermined ratio in all candidates, a means for determining a candidate satisfying a predetermined condition as a reading is included.

  According to the eighth aspect, when the ratio of the number of characters matching the second information and the candidate to the number of characters of the second information and the candidate is smaller than a predetermined ratio in all candidates The reading is determined as a candidate that satisfies a predetermined condition. For example, a candidate that is frequently used for the first and last names is determined as a reading. Therefore, even when the personal name is not included in the e-mail address, or even when a part of the e-mail address coincides with a partial reading of the personal first and last name, an accurate reading can be determined.

  In the information processing apparatus according to the ninth aspect of the invention, in addition to the configuration of the eighth aspect of the invention, the determining means includes means for determining, as a reading, a candidate that is frequently used for first and last names.

  According to the ninth aspect, when the ratio of the number of matching characters is smaller than a predetermined ratio in all candidates, a candidate that is frequently used for the first and last names is determined as a reading. Therefore, even if an email address does not include an individual's first and last name, or a part of the email address coincides with a part of the reading of an individual's first and last name, a candidate who is likely to be an accurate reading Can be determined to read.

  In the information processing device according to the tenth invention, in addition to the configuration of any one of the first to seventh inventions, the determining means includes the second information and the second information for any number of candidate characters. If the ratio of the number of characters that match the candidate is smaller than a predetermined ratio for all candidates, a means for determining reading based on the second information is included.

  According to the tenth aspect, when the ratio of the number of characters matching the second information and the candidate to the number of characters of the second information and the candidate is smaller than a predetermined ratio in all candidates The reading is determined based on the second information. For example, a kana corresponding to the second information is determined as a reading. Therefore, it is possible to determine readings of first and last names, and to determine readings that are highly likely to be accurate readings even when accurate reading candidates are not generated.

  The information processing apparatus according to the eleventh invention further includes means for generating a pseudonym corresponding to the second information in addition to the configuration of any one of the first to tenth inventions. The generating means includes means for generating candidates with pseudonyms. The determining means includes means for determining a reading based on the result of collating the corresponding kana and the candidate.

  According to the eleventh aspect, the reading can be determined based on the result of collating the kana corresponding to the second information with the kana representing the candidate reading.

  In the information processing apparatus according to the twelfth invention, in addition to the configuration of any one of the first to eleventh inventions, the second information includes a Roman character. The generating means includes means for generating candidates in Roman letters.

  According to the twelfth aspect, the reading can be determined based on the result of collating the Roman character included in the second information with the Roman character representing the candidate reading.

  In the information processing apparatus according to the thirteenth aspect, in addition to the configuration of any one of the first to twelfth aspects, the second information is a character for identifying an individual included in the mail address.

  According to the thirteenth invention, the second information is a character for identifying an individual included in the mail address. In many cases, characters for identifying an individual included in an e-mail address include all or part of a person's full name in Roman letters. Therefore, it is possible to accurately generate readings of first and last names.

  In the information processing apparatus according to the fourteenth invention, in addition to the configuration of any one of the first to thirteenth inventions, the acquisition means acquires character information by recognizing captured image information. including.

  According to the fourteenth aspect, character information is acquired by recognizing captured image information. Therefore, for example, kanji representing a first and last name, which is character information, and an e-mail address can be acquired from image information obtained by capturing an image such as a business card in which the first and last name and e-mail address are described.

  In the information processing apparatus according to the fifteenth aspect, in addition to the configuration of any one of the first to fourteenth aspects, the first information associates kanji with readings when kanji is used for first and last names. Information.

  According to the fifteenth aspect of the invention, the first information is information that associates kanji with readings when kanji is used for first and last names. Therefore, the capacity for storing the first information can be reduced as compared with the case of storing all readings of kanji.

  An information processing method according to a sixteenth aspect of the invention is an information processing method performed by an information processing apparatus including a storage unit, an acquisition unit, a generation unit, and a determination unit. This information processing method includes the step of storing, using the storage means, first information that associates kanji and phonetic characters representing the reading of kanji, and includes kanji representing an individual's first and last name and an individual's e-mail address. Obtaining the second information using the obtaining means, generating the kanji reading candidates representing the first and last names based on the first information using the creating means, and the second information Determining a kanji reading representing the first and last names based on the result of collating with the candidate using a determining means.

  According to the sixteenth invention, the kanji representing the personal first name and the second information included in the personal e-mail address are obtained using the obtaining means. For example, in the case of a business card, an individual's full name and e-mail address are often written on the same surface, either the front surface or the back surface. Therefore, when acquiring the kanji representing the first and last name and the second information from the image information obtained by capturing the business card, it is only necessary to acquire the image information of one side of the business card. A kanji reading candidate representing the acquired first and last name of the individual is generated based on first information stored in advance as a phonetic character. Based on the result (for example, the number and position of matching characters) of the generated candidate and the second information (for example, personal identification information) included in the acquired e-mail address, the reading of kanji representing the first and last names Is determined. Email addresses often include romanized readings of all or part of an individual's full name. Therefore, it is possible to determine the reading of first and last names from the generated candidates. As a result, it is possible to provide an information processing method capable of efficiently generating first and last name readings.

  A program according to the seventeenth invention is a program for realizing the information processing method according to the sixteenth invention. Thus, by installing the program according to the seventeenth invention on, for example, a computer, the computer can execute the information processing method according to the sixteenth invention.

  A recording medium according to an eighteenth aspect of the invention is a computer-readable recording medium recording the program according to the seventeenth aspect of the invention. Thus, by installing the program according to the seventeenth invention from the recording medium according to the eighteenth invention, for example, on the computer, the computer can execute the information processing method according to the sixteenth invention.

  Hereinafter, embodiments of the present invention will be described with reference to the drawings. In the following description, the same parts are denoted by the same reference numerals. Their names and functions are also the same. Therefore, detailed description thereof will not be repeated. In the present embodiment, an information processing apparatus that generates a reading of first and last names recognized from business card image information will be described. However, the information processing apparatus according to the present invention is not limited to this. For example, the information processing apparatus according to the present invention can be applied to an apparatus that generates readings of first and last names included in electronic mail information received from a communication destination.

  With reference to FIG. 1, an information processing apparatus 100 according to the present embodiment will be described. The information processing apparatus 100 is connected to the input device 200 and the output device 300.

  The input device 200 includes an image scanner, a CCD (Charge Coupled Device), and the like. The input device 200 inputs an image of a subject such as a business card, converts it into image information, and transmits the image information to the information processing device 100. The input device 200 may be a device that inputs image information of a subject from a communication destination or a recording medium.

  The output device 300 is configured by a liquid crystal display device or the like, and outputs readings and other character strings generated by the information processing device 100.

  The information processing apparatus 100 includes a storage unit 120, a character recognition unit 130, an information acquisition unit 140, a generation unit 150, a determination unit 160, and a control unit 110 connected to these units.

  The storage unit 120 stores intermediate data of processing executed by each unit. The storage unit 120 stores a character recognition information unit 122 and a kanji reading information unit 124 in advance.

  When the character recognition unit 130 receives the control signal from the control unit 110, the character recognition unit 130 recognizes the image information received from the input device 200 as character information using the character recognition information unit 122. OCR is used as a character recognition technique. Other techniques may be used.

  When the information acquisition unit 140 receives the control signal from the control unit 110, the information acquisition unit 140 acquires first and last name information and mail address information from the character information recognized by the character recognition unit 130.

  When the generation unit 150 receives the control signal from the control unit 110, the generation unit 150 generates a candidate for reading the first and last name information acquired by the information acquisition unit 140 using the kanji reading information unit 124.

  When receiving the control signal from the control unit 110, the determination unit 160 collates the reading candidate generated by the generation unit 150 with the mail address information, and determines the reading of the first and last name information from the reading candidate of the first and last name information.

  The control unit 110 controls the entire information processing apparatus 100 by transmitting a control signal to each unit and controlling the processing of each unit, or calculating based on information from each unit.

  The kanji reading information unit 124 will be described with reference to FIG. As shown in FIG. 2, the kanji reading information unit 124 stores kanji and roman characters representing the reading in association with each other. In order to reduce the storage capacity, the kanji reading information unit 124 stores not only all readings for kanji, but only readings when kanji are used for surnames and first names. When there are a plurality of readings, they are stored in the descending order of frequency used for first and last names. This is because, for example, in the case of “child”, the frequency of reading when used as first and last names is higher in “ko” than in “shi”.

  The contents of the kanji reading information unit 124 are not limited to the information shown in FIG. 2 as long as the information represents kanji and its reading. For example, the kanji reading information unit 124 may be information that stores a reading for a surname and a reading for a name separately for one kanji. In addition, the kanji reading information unit 124 may be information in which frequency information used for reading a surname or first name is stored in advance. The kanji reading information unit 124 has different reading frequencies at positions used for first and last names (for example, “good” for the first character of the first name and “mi” for the second and subsequent characters) )), Information that further stores reading frequency information according to the position used for the first and last names may be used. The kanji reading information section 124 may be information that stores readings in word units of surnames and first names, instead of information for each kanji.

  With reference to FIG. 3, a control structure of a program executed when control unit 110 constituting information processing apparatus 100 according to the present embodiment generates reading candidates will be described.

  In step (hereinafter, step is abbreviated as S) 100, control unit 110 determines whether or not image information transmitted from input device 200 has been received. If received (YES in S100), the process proceeds to S102. Otherwise (NO in S100), the process returns to S100.

  In S102, control unit 110 transmits a control signal to character recognition unit 130 so that the received image information is recognized as character information.

  In S104, control unit 110 transmits a control signal to information acquisition unit 140 so as to acquire first and last name information from the recognized character information. For example, the control unit 110 transmits a control signal so as to acquire a character string in which characters frequently used for first and last names are continuous as first and last name information. In addition, the acquisition method of full name information is not limited to this. For example, when receiving image information of a business card, the control unit 110 transmits a control signal so as to acquire, as first and last name information, a character string of the largest size existing near the center among the received image information of the business card. It may be.

  In S106, control unit 110 transmits a control signal to information acquisition unit 140 so as to acquire mail address information from the recognized character information. For example, the control unit 110 transmits a control signal so as to acquire, as mail address information, a character string obtained by removing a keyword such as “E-mail:” from a series of character strings including “@” in the received image information. Send. In addition, the acquisition method of e-mail address information is not limited to this. For example, the control unit 110 may transmit a control signal so as to acquire a character string including a keyword such as “E-mail:” as mail address information.

  In S108, control unit 110 transmits a control signal to information acquisition unit 140 so as to divide the acquired first and last name information into last name information and first name information. For example, the control unit 110 transmits a control signal so as to divide information described before the blank character of the first name and last name information as last name information and information described after the blank character as first name information. If there is no blank character, the dividing position may be near the center of the character string of the surname information, or before the kanji that is used more frequently than the surname. If a plurality of positions to be divided are considered, the following steps may be performed for a plurality of candidates, and the best matching may be determined by a candidate determination process described later.

  In S110, control unit 110 stores last name information and first name information in storage unit 120. In S112, control unit 110 reads one character of the surname information from storage unit 120.

  At S <b> 114, control unit 110 transmits a control signal to generation unit 150 so as to search for reading of one character of the read surname information using kanji reading information unit 124.

  In S116, control unit 110 determines whether or not all character readings included in the surname information have been searched. If all character readings are searched (YES in S116), the process proceeds to S118. Otherwise (NO in S116), the process returns to S112.

  At S118, control unit 110 transmits a control signal to generation unit 150 so as to generate a candidate for reading last name information (hereinafter also referred to as a surname candidate) by combining the retrieved readings. When there are a plurality of retrieved readings, the control unit 110 generates surname candidates for all combinations.

  In S120, control unit 110 reads one character of the name information from storage unit 120. In S122, control unit 110 transmits a control signal to generation unit 150 so as to search for one character of the read name information using kanji reading information unit 124.

  In S124, control unit 110 determines whether or not all character readings included in the name information have been searched. If all character readings are searched (YES in S124), the process proceeds to S126. Otherwise (NO in S124), the process returns to S120.

  In S126, control unit 110 transmits a control signal to generation unit 150 so as to generate a candidate for reading name information (hereinafter also referred to as a name candidate) by combining the retrieved readings. When there are a plurality of retrieved readings, the control unit 110 generates name candidates for all combinations.

  With reference to FIG. 4, a control structure of a program executed when control unit 110 configuring information processing apparatus 100 according to the present embodiment determines reading of first name and last name information will be described.

  In S200, control unit 110 transmits a control signal to determination unit 160 so as to acquire personal identification information from the mail address information. The personal identification information is a character string for identifying an individual included in the mail address information. For example, the control unit 110 acquires a character string before “@” included in the mail address information as personal identification information.

  In S202, control unit 110 transmits a control signal to determination unit 160 so as to divide the acquired personal identification information. For example, the control unit 110 transmits a control signal so as to divide the personal identification information before and after the division delimiter. The division delimiter is a character used to delimit characters included in the mail address, and examples thereof include “_”, “.”, “-”, And numbers. In S200 and S202, the personal identification information is divided after obtaining the personal identification information from the mail address information. However, the personal identification information may be obtained after the mail address information is divided.

  In S204, control unit 110 stores the divided personal identification information in storage unit 120. In S206, control unit 110 reads one of the divided pieces of personal identification information from storage unit 120. In S208, control unit 110 reads a surname candidate from storage unit 120.

  In S300, control unit 110 calculates the degree of coincidence between the read personal identification information and the surname candidate. Details of this process will be described later.

  In S210, control unit 110 determines whether or not the matching degrees of all name candidates have been calculated. When the matching degrees of all name candidates are calculated (YES in S210), the process proceeds to S212. Otherwise (NO in S210), the process returns to S208.

  In S212, control unit 110 reads name candidates from storage unit 120. In S300, control unit 110 calculates the degree of coincidence between the read personal identification information and the name candidate. In S214, control unit 110 determines whether or not the matching degrees of all name candidates have been calculated. When the matching degrees of all name candidates are calculated (YES in S214), the process proceeds to S216. Otherwise (NO in S214), the process returns to S212.

  In S216, control unit 110 determines whether or not all the matching degrees of the divided personal identification information have been calculated. When all the matching degrees are calculated (YES in S216), the process proceeds to S400. Otherwise (NO in S216), the process returns to S206.

  At S400, control unit 110 transmits a control signal to determination unit 160 so as to determine reading of the first and last names. Details of this process will be described later.

  At S218, control unit 110 converts the determined reading candidate from Roman characters to Hiragana using the correspondence information between Roman characters and Hiragana stored in storage unit 120 as shown in FIG. In addition, although the Roman character in FIG. 5 is described by the Hebon type | formula, a ceremonial type may be sufficient. In S220, control unit 110 outputs the converted reading to output device 300.

  Referring to FIG. 6, a control structure of a program executed when control unit 110 constituting information processing apparatus 100 according to the present embodiment calculates the degree of coincidence between personal identification information and a surname candidate or first name candidate. explain. In the description of this structure, for the sake of convenience, the surname candidate and the first name candidate are simply described as candidates without distinction.

  In S302, control unit 110 counts the number of characters that match the personal identification information and the candidate as the number of matches. Note that the controller 110 does not count the number of matches when the number of consecutively matched characters is 1 or less except for the beginning and the end.

  In S304, control unit 110 determines whether the personal identification information or the first or last character of the candidate matches. If they match (YES in S304), the process proceeds to S306. Otherwise (NO in S304), the process proceeds to S308.

  In S306, control unit 110 adds 1 to the number of matches. In S308, control unit 110 stores a position (hereinafter referred to as a matching position) in the personal identification information of a character that matches the personal identification information and the candidate.

  With reference to FIG. 7, a control structure of a program executed when control unit 110 configuring information processing apparatus 100 according to the present embodiment determines reading of first and last names will be described.

  In S402, control unit 110 determines whether there is a surname candidate whose ratio of the number of matches to the number of characters of the personal identification information is greater than a predetermined ratio. For example, the control unit 110 determines whether there is a surname candidate whose rate of matching is greater than 0 percent (that is, whether there is a surname candidate including a matching character). Note that the ratio of the number of matches is not limited to 0 percent. In addition, the control unit 110 may determine whether there is a candidate whose ratio of the number of matches to the number of characters of the surname candidate is greater than a predetermined ratio. If there is a surname candidate larger than the predetermined ratio (YES in S402), the process proceeds to S404. Otherwise (NO in S402), the process proceeds to S406.

  In S404, control unit 110 transmits a control signal to determination unit 160 so as to determine the surname candidate having the largest number of matches as the reading of the last name information.

  In S406, control unit 110 transmits a control signal to determining unit 160 so as to determine a surname candidate that is frequently used as a surname as a reading of the last name information. That is, since the kanji reading information unit 124 stores readings in the order of frequency used for the first and last names, the control unit 110 determines the first reading stored in the kanji reading information unit 124 as the reading of the last name information. Then, a control signal is transmitted to the determination unit 160.

  In S408, control unit 110 extracts name candidates that do not overlap with the surname candidates whose matching positions are determined. In S410, control unit 110 determines whether or not there are name candidates in which the ratio of the number of matches to the number of characters in the personal identification information is greater than a predetermined ratio among the extracted name candidates. For example, the control unit 110 determines whether or not there is a name candidate whose rate of matching is greater than 0 percent (that is, whether or not there is a name candidate including a matching character). Note that the ratio of the number of matches is not limited to 0 percent. In addition, the control unit 110 may determine whether or not there is a candidate whose ratio of the number of matches to the number of characters of the name candidate is greater than a predetermined ratio. If there are name candidates greater than a predetermined ratio (YES in S410), the process proceeds to S412. Otherwise (NO in S410), the process proceeds to S414.

  In S412, control unit 110 transmits a control signal to determination unit 160 so as to determine reading of name information for the name candidate having the largest number of matches.

  In S414, control unit 110 transmits a control signal to determination unit 160 so as to determine reading of name information for name candidates that are frequently used for names. That is, since the kanji reading information unit 124 stores the readings in the order in which the readings are used in the first and last names, the control unit 110 determines the reading stored first in the kanji reading information unit 124 as the reading of the name information. Then, a control signal is transmitted to the determination unit 160.

  At S416, control unit 110 transmits a control signal to determination unit 160 so as to determine the reading of the name information by combining the reading of the last name information and the reading of the name information.

  An operation of information processing apparatus 100 according to the present embodiment based on the above-described structure and flowchart will be described.

  First, the operation when the business card shown in FIG. 8 is recognized will be described. This business card has “Kumida Miko”, which is a kanji for first and last names, and an email address “ykakuta@xyz.com” on the front. Therefore, just by inputting image information of the image on the surface of the business card to the input device 200, “Miko Tsunoda” and “ykakuta@xyz.com” are input to the information processing device 100.

  When image information is received from the input device 200 (YES in S100), “XYZ”, “General Affairs Department”, “Miko Tsunoda”, “XYZ Co., Ltd.”, “Abano-ku, Osaka City "," Telephone (06) 1234-5678 "," FAX (06) 1234-9999 "," E-mail: ykakuta@xyz.com "are recognized (S102). Among these character information, “Miko Tsunoda” in which characters frequently used in first and last names are consecutive is acquired as first name and last name information (S104), and “E-mail: ykakuta @ including“ E-mail ”and“ @ ”is acquired. Of “xyz.com”, “ykakuta@xyz.com” excluding “E-mail:” is acquired as mail address information (S106).

  The first and last name information “Miko Kakuda” is divided into last name information “Kakuda” written before the blank character and first name information “Miko” written after the blank character (S108) and stored in the storage unit 120. (S110).

  Among the “Kakuda”, “Kaku” and “Tana” readings are sequentially searched (S112 to S116), and as shown in FIG. 9A, a plurality of surname candidates such as “tsunota” and “tsunoda” are generated. (S118). Similarly for “Meiko”, name candidates “miko” and “yoshiko” are generated as shown in FIG. 9B (S120 to S126).

  Of the mail address information “ykakuta@xyz.com”, the character string “ykakuta” before “@” is acquired as personal identification information (S200). This “ykakuta” is the first letter of “yoshiko” in “Roman” for “Yoshiko” in “Kutata Yoshiko”, which is the reading of “Miko Tsunoda”, and “Kuta” in Roman It is a combination of “kakuta” expressed in.

  The degree of coincidence between “ykakuta”, the surname candidate and the first name candidate is calculated (S300). As shown in FIG. 9A, in the case of the surname candidate “tsunota”, two characters of “ta” that match “ykakuta” are counted as the number of matches (S302), and the last character “a” is Since there is a match, the number of matches is incremented by one (S306), and the total number of matches is “3”. Further, it is stored that the matching position is from the sixth character to the seventh character in “ykakuta” (S308). When such processing is performed for all surname candidates (YES in S210), the matching degree of first name candidates is calculated in the same manner (S212, S300).

  When the calculation results as shown in FIGS. 9A and 9B are obtained in this way, a process for determining reading of the first and last names is performed (S400).

  The reading of the surname is determined for the largest surname candidate “kakuta” whose total number of matches is “7” (YES in S402, S404). Since the matching position of the surname candidate “kakuta” is “2-7 (second to seventh characters in“ ykakuta ”)”, the matching position does not overlap “1-1 (first character in“ ykakuta ”)” “Yoshiko” is extracted (S408). Since the extracted name candidate is only “yoshiko” and the candidate having the largest number of matches, the reading of the name is determined as “yoshiko” (YES in S410, S412). The reading of the name information is determined as “kakuta yoshiko” (S416). “Kakuta yoshiko” is converted into “Kakuta Yoshiko” (S218) and output to the output device 300 (S220).

  In this way, the surname candidate and the surname candidate are generated separately, and the reading of the surname and the surname is determined so that the match positions of the surname candidate and the surname candidate and the personal identification information “ykakuta” do not overlap. Therefore, it is possible to suppress duplication of the surname candidates and the characters matching the first name candidates in the personal identification information “ykakuta”. This allows the correct reading of the surname and first name to be determined.

  Second, the operation when the business card shown in FIG. 10 is recognized will be described. The same processing as when the business card shown in FIG. 8 is recognized (S100 to S124), surname candidates and name candidates for the name information “Miko Tsunoda” are generated (S118, S126), and the personal identification information “yo-” kakuta "is acquired (S200). In addition, this "yo-kakuta" expresses "yo" of "yoshiko" which expressed "Yoshiko" of "Kutata Yoshiko" which is the reading of "Miko Tsunoda" in Roman letters and "Kuta" in Roman letters. "Kakuta" is combined with "-".

  “Yo-kakuta” is divided into “yo” and “kakuta” before and after “-” (S202). As shown in FIG. 11, the degree of coincidence between “yo” and each surname candidate and each first name candidate is calculated (S300). As shown in FIG. 12, the degree of coincidence between “kakuta” and each surname candidate and each first name candidate is calculated (S216, S300).

  As shown in FIG. 11A, the first personal identification information “yo” and each surname candidate have no matching character, and the total number of matches of all candidates is calculated as zero.

  As shown in FIG. 11B, the first personal identification information “yo” and the name candidate “miko” are calculated such that “o” at the end matches and the total number of matches is “2” ( S302, S306). The first personal identification information “yo” and the name candidate “yoshiko” are calculated such that the leading “yo” matches and the total number of matches is “3” (S302, S306).

  The degree of coincidence is similarly calculated for the second personal identification information “kakuta” (S400), resulting in a calculation result as shown in FIG.

  The surname candidate “kakuta” that is “8”, which completely matches the second personal identification information “kakuta” and has the largest total number of matches, is determined as the last name reading (YES in S402, S404). The first personal identification information “kakuta” does not overlap with the first to sixth characters (S408). The name “yoshiko” with the largest total number of matches with the first personal identification information “yo” is the name. (YES in S410, S412).

  Third, the operation when the business card shown in FIG. 13 is recognized will be described. The same processing as when the business card shown in FIG. 8 is recognized (S100 to S124), surname candidates and first name candidates for the name information “Ken Tanaka” are generated (S118, S126), and the personal identification information “ktanaka”. Is acquired (S200). This “ktanaka” is the first letter of “ken”, which is “Ken” in “Tanaka Ken”, which is the reading of “Tanaka Ken”, and “Tanaka” in Roman letters. It is a combination of "tanaka".

  FIG. 14 shows the calculation result of the degree of coincidence. In this case, the surname candidate “tanaka” having the largest total number of matches is determined to read the surname (YES in S402, S404). The name candidate “takeshi” is “2” having the same total number of matches as the other name candidates “ken”, but the match position “2-3” of the name candidate “takeshi” (from the second character of “ktanaka” is 3 "Character") is duplicated with the matching position "2-7 (2nd to 7th characters of" ktanaka ")" of the determined surname candidate "tanaka", so "takeshi" is acquired as a name candidate Instead, “ken” is acquired (S408). Thus, since the reading of the surname candidate and the surname candidate is determined so that the matching positions in the personal identification information do not overlap, it is possible to determine the correct reading of the surname and the surname.

  Fourth, the operation when the business card shown in FIG. 15 is recognized will be described. The same processing as when the business card shown in FIG. 8 is recognized (S100 to S124), surname candidates and name candidates for the name information “Ken Tanaka” are generated (S118, S126), and the personal identification information “ke -tanaka "is acquired (S200). This “ke-tanaka” is a “ken” that represents “Ken” of “Tanaka Ken” in Roman letters, “tanaka” that represents “Tanaka” in Roman letters, and “-”. It is a combination.

  “Ke-tanaka” is divided into “ke” and “tanaka” before and after “-” (S202). The calculation results of the degree of coincidence are shown in FIGS.

  As shown in FIG. 16A, the first personal identification information “ke” and each surname candidate do not have a matching character, and the total number of matches of all candidates is calculated as zero.

  As shown in FIG. 16B, the first personal identification information “ke” and the name candidate “ken” match the two characters “ke” and the first character “k”. The total number of matches is calculated as “3”. The first personal identification information “ke” and the name candidate “takeshi” are calculated as “2” because the two characters “ke” in the middle of “takeshi” match. Thus, even if the number of matching characters is the same, “ken” that matches the first character is calculated to have a larger total number of matches than “takeshi” that matches the middle character. As a result, it is possible to prioritize “ken”, which is an accurate reading, and accurately determine the reading of the first and last names.

  The degree of coincidence is similarly calculated for the second personal identification information “tanaka”, and the calculation result as shown in FIG. 17 is obtained.

  The surname candidate “tanaka”, which is “8”, which completely matches the second personal identification information “tanaka” and has the largest total number of matches, is determined as the last name reading (YES in S402, S404). The matching position does not overlap with the determined surname candidate “tanaka” (S 408), “ken”, which is “3” having the largest number of matches with the first personal identification information “ke”, is determined as the reading of the first name. (YES in S410, S412).

  In the description of the first to fourth operations, the case has been described where there are candidates for which the ratio of the number of matches is greater than a predetermined ratio (YES in S402, YES in S410). On the other hand, if there is no candidate whose ratio of the number of matches is larger than a predetermined ratio (NO in S402, NO in S410), a candidate that is frequently used for the first and last names is read out of each candidate. (S406, S414). That is, since the kanji reading information unit 124 stores readings in the descending order of frequency used for first and last names, the candidate that combines the readings stored first in the kanji reading information unit 124 is determined to be reading. For this reason, even if there is no candidate whose rate of coincidence is greater than a predetermined rate, it is possible to determine a candidate that is highly likely to be accurate reading.

  As described above, according to the information processing apparatus according to the present embodiment, the name information and the personal identification information of the mail address are acquired from the image information of the business card. The correct reading of the name is determined as a result of collating the kanji reading candidates representing the name with the personal identification information. Therefore, it is possible to efficiently determine the reading of first and last names.

  In the above-described embodiment, the kanji reading information unit 124 stores kanji and a roman character representing the reading in association with each other as shown in FIG. 2, but as shown in FIG. It may be information that represents kanji and their readings in hiragana. In this case, first name surname and first name candidate are generated as hiragana, and the Roman characters contained in the personal identification information are converted into hiragana, and the degree of matching of the hiragana between each candidate and the personal identification information is calculated. Can be determined.

  Further, in the above-described embodiment, if there is no candidate whose ratio of the number of matches is greater than a predetermined ratio (NO in S402, NO in S410), a candidate that is frequently used for a surname or first name is read. (S406, S414). On the other hand, for example, a character obtained by converting a Roman character included in the personal identification information into a kana may be determined as a reading of the surname. This makes it possible to determine the reading of first and last names, and to determine a reading that is highly likely to be an accurate reading even when an accurate reading candidate is not generated.

  The embodiment disclosed this time should be considered as illustrative in all points and not restrictive. The scope of the present invention is defined by the terms of the claims, rather than the description above, and is intended to include any modifications within the scope and meaning equivalent to the terms of the claims.

It is a block diagram which shows the structure of the information processing apparatus 100 which concerns on embodiment of this invention. It is a figure (the 1) which shows the information memorize | stored in the information processing apparatus 100 which concerns on embodiment of this invention. It is a flowchart (the 1) which shows the control structure of the control part which comprises the information processing apparatus 100 which concerns on embodiment of this invention. It is a flowchart (the 2) which shows the control structure of the control part which comprises the information processing apparatus 100 which concerns on embodiment of this invention. It is FIG. (2) which shows the information memorize | stored in the information processing apparatus 100 which concerns on embodiment of this invention. It is a flowchart (the 3) which shows the control structure of the control part which comprises the information processing apparatus 100 which concerns on embodiment of this invention. It is a flowchart (the 4) which shows the control structure of the control part which comprises the information processing apparatus 100 which concerns on embodiment of this invention. It is a figure (the 1) showing the business card which the information processing apparatus 100 which concerns on embodiment of this invention acquires as image information. It is a figure (the 1) which shows the result of the coincidence degree which the information processing apparatus 100 which concerns on embodiment of this invention calculated. It is FIG. (The 2) showing the business card which the information processing apparatus 100 which concerns on embodiment of this invention acquires as image information. It is FIG. (2) which shows the result of the coincidence degree which the information processing apparatus 100 which concerns on embodiment of this invention calculated. It is a figure (the 3) which shows the result of the coincidence degree which the information processing apparatus 100 which concerns on embodiment of this invention calculated. It is FIG. (The 3) showing the business card which the information processing apparatus 100 which concerns on embodiment of this invention acquires as image information. It is FIG. (The 4) which shows the result of the coincidence degree which the information processing apparatus 100 which concerns on embodiment of this invention calculated. It is FIG. (The 4) showing the business card which the information processing apparatus 100 which concerns on embodiment of this invention acquires as image information. It is FIG. (5) which shows the result of the coincidence degree which the information processing apparatus 100 which concerns on embodiment of this invention calculated. It is FIG. (6) which shows the result of the coincidence degree which the information processing apparatus 100 which concerns on embodiment of this invention calculated. It is FIG. (3) which shows the information memorize | stored in the information processing apparatus 100 which concerns on embodiment of this invention.

Explanation of symbols

  DESCRIPTION OF SYMBOLS 100 Information processing apparatus, 110 Control part, 120 Storage part, 122 Character recognition information part, 124 Kanji reading information part, 130 Character recognition part, 140 Information income part, 150 Generation part, 160 Determination part.

Claims (18)

Means for preliminarily storing first information associating kanji and phonetic characters representing the reading of the kanji;
An acquisition means for acquiring kanji representing an individual's first and last name and second information included in the individual's email address;
Generating means for generating candidates for reading kanji representing the first and last names based on the first information;
An information processing apparatus comprising: a determination unit configured to determine reading of a kanji character representing the first and last name based on a result of collating the second information with the candidate.   The information processing apparatus according to claim 1, wherein the determining unit includes a unit for determining the reading based on a number of characters that match the second information and the candidate.   The information processing apparatus according to claim 2, wherein the determining means includes means for determining a candidate having the largest number of matching characters as the reading.   The information processing apparatus according to claim 1, wherein the determination unit includes a unit for determining the reading based on a character position that matches the second information and the candidate.   The information processing apparatus according to claim 4, wherein the determining unit includes a unit that prioritizes a candidate whose leading character matches the leading character of the second information and determines the reading.   The information processing apparatus according to claim 4, wherein the determination unit includes a unit that prioritizes a candidate whose last character matches the last character of the second information and determines the reading. The generating means includes means for generating a surname candidate that is a reading of a kanji representing a surname and a name candidate that is a reading of a kanji representing a first name,
In the second information, the determination unit may be configured such that a character position matching the last name candidate and the second information is different from a character position matching the first name candidate and the second information in the second information. The information processing apparatus according to claim 4, further comprising means for determining the reading.   The determination means has a ratio of the number of characters matching the second information and the candidate with respect to the number of characters of the second information and the candidate is smaller than a predetermined ratio for all candidates. The information processing apparatus according to claim 1, further comprising means for determining a candidate satisfying a predetermined condition as the reading.   The information processing apparatus according to claim 8, wherein the determination unit includes a unit for determining the candidate that is frequently used for first and last names as the reading.   The determination means has a ratio of the number of characters matching the second information and the candidate with respect to the number of characters of the second information and the candidate is smaller than a predetermined ratio for all candidates. The information processing apparatus according to claim 1, further comprising means for determining the reading based on the second information. The information processing apparatus further includes means for generating a pseudonym corresponding to the second information,
The generating means includes means for generating the candidate with a pseudonym,
The information processing apparatus according to claim 1, wherein the determining unit includes a unit for determining the reading based on a result of collating the corresponding kana and the candidate. The second information includes a romaji,
The information processing apparatus according to claim 1, wherein the generation unit includes a unit for generating the candidate in Roman letters.   The information processing apparatus according to claim 1, wherein the second information is a character for identifying an individual included in the mail address.   The information processing apparatus according to claim 1, wherein the acquisition unit includes a unit for acquiring character information by recognizing captured image information.   The information processing apparatus according to claim 1, wherein the first information is information in which the kanji is associated with reading when the kanji is used as a first name and last name. An information processing method performed by an information processing apparatus including a storage unit, an acquisition unit, a generation unit, and a determination unit,
Storing, using the storage means, first information associating a kanji with a phonetic character representing a reading of the kanji;
Using the acquisition means to acquire kanji representing an individual's first and last name and second information included in the individual's email address;
Generating a kanji reading candidate representing the first and last names based on the first information using the generating means;
An information processing method including: determining, using the determining means, a reading of a kanji character representing the first and last name based on a result of collating the second information with the candidate.   The program which implement | achieves the information processing method of Claim 16.   The computer-readable recording medium which recorded the program of Claim 17.

Images