JP2012194264A - Automatic transaction device, voice reading-aloud method, and voice reading-aloud program - Google Patents

Abstract

PROBLEM TO BE SOLVED: To restrict occurrence of inaccurate reading-aloud.SOLUTION: In an automatic transaction device 1, a phonogram string which expresses the phonetic of a target character string from the target character string containing Kanji characters is compared with a Kana character string which expresses the Furigana of the target character string and consists of voiceless sounds, voiced sounds to the voiceless sounds, and at least voiceless sounds among p-sounds, and any one of the phonogram string and the Kana character string is selected in order to read aloud the target character string based on the comparison result. Thereby, occurrence of inaccurate reading-aloud due to unnatural pronunciation of the target character string aimed at reading-aloud is reduced.

Description

  The present invention relates to an automatic transaction apparatus, a speech reading method, and a speech reading program.

  In financial institutions such as banks, the national bank association provides a predetermined format for financial institutions nationwide. The format includes, for example, a code of a financial institution and each branch of the financial institution, a name represented by kanji and a kana for the kanji, and a standard communication protocol. In a financial institution, for example, for printing a passbook in an automatic transaction device (ATM (Automated Teller Machine)), kanji of a financial institution / branch included in this format and a kana corresponding to the kanji are used.

An ATM installed in a financial institution or the like has a function of performing guidance such as voice operation so that even a visually impaired user can operate the ATM (for example, see Patent Document 1).
Even in this case, the name of the financial institution and the branch issued during the guidance is the voice information synthesized from the kanji such as the name of the financial institution and the branch included in the format described above and the kana corresponding to the kanji. Read based on.

  Note that the hiragana of the format is composed of 50 large letters, that is, only a clear sound, a muddy sound for a clear sound, and a semi-voiced sound for a clear sound. In other words, all the small letters (a, i, u, e, o, o, u, u, u, ヮ, t) such as stuttering and prompting sounds are overwritten letters (a, i, u, e, o, ya, Yu, Yo, Wa, Tsu).

Japanese Unexamined Patent Publication No. 1-163894

However, there are the following problems when using speech information synthesized as described above.
When using names represented in Chinese characters of financial institutions and branches, processing for temporarily converting the Chinese characters to kana is performed. However, in such a process, the correct kana character corresponding to the kanji is not necessarily converted (for example, “Kunitachi” for “Kunitachi” in “National”). There was a problem that the names of branches were not read out correctly.

  Further, for such a problem, a method of synthesizing speech from a kana corresponding to a kanji representing the names of financial institutions and branches can be considered. However, since all the hiragana characters in the format are composed of large letters, some hiragana characters cannot be read out correctly from the voice information synthesized by speech (for example, for “Tokyo” in “Tokyo”). “Tokiyo”).

  The present invention has been made in view of these points, and an object thereof is to provide an automatic transaction apparatus, a speech reading method, and a speech reading program that suppress the occurrence of inaccurate reading.

  To achieve the above object, in an automatic transaction apparatus that performs voice guidance, conversion means for converting a target character string including kanji into a phonetic character string representing a phonetic of the target character string, and the converted phonetic character In order to read out the target character string based on a comparison result between a string and a kana character string composed of at least a clear sound of a clear sound, a muffled sound for a clear sound, and a semi-voiced sound for a clear sound. There is provided an automatic transaction apparatus having a comparison / determination means for selecting the character string.

  In order to achieve the above object, a speech reading method and a speech reading program similar to those of the automatic transaction apparatus are provided.

  By such an automatic transaction apparatus, a speech reading method and a speech reading program, the occurrence of inaccurate reading is suppressed.

It is a conceptual diagram of the automatic transaction apparatus which concerns on 1st Embodiment. It is a figure which shows the example of an external appearance of ATM which concerns on 2nd Embodiment. It is a figure which shows the hardware structural example of ATM which concerns on 2nd Embodiment. It is a block diagram showing the function with which ATM which concerns on 2nd Embodiment is provided. It is a figure which shows the example of the financial institution information with which ATM which concerns on 2nd Embodiment is provided. It is a figure which shows the voice guidance information with which ATM which concerns on 2nd Embodiment is provided. It is a flowchart which shows the audio | voice reading process procedure of ATM which concerns on 2nd Embodiment. It is a flowchart which shows the character string comparison processing procedure of ATM which concerns on 2nd Embodiment. It is FIG. (1) for demonstrating typically the comparison of the character string which concerns on 2nd Embodiment. It is FIG. (2) for demonstrating typically the comparison of the character string which concerns on 2nd Embodiment.

Hereinafter, embodiments will be described with reference to the drawings.
[First Embodiment]
FIG. 1 is a conceptual diagram of an automatic transaction apparatus according to the first embodiment.

The automatic transaction apparatus 1 can suppress the occurrence of inaccurate reading when reading a voice.
Such an automatic transaction apparatus 1 includes dictionary information storage means 2, character string information storage means 3, and waveform information storage means 4. Furthermore, a conversion unit 5, a comparison determination unit 6, a generation unit 7, and an audio output unit 8 are provided. Note that the conversion means 5, the comparison determination means 6, the generation means 7, and the voice output means 8 are executed by executing a voice reading program by a CPU (Central Processing Unit) (not shown) included in the automatic transaction apparatus 1. The processing function is realized.

  The dictionary information storage means 2 stores a dictionary to be referred to when the text information of the vocabulary subject to speech synthesis processing is parsed, and the notation, kana reading, accent information, part of speech information for each vocabulary in advance. Are registered in association with each other.

  The character string information storage means 3 represents a kana character set in advance for a target character string including kanji characters, and preliminarily stores information on a kana character string composed of at least a clear sound of a clear sound, a muddy sound for a clear sound, and a semi-voice sound for a clear sound. Storing. The character string information storage means 3 includes, for example, a name (target character string) including a kanji of a financial institution and a branch to be read out in a format related to a financial institution of the National Bankers Association, and a kana character representing a kana corresponding to the name Hold the column. As mentioned above, the hiragana characters in the format are all small letters (a, i, u, e, o, ya, u, o, ヮ, tsu) such as stuttering and prompting sounds. A, i, u, d, o, ya, yu, yo, wa, tsu). Therefore, in this format, for example, “Azamino (kana character string)” for the branch name “Azamino (target character string)” and “Tranomone (kana character) for the branch name“ Toranomon (target character string) ”. Column) ”,“ Kyoto (kana character string) ”, etc., for the branch name“ Kyoto (target character string) ”.

  The waveform information storage means 4 stores speech waveform information corresponding to kana characters included in the phonetic character string and the kana character string, respectively. Note that the phonetic character string is composed of, for example, kana characters and special characters (colon, comma, quotes, percent symbols, etc.) that are phonetic symbols that indicate the reading and accent of the kana characters.

The conversion means 5 converts the target character string including the kanji into a phonetic character string representing the phonetic of the target character string.
The comparison determination means 6 represents a phonetic character string converted by the conversion means 5 and a kana character of the target character string, and a kana character string composed of at least a clear sound, a muddy sound for a clear sound, and a semi-voice sound for a clear sound, One of the character strings is selected to read out the target character string based on the comparison result of.

The generation unit 7 generates voice information of the target character string based on the character string selected by the comparison determination unit 6 based on the waveform information storage unit 4.
The voice output unit 8 reads out the target character string based on the voice information generated by the generation unit 7.

A voice reading method of the automatic transaction apparatus 1 having such a configuration will be described.
First, the conversion means 5 converts a target character string including a kanji character to be read out into a phonetic character string representing a phonetic of the target character string based on the dictionary information storage means 2.

  The comparison determination means 6 compares the converted phonetic character string with the kana character string representing the kana of the target character string to be read out stored in the character string information storage means 3. At the time of comparison, the phonetic character string and the kana character string are compared from the beginning to the end only for the kana characters one by one based on the matching condition. As a specific example of the matching condition, for example, when the kana character included in the phonetic character string to be compared is a large character corresponding to the small character of the kana character string, If the long sound and the kana character string “o” or “u” are to be compared, they match each other.

  The comparison determination means 6 selects either the phonetic character string or the kana character string to read out the target character string to be read out based on the result of such comparison. For example, when the phonetic character string and the kana character string match as a result of the comparison, the phonetic character string is selected for reading out the target character string. On the other hand, if they do not match, a kana character string is selected.

  Based on the speech waveform stored in the waveform information storage unit 4, the generation unit 7 generates speech information of the target character string to be read out from the character string selected by the comparison determination unit 6, and the speech output unit 8 Based on the voice information, the target character string to be read out is read out.

  Thus, in the automatic transaction apparatus 1, the phonetic character string which represents the phonetic sound of the said target character string from the target character string containing a kanji character, the kana of the target character string, the muffled sound for the clear sound and the clear sound, and the semi-voice sound for the clear sound In order to compare at least a kana character string composed of a clear sound and read out the target character string based on the comparison result, either the phonetic character string or the kana character string is selected. That is, in this way, when the phonetic character string and the kana character string of the target character string match, the phonetic character string corresponds to the phonetic kana preset in the target character string, If the phonetic character string does not match the kana character string, the phonetic character string does not correspond to the phonetic name preset in the target character string. For this reason, when the phonetic character string and the kana character string match, the phonetic character string is selected, and when they do not match, the kana character string is selected and the target character string is read correctly. As a result, it is possible to prevent the target character string to be read out from being read out inaccurately with an unnatural pronunciation.

[Second Embodiment]
In the second embodiment, the first embodiment will be described more specifically.
FIG. 2 is a diagram illustrating an external appearance example of an ATM according to the second embodiment.

  The ATM 10 has a display means 30a for displaying a screen relating to transactions, a card processing unit having a card insertion / discharge port 40a on the front surface, a bankbook processing unit having a bankbook insertion / discharge port 50a, and a banknote processing having a banknote insertion / discharge port 60a. The unit has a coin processing unit having a coin insertion / discharge port 70a. Furthermore, ATM10 is used by the user with a visual impairment, and has the receiver 80 which carries out the voice guidance of the instruction | indication of the operation procedure and operation input to the said user.

Next, the hardware configuration of the ATM 10 having such an appearance will be described.
FIG. 3 is a diagram illustrating a hardware configuration example of the ATM according to the second embodiment.
The ATM 10 includes a control unit 20 that executes processing related to financial transactions such as transfer, deposit, remittance, and bookkeeping performed in the ATM 10, and a display unit 30 that receives an operation input from a user and displays an output result. Further, the ATM 10 includes a card processing unit 40 having a card insertion / discharge port 40a, a bankbook processing unit 50 having a bankbook insertion / discharge port 50a, and a banknote processing unit 60 having a banknote insertion / discharge port 60a, as described in FIG. , And a coin processing unit 70 having a coin insertion / discharge port 70a, and a receiver 80 used by a visually impaired user.

  The control unit 20 includes a CPU 20a, a RAM (Random Access Memory) 20b, an HDD (Hard Disk Drive) 20c, a graphic interface 20d, a host communication control means 20e, and an input / output interface 20f, each of which is a bus 20g. Are connected to each other.

The CPU 20a comprehensively controls the entire ATM 10 by executing various programs stored in a storage medium such as the HDD 20c.
The RAM 20b temporarily stores at least one of an OS (Operating System) and programs to be executed by the CPU 20a. The RAM 20b stores various data necessary for processing by the CPU 20a.

  The HDD 20c stores the OS and application programs on the ATM 10. The HDD 20c stores various data necessary for processing by the CPU 20a.

  A display unit 30 is connected to the graphic interface 20d. The graphic interface 20d causes the display unit 30a of the display unit 30 to display an image in accordance with a command from the CPU 20a. In addition, the graphic interface 20 d acquires information corresponding to the operation input detected by the input detection unit 30 b of the display unit 30.

  The host communication control means 20e is connected to, for example, a LAN (Local Area Network) so as to be communicable with a host computer (not shown) of a management center that manages ATMs installed in various places. Can communicate.

  A card processing unit 40, a passbook processing unit 50, a banknote processing unit 60, a coin processing unit 70, and a receiver 80 are connected to the input / output interface 20f. The input / output interface 20f, the card processing unit 40, the bankbook processing unit 50, the banknote processing unit 60, the coin processing unit 70, and the receiver 80 exchange signals with the CPU 20a via the bus 20g.

The display unit 30 is, for example, a touch panel including a display unit 30a and an input detection unit 30b.
The display means 30a is, for example, an LCD (Liquid Crystal Display) that displays a screen based on screen information in accordance with the processing of the ATM 10.

  The input detection unit 30b detects a user's touch on the screen displayed by the display unit 30a. Specifically, the position information of the operation input is detected from an infrared LED (Light Emitting Diode) in which the infrared beam is blocked by the operation input by the user and the phototransistor, and the position information is transmitted to the control unit 20. To do. Note that the CPU 20a of the control unit 20 identifies the processing content displayed on the screen corresponding to the detected touch position, and executes the processing.

  The card processing unit 40 records a processing result of a predetermined process in accordance with the execution instruction of the card inserted by the user through the card insertion / discharge port 40a. Further, the card processing unit 40 discharges the card for which the processing result has been recorded from the card insertion / discharge port 40a and returns it to the user.

  The passbook processing unit 50 transports the user's passbook inserted from the passbook insertion / discharge port 50a to a predetermined position, and records the processing result of the predetermined processing in accordance with an execution instruction from the user. Further, the passbook processing unit 50 conveys the passbook for which processing result recording has been completed, discharges it from the passbook insertion / discharge port 50a, and returns it to the user.

  The banknote processing unit 60 and the coin processing unit 70 receive banknotes and coins inserted from the banknote insertion / discharge port 60a and the coin insertion / discharge port 70a, and count the banknotes and coins, respectively. Moreover, according to a predetermined | prescribed execution instruction | indication of a user, the banknote and coin of the amount according to a process can be discharge | released from the banknote insertion / discharge port 60a and the coin insertion / discharge port 70a.

  The receiver 80 includes a speaker for outputting voice guidance, and is arranged with push-phone type numeric keypad buttons. The visually impaired user lifts the handset 80, is guided by voice guidance heard from the speaker of the handset 80, operates the buttons on the pushphone, and executes transaction processing according to the buttons on the pushphone, The amount of money can be entered.

  The card processing unit 40, the passbook processing unit 50, the banknote processing unit 60, and the coin processing unit 70 are controlled by the CPU 20a of the control unit 20 in accordance with an operation input through the display unit 30 by the user. Further, in response to the user's button operation on the handset on the handset 80, a predetermined process is executed by the CPU 20a of the control unit 20, and the card processing unit 40, the passbook processing unit 50, the banknote processing unit 60, and the coin processing unit 70 are Each is controlled.

Next, the control function provided in the ATM 10 having such a hardware configuration will be described.
FIG. 4 is a block diagram showing functions provided in the ATM according to the second embodiment. FIG. 5 is a diagram illustrating an example of financial institution information included in the ATM according to the second embodiment, and FIG. 6 is a diagram illustrating voice guidance information included in the ATM according to the second embodiment. In particular, FIG. 6 shows an example of voice guidance information. In this case, after the information related to the transfer is input, the utterance information for confirming the transfer content is shown.

  The ATM 10 is capable of suppressing the occurrence of inaccurate speech reading when performing voice guidance for a visually impaired user and executing guidance related to operations.

First, the ATM 10 includes a financial institution information storage unit 21, an output character string information storage unit 22, and a voice guidance information storage unit 23.
The financial institution information storage means 21 is a format provided by the National Bank Association, and stores information related to financial institutions. For example, as shown in FIG. 5, information on a financial institution code, a transfer name of a financial institution associated with the financial institution code, and a name displayed in kanji are stored. Similarly, for the branch of the financial institution, information on the branch code and the name of the branch represented in kana and kanji is stored. As mentioned above, all of the small letters (a, i, u, e, o, ya, u, yo, ヮ, tsu) such as stuttering and prompting sound are used for the format provided by the National Bank Association. Expressed in large letters (a, i, u, d, o, ya, yu, yo, wa, tsu). Specifically, in the case of FIG. 5, for the financial institution “Kyushu Bank”, the information of the financial institution code “0111” and the names of the financial institutions “Kyushu Yuginko (Kana)” and “Kyushu Bank (Kanji)” is stored. Has been. As for the branch name “national branch”, information of branch code “0024”, branch names “Kunitachi Shiten (shake pseudonym)” and “national branch (kanji)” is stored.

  The output character string information storage unit 22 stores information on a phonetic character string or a kana character string corresponding to the names of financial institutions and branches to be read selected by the comparison determination unit 27 described later from the comparison result. In addition, information of a phonetic character string corresponding to how to read the numbers and a kana character string representing the names of the client and the recipient are stored by the conversion means 26c described later.

  The voice guidance information storage means 23 holds utterance information representing the utterance contents voice-guided to the visually impaired user from the ATM 10. For example, before the transfer is executed, in order to confirm the transfer contents, as shown in FIG. 6, utterance information is held in association with each identification number. Specifically, the identification number “011” holds speech information “Now processing. Please wait.” Further, the identification number “003” holds the utterance information “Branch name is (b. Branch name)”. Note that “(b. Branch name)” is merged with the branch name voice-synthesized by the ATM 10.

  Further, the ATM 10 includes a card reading unit 24, a searching unit 25, a voice synthesizing unit 26, a comparison / determination unit 27, an utterance information merging unit 28, a numeric information notifying unit 29, an input receiving unit 31, an utterance information notifying unit 32, and an audio output control. Means 33 are provided.

  The card reading unit 24 reads information recorded on the transfer card inserted by the user into the card insertion / discharge port 40 a of the card processing unit 40. The transfer card is one in which transfer information such as a payee and its bank account, the name and telephone number of the client is recorded magnetically in advance. Note that the names of the recipient and the client are recorded as kana character strings representing the kana. By simply inserting such a transfer card into the card processing unit 40, a transfer operation can be performed by designating a recorded transfer destination, thereby simplifying the transfer operation. The card reading means 24 stores in the output character string information storage means 22 the kana character strings representing the names of the recipients and clients read in this way, and information relating to numbers such as account numbers and telephone numbers is notified by numerical information. The means 29 is notified, and other information is notified to the search means 25.

  The search means 25 uses the name of the financial institution or branch name to be read out corresponding to the financial institution or branch code stored in the transfer card read by the card reading means 24 and the name expressed in kanji. Search from the information storage means 21.

  The voice synthesizing means 26 is a voice guidance information storing means 23 in which the name of the financial institution or branch to be read out and the name expressed in kanji, the numbers, and the names of the recipient and the client are merged at a predetermined location. Synthesizes speech information from the utterance content held by. Further, the speech synthesis unit 26 further includes a dictionary information storage unit 26a, a waveform information storage unit 26b, a conversion unit 26c, and a generation unit 26d in order to realize such a function.

  The dictionary information storage means 26a stores a dictionary to be referred to when synthesizing the text information of the vocabulary and numbers to be subjected to speech synthesis processing, and the notation, kana reading, accent, and the like for each vocabulary and numbers in advance. Information and detailed information of part-of-speech information are registered in association with each other.

  As described above, the waveform information storage unit 26b stores the information of the speech waveform corresponding to the kana characters constituting the phonetic character string and the kana character string. Note that the phonetic character string is composed of, for example, kana characters and special characters (colon, comma, quotes, percent symbols, etc.) that are phonetic symbols that indicate the reading and accent of the kana characters.

  Based on the dictionary information stored in the dictionary information storage unit 26a, the conversion unit 26c is a phonetic character string representing a phonetic representation of a name represented in Chinese characters of a financial institution or branch and a number such as a telephone number or an account number. Respectively. The conversion unit 26 c notifies the comparison determination unit 27 of the phonetic character string corresponding to the name, and causes the output character string information storage unit 22 to store the numerical phonetic character string. A specific example of the phonetic character string will be described later.

  The generating means 26d has a phonetic character string or a kana character string corresponding to a name expressed in Chinese characters of a financial institution or a branch, numbers such as an account number and a telephone number, and names of a recipient and a client. Utterances merged with locations where the names and numbers of institutions or branches are entered (for example, (a. Bank name), (b. Branch name), (c. Account number), (e. Recipient name), etc. in FIG. 6) About the information, based on the waveform information stored in the waveform information storage means 26b, the voice of the utterance information including the name of the financial institution or branch, the number, the name of the recipient and the client, respectively, as necessary Generate information.

  Then, the comparison determination means 27 compares the phonetic character string converted by the conversion means 26c of the speech synthesis means 26 with the kana character string representing the kana of the target character string, and based on the comparison result, To read out the target character string, either a phonetic character string or a kana character string is selected. The comparison / determination means 27 causes the output character string information storage means 22 to store the phonetic character string or the kana character string corresponding to the selected target character string.

  The utterance information merging means 28 is a financial institution of utterance contents held by the voice guidance information storage means 23, which is a phonetic character string or kana character string corresponding to the name of the financial institution or branch stored in the output character string information storage means 22. Or merge with the location where the branch name is entered. Further, the phonetic character string representing the phonetic number of the phone number and the like stored in the output character string information storage unit 22 is merged with the place where the phone number of the utterance content held in the voice guidance information storage unit 23 is entered. Further, the kana character string indicating the name and the name of the recipient and the client stored in the output character string information storage means 22 is merged with the place where the requester and the recipient of the utterance content held in the voice guidance information storage means 23 are entered. To do. Further, the utterance information merging means 28 merges the information stored in the output character string information storage means 22 with the utterance information stored in the voice guidance information storage means 23 in accordance with the progress of the transaction processing executed by the ATM 10. Then, the generation unit 26d of the speech synthesis unit 26 is notified. For example, before the ATM 10 executes the transfer process, in order to confirm the transfer information and the transfer amount acquired from the transfer card, the utterance information obtained by combining the transfer information and the transfer amount is sent to the generation unit 26d of the speech synthesis unit 26. Notification is performed.

  The number information notification means 29 sends the numbers corresponding to the telephone number, account number, etc. read from the transfer card by the card reading means 24 and the transfer amount received by the input receiving means 31 to the conversion means 26c of the speech synthesis means 26. Notice.

  The input accepting unit 31 accepts an operation input for a button on the push phone of the receiver 80. Further, the input receiving means 31 notifies the numeric information notifying means 29 of the information on the amount of money transferred by the operation input to the button of the push phone of the receiver 80.

  The utterance information notification means 32 selects the utterance information corresponding to the operation input to the button of the push phone received by the input reception means 31 from the voice guidance information storage means 23, and generates the selected utterance information by the generation means 26 d of the voice synthesis means 26. Notify Further, the utterance information notification means 32 selects the utterance information from the voice guidance information storage means 23 in accordance with the progress of the transaction processing executed by the ATM 10, and sends the selected utterance information to the generation means 26 d of the voice synthesis means 26. Notice. For example, after accepting the transfer card at the ATM 10, the utterance information that prompts the user to enter the personal identification number is requested, and the utterance information that indicates the completion of the transaction when the transaction is terminated. Is selected from the voice guidance information storage means 23.

  When the voice information is notified from the generating means 26d of the voice synthesizing means 26, the voice output control means 33 notifies the receiver 80 of a voice output request based on the voice information, and the voice information is output from the speaker of the receiver 80. The sound based on is output.

A voice reading process of the names of financial institutions and branches performed by the ATM 10 having such a configuration will be described.
FIG. 7 is a flowchart showing the ATM voice reading processing procedure according to the second embodiment. Note that FIG. 7 mainly shows a speech-to-speech process relating to the names of financial institutions and their branches.

  First, when a visually impaired user lifts the handset 80 of the ATM 10, the handset 80 is pressed so that the user presses the button of the push phone corresponding to the execution of each transaction of the handset 80 in order to cause the user to select a desired transaction. Voice guidance is output from the speaker.

  For example, when the user presses a button of a push phone corresponding to a transfer transaction of the receiver 80, the input reception unit 31 receives a signal corresponding to the press, and the utterance information notification unit 32 corresponds to the signal. The utterance information that prompts the insertion of the transfer card is selected from the voice guidance information storage means 23. Based on the waveform information storage unit 26b, the generation unit 26d of the speech synthesis unit 26 generates speech information from utterance information that prompts insertion of a transfer card. The voice output control means 33 notifies the receiver 80 of a voice output request based on the voice information, and voice guidance is made so that a transfer card is inserted from the speaker of the receiver 80 to the user.

  [Step S11] When the user inserts the transfer card into the card insertion / discharge port 40a of the card processing unit 40, the card reading means 24 records the card from the card insertion / discharge port 40a of the card processing unit 40 to the user's transfer card. Read the bank code and branch code of the bank account that is being transferred.

  [Step S12] The search unit 25 searches the financial institution information storage unit 21 for the name of the financial institution or branch to be read out and the name represented in kanji corresponding to the financial institution code and branch code read in step S11. To do.

  [Step S13] Based on the dictionary information stored in the dictionary information storage unit 26a, the conversion unit 26c of the speech synthesizer 26 converts the name of the financial institution or branch searched by the search unit 25 into the name of the name. Convert to phonetic string representing phonetic.

  [Step S20] The comparison / determination means 27 compares the phonetic character string converted in Step S13 with the kana character string representing the kana of the name represented by the kana of the financial institution or branch. In order to read out the name of the financial institution or branch based on the result of this comparison, either the phonetic character string or the kana character string is selected.

  [Step S14] The comparison determination unit 27 stores the phonetic character string or the kana character string corresponding to the name of the financial institution or branch selected in Step S20 in the output character string information storage unit 22.

[Step S15] The search unit 25 determines whether or not there is a name of another financial institution or branch that needs to be converted by the conversion unit 26c of the speech synthesis unit 26.
If it is determined that there is a name to be converted, the process proceeds to step S12, and if not, the process proceeds to step S16.

  [Step S16] The utterance information merging means 28 utters the utterance character string or kana character string stored in the output character string information storage means 22 in step S14 and corresponds to the processing content held by the voice guidance information storage means 23. It is merged with the contents and notified to the speech synthesis means 26.

  The generating unit 26d of the speech synthesizing unit 26 is based on the waveform information stored in the waveform information storing unit 26b, and the phonetic character string of the name represented in kanji of the notified financial institution or branch, or the financial institution or branch The voice information is generated from the utterance information in which the kana character string of the name represented by the above-mentioned name is merged at a predetermined utterance position.

  [Step S17] The voice output control means 33 notifies the receiver 80 of an output request for the voice information together with the voice information notified from the generating means 26d of the voice synthesizer 26, and the voice guidance based on the utterance information from the receiver 80. Is output.

Based on such a reading process, the user can perform an operation in response to voice guidance output from the speaker of the receiver 80.
Furthermore, the character string comparison process performed by the comparison determination unit 27 in step S20 will be described.

  In the character string comparison process, a matching condition is newly provided in addition to the case of matching when comparing the phonetic character string and the kana character string. First, special characters representing phonetic symbols in phonetic character strings are not used for comparison. Furthermore, when the selected character of the phonetic character string is a small letter and corresponds to the selected kana character of the kana character string, the selected character of the phonetic character string is a long sound, and the kana character string is In the case of “o” or “c”, they shall match. Under such conditions, the phonetic character string and the kana character string are compared.

FIG. 8 is a flowchart showing an ATM character string comparison processing procedure according to the second embodiment.
[Step S21] The comparison determination unit 27 includes a phonetic character string converted from the name represented in Chinese characters of the financial institution or branch notified from the conversion unit 26c of the speech synthesis unit 26, and the financial institution information storage unit 21. The first kana character (i = 1) with the kana character string of the name represented by the stored kana of the financial institution or branch to be stored is selected.

[Step S22] The comparison / determination means 27 determines whether or not the character selected in the phonetic character string is a special character.
If it is determined that the selected character is a special character, the process proceeds to step S27. If the selected character is not a special character, the process proceeds to step S23.

[Step S23] The comparison determination unit 27 compares the characters selected from the phonetic character string and the kana character string.
[Step S24] As a result of the comparison in step S23, if the selected characters match, the process proceeds to step S27, and if not, the process proceeds to step S25.

  [Step S25] As a result of the comparison in Step S23, if the selected character of the phonetic character string is a small character and corresponds to the selected Kana character of the Kana character string, the process proceeds to Step S27 as a match. If not, the process proceeds to step S26.

  [Step S26] As a result of the comparison in Step S23, if the selected character of the phonetic character string is a long sound and the kana character string is “o” or “c”, the process proceeds to Step S27 as a match. If the kana character string is not “o” or “c”, the process proceeds to step S29b.

[Step S27] The comparison / determination means 27 determines whether or not the characters of the selected phonetic character string and kana character string are the last character.
As a result of the determination, if it is the last character, the process proceeds to step S29a. If there is still a character to be compared instead of the last character, the process proceeds to step S28.

[Step S28] The comparison / determination means 27 newly selects a character (i = i + 1) adjacent in the end direction of the selected phonetic character string and kana character string.
[Step S29a] The comparison determination unit 27 selects a phonetic character string.

[Step S29b] The comparison determination unit 27 selects a kana character string.
Next, specific examples of the speech reading process and the character string comparison process performed in the ATM 10 will be described. In the following, a case will be described as an example in which the transfer content and the transfer amount are input to the ATM 10 and the transfer content is confirmed before the transfer transaction is executed.

  First, a user who is visually impaired inserts a transfer card into the card insertion / discharge port 40 a of the card processing unit 40. The transfer card includes, for example, the account number of the transfer destination, the financial code “0111” of “Kyushu Bank” which is a financial institution corresponding to the account number, and the branch code “National Branch” of the branch “ 0024 ”, the name of the recipient and the client (a pseudonym), and the telephone number of the client are magnetically recorded.

  Then, the card reading means 24 reads the transfer destination financial institution code “0111” and the branch name code “0024” recorded on the transfer card of the user from the card insertion / discharge port 40 a of the card processing unit 40. The names of the client and recipient are also read and stored in the output character string information storage unit 22, and the account number of the transfer destination and the telephone number of the client are notified to the number information notification unit 29 (step S11).

  The search means 25 corresponds to the financial institution code “0111” and the branch code “0024”, and the names expressed in kanji and “Kyushu Yuginko”, “Kyushu Bank”, “Kunitachi Shiten”. And “national branch”) are searched from the financial institution information storage means 21 of FIG. 5 (step S12).

  Next, the converting means 26c of the speech synthesizing means 26 first converts “Kyushu Bank”, which is the name of the financial institution, into a phonetic character string based on the dictionary information stored in the dictionary information storing means 26a. Similarly, the conversion means 26c of the speech synthesizer 26 also converts the account number of the transfer destination and the telephone number of the client into a phonetic character string and stores it in the output character string information storage means 22 (step S13). .

  The comparison determination means 27 compares the kana character string (“Kyushuyuginko”) stored in the financial institution information storage means 21 with the converted phonetic character string for “Kyushu Bank” (step S20). .

Here, the comparison process of “Kyushu Bank” will be specifically described.
FIG. 9 is a diagram for schematically explaining comparison of character strings according to the second embodiment. 9A represents a phonetic character string corresponding to the target character string to be read out, and FIG. 9B represents a kana character string corresponding to the target character string.

  It is a phonetic character string corresponding to “Kyushu Bank” in the name (target character string) represented in Kanji of the financial institution shown in FIG. The apostrophe and comma of the eighth and twelfth characters from the left in FIG. 9A are special characters of phonetic symbols representing “Ginkgo” and phonetic symbols representing accents.

First, the comparison determination means 27 selects the first kana character “ki” of the phonetic character string and the kana character string of “Kyushu Bank” (step S21).
It is determined that both kana characters are “ki”, are not special characters, and match, and “k” of the kana character is not the last character (steps S22, S23, S24, S27).

Next, the comparison / determination means 27 selects the kana characters “yu” and “yu” next to “ki” in the phonetic character string and the kana character string, respectively (step S28).
The selected phonetic character string “u” is not a special character, the phonetic character string “u” corresponds to the small letter “yu” in the kana character string, and the character phonetic character string “u” It is determined that it is not a character (steps S22, S23, S24, S25, S27).

Next, the comparison / determination means 27 selects the long sound “-” and “c” next to the phonetic character string “yu” and the kana character string “yu” (step S28).
In this case as well, “-” in the selected phonetic character string is a long sound and not a special character, and the phonetic character string “-” does not match the “k” in the kana character string. It is determined that “-” in the phonetic character string is not the last character (steps S22, S23, S24, S25, S26, S27).

  The fourth to seventh characters of the phonetic character string are similarly compared and matched. Next, when the comparison judging unit 27 selects “'” of the phonetic character string, it is a special character. The character is skipped and the next “n” is selected (steps S22, S27, S28).

  When the comparison determination means 27 selects “,” at the end of the phonetic character string, since it is a special character, it skips the character, and since the special character is the end character, A phonetic character string is selected (steps S22, S27, S29a).

The comparison determination unit 27 stores the phonetic character string of the financial institution name “Kyushu Bank” selected in this way in the output character string information storage unit 22 (step S14).
The search means 25 searches for a “national branch” to be read out other than “Kyushu Bank”.

  Next, as shown in FIG. 10, the comparison / determination unit 27 sets the kana character string (“Kunitachi Stain”) stored in the financial institution information storage unit 21 and the converted phonetic character string for “national branch”. Are compared (step S20).

The comparison process of “national branch” will be specifically described.
FIG. 10 is a diagram for schematically explaining comparison of character strings according to the second embodiment. 10A shows a phonetic character string corresponding to the target character string to be read out, and FIG. 10B shows a kana character string corresponding to the target character string.

First, the comparison determination unit 27 selects the first kana characters “K” and “K” of the phonetic character string and the kana character string of “National branch” (step S21).
The comparison judging means 27 is that both of the kana characters are “ko” and “ku”, which are not special characters, but they do not match, and the phonetic character string “ko” is “k” in the kana character string. It is determined that the phonetic character string “ko” is not a long sound, and the kana character string of the branch name “National branch” is selected (steps S22 to S26, S29b).

The comparison determination means 27 stores the kana character string of the branch name “national branch” selected in this way in the output character string information storage means 22 (step S14).
Although not shown after the transfer information is acquired from the transfer card in this way, the utterance information notification unit 32 selects the utterance information that prompts the user to input the transfer amount from the voice guidance information storage unit 23, and the generation unit 26d Based on the voice information generated from the speech information, the user is instructed to input the transfer amount from the speaker of the receiver 80.

  When the user inputs the transfer amount with the button of the push phone of the receiver 80, the input receiving unit 31 receives the transfer amount and notifies the number information notification unit 29 of the transfer amount. The conversion unit 26 c converts the transfer amount notified from the numeric information notification unit 29 into a phonogram string by the method described above and stores it in the output character string information storage unit 22.

  When the input of such transfer information and transfer amount to the ATM 10 is completed, the utterance information merging means 28 reads the phonetic character string of “Kyushu Bank” and the “national branch” stored in the output character string information storage means 22. The kana character string and the phonetic character string converted from the account number of the transfer destination, the telephone number of the client, and the transfer amount are merged into the utterance content corresponding to the processing content held by the voice guidance information storage means 23. Then, the voice synthesis means 26 is notified.

  Based on the waveform information stored in the waveform information storage unit 26b, the generation unit 26d of the speech synthesis unit 26 is based on the name of the financial institution and branch, the account number of the client, the name of the recipient and the client, and the client's telephone. Voice information is generated from the utterance information in which the number and the transfer amount are merged (step S16).

  The voice output control means 33 notifies the receiver 80 of the voice information output request together with the voice information, the name of the financial institution and the branch, the account number of the client, the name of the receiver and the client, and the client's name. Voice guidance based on the utterance information shown in FIG. 6 in which the telephone number and the transfer amount are merged is output from the speaker of the receiver 80 (step S17).

  As described above, in ATM 10, the comparison / determination means 27 represents the phonetic character string representing the phonetic of the target character string from the target character string including the kanji, and the kana of the target character string. In order to compare the kana character string composed of at least a clear sound of the semi-voiced sound based on the matching condition and to read out the target character string based on the comparison result, either the phonetic character string or the kana character string is It was made to choose. In this way, when the phonetic character string and the kana character string of the target character string match, the phonetic character string corresponds to the phonetic kana preset in the target character string, If the phonetic character string and the kana character string do not match, the phonetic character string does not correspond to the phonetic name preset in the target character string. For this reason, when the phonetic character string and the kana character string match, the phonetic character string is selected, and when they do not match, the kana character string is selected and the target character string is read correctly. As a result, the target character string to be read aloud is prevented from being read out incorrectly with an unnatural pronunciation.

  The above processing functions can be realized by a computer. In that case, a program describing processing contents of functions that the automatic transaction apparatus 1 and the ATM 10 should have is provided. By executing the program on a computer, the above processing functions are realized on the computer. The program describing the processing contents can be recorded on a computer-readable recording medium. Examples of the computer-readable recording medium include a magnetic storage device, an optical disk, a magneto-optical recording medium, and a semiconductor memory. Magnetic storage devices include hard disk devices (HDD), flexible disks (FD), magnetic tapes, and the like. Optical discs include DVD (Digital Versatile Disc), CD (Compact Disc) -ROM (Read Only Memory) / RW (Re-Writable), and the like. Magneto-optical recording media include MO (Magneto-Optical disk).

  When distributing the program, for example, a portable recording medium such as a DVD or a CD-ROM in which the program is recorded is sold. It is also possible to store the program in a storage device of a server computer and transfer the program from the server computer to another computer via a network.

  The computer that executes the program stores, for example, the program recorded on the portable recording medium or the program transferred from the server computer in its own storage device. Then, the computer reads the program from its own storage device and executes processing according to the program. The computer can also read the program directly from the portable recording medium and execute processing according to the program. In addition, each time a program is transferred from a server computer connected via a network, the computer can sequentially execute processing according to the received program.

  Further, at least one means of the above processing function can be realized by an electronic circuit such as a DSP (Digital Signal Processor), an ASIC (Application Specific Integrated Circuit), or a PLD (Programmable Logic Device).

DESCRIPTION OF SYMBOLS 1 Automatic transaction apparatus 2,26a Dictionary information storage means 3 Character string information storage means 4,26b Waveform information storage means 5,26c Conversion means 6,27 Comparison determination means 7,26d Generation means 8 Voice output means 10 ATM
20 control unit 20a CPU
20b RAM
20c HDD
20d graphic interface 20e host communication control means 20f input / output interface 20g bus 21 financial institution information storage means 22 output character string information storage means 23 voice guidance information storage means 24 card reading means 25 search means 26 speech synthesis means 28 utterance information merging means 29 Number information notification means 30 Display unit 30a Display means 30b Input detection means 31 Input reception means 32 Speech information notification means 33 Voice output control means 40 Card processing unit 40a Card insertion / discharge port 50 Passbook processing unit 50a Passbook insertion / discharge port 60 Banknote Processing unit 60a Bill insertion / discharge port 70 Coin processing unit 70a Coin insertion / discharge port 80 Handset

Claims (9)

In an automatic transaction apparatus that provides voice guidance,
Conversion means for converting a target character string including kanji into a phonetic character string representing a phonetic of the target character string;
Based on the comparison result between the converted phonetic character string and the kana character string that represents the kana character of the target character string and is composed of at least a clear sound, a muddy sound for a clear sound, and a semi-voice sound for a clear sound A comparison / determination means for selecting one of the character strings to read out the column;
An automatic transaction apparatus comprising: The phonetic character string is composed of kana characters and special characters representing phonetic symbols representing the pronunciation of the kana characters,
The automatic transaction apparatus according to claim 1. The comparison determination unit compares the kana characters constituting the phonetic character string and the kana character string one character at a time from the first character to the last character based on a matching condition.
The automatic transaction apparatus according to claim 2. The matching condition matches when the kana character constituting the phonetic character string is a stuttering or prompting character corresponding to the clear sound of the kana character string for the character to be compared. ,
The automatic transaction apparatus according to claim 3. The matching condition is that if the character representing the long sound of the kana character constituting the phonetic character string corresponds to “U” or “O” of the kana character string for the character to be compared, To
The automatic transaction apparatus according to claim 3. When the phonetic character string and the kana character string match, the comparison determination unit selects the phonetic character string to read out the target character string.
The automatic transaction apparatus according to claim 1. When the phonetic character string and the kana character string do not match, the comparison determination unit selects the kana character string to read out the target character string.
The automatic transaction apparatus according to claim 1. In the voice reading method for voice guidance,
The conversion means converts the target character string including kanji into a phonetic character string representing the phonetic character of the target character string,
The comparison determination means represents a comparison result between the converted phonetic character string and a kana character string that represents the kana of the target character string and is composed of at least a clear sound, a muddy sound for a clear sound, and a semi-voiced sound for a clear sound. Select any string to read the target string based on
A speech-to-speech method characterized by that. Computer
Conversion means for converting a target character string including kanji into a phonetic character string representing a phonetic of the target character string;
Based on the comparison result between the converted phonetic character string and the kana character string that represents the kana character of the target character string and is composed of at least a clear sound, a muddy sound for a clear sound, and a semi-voice sound for a clear sound A comparison / determination means for selecting one of the character strings to read out the column;
A speech-to-speech program characterized by functioning as

Images