JP4760452B2 - Speech training apparatus, speech training system, speech training support method and program - Google Patents

Description

The present invention, utterance training devices, speech training system, a utterance training support method, and a program, in particular, utterance training apparatus for supporting an enhanced speech technology, speech training system, relates utterance training support method, and a program.

  Contact centers for responding to inquiries from customers by telephone are provided at various locations. In this contact center, the operator responds to the customer by telephone.

  The operator needs to utter a phrase to be uttered. For example, the operator generally informs the customer of his / her affiliation and name at the beginning of a call with the customer, and informs his / her affiliation and name at the end of the call. In this case, the operator needs to reliably speak his / her affiliation and name.

  Moreover, the operator needs to suppress the utterance of a specific phrase. For example, there is an operator who frequently utters connected words such as “um”, “noh”, or “um”. If the operator speaks the connecting words frequently, it is difficult for the customer to hear the operator's speech.

  It is effective for the operator to confirm his / her utterance or listen to his / her own utterance evaluation in order to improve his / her utterance technique such as uttering a word / phrase to be uttered.

  Conventionally, an operator administrator (hereinafter referred to as a supervisor), for example, listens to a recording of a call between an operator and a customer, evaluates whether the utterance of the operator is appropriate, and communicates the evaluation to the operator. It was.

  However, the evaluation method in which a supervisor listens to a call recording and evaluates an utterance takes time. For this reason, the supervisor evaluated the utterance of the operator about once every several months. Even if an operator listens to his / her speech evaluation about once every several months, it is difficult to improve his / her speech technology.

  Japanese Patent Laid-Open No. 2004-309631 describes a dialogue practice support device that can automatically evaluate a user's utterance. This dialogue practice support device stores a pseudo question to the user and obtains a user's answer to the pseudo question. The dialogue practice support device extracts a specific term from the answer and presents the term.

As a result, it is possible to automatically confirm whether or not the user utters a specific term, and it is possible to shorten the time for evaluating the utterance of the operator.
Japanese Patent Application Laid-Open No. 2004-309631

  In the dialogue practice support device described in Patent Document 1, an operator must listen to a pseudo question and respond to the question. For this reason, the operator must improve the speech technology using the dialogue practice support device separately from the normal work such as reception with the customer. For this reason, an operator's burden increases.

  An object of the present invention is to provide a voice branching device, a speech training device, a voice branching method, a speech training support method, and a program capable of reducing a user's burden for improving speech technology.

The utterance training system of the present invention is an utterance training system including an utterance training apparatus, and an utterance training apparatus connected to the utterance training apparatus and a destination telephone, wherein the utterance training apparatus includes An acquisition unit that acquires an utterance to a telephone; and a branch unit that outputs the utterance acquired by the acquisition unit to both the telephone and the utterance training apparatus, wherein the utterance training apparatus specifies a user An object storage unit that stores an identifier and an extraction target phrase to be extracted from the user's utterance specified by the identifier in association with each identifier, an input unit that receives the identifier, and an output from the voice branching device A recognition unit that converts the received utterance into character information, and an extraction unit that extracts an extraction target phrase stored in the target storage unit from the character information converted by the recognition unit; See containing and a presentation unit that presents the extracted target phrase extracted by the extracting unit, the extracting unit, from the character information which the recognition unit is converted, associated with the identifier of the input unit accepts target storage unit The extracted word / phrase to be extracted is extracted.

In the speech training support method performed by the speech training system of the present invention, an identifier for identifying a user and an extraction target phrase to be extracted from the user's speech identified by the identifier are stored in association with each identifier. An utterance training support method performed by an utterance training system including an utterance training apparatus including a target storage unit, and an audio branching apparatus connected to the utterance training apparatus and a telephone of a call destination, wherein the utterance training apparatus includes: Both the input step of accepting the identifier, the speech acquisition device in which the speech branching device acquires speech to the telephone, and the speech branching device sends the acquired speech to both the telephone and the speech training device. Output step, the utterance training device accepts the outputted utterance, and the utterance training device accepts the utterance training device. A recognition step of converting the speech into text information, the speech training apparatus, from the converted character information, the extraction step of extracting an extraction target phrase associated with said accepted identifier in the object storage unit The speech training device includes a phrase presentation step of presenting the extracted extraction target phrase.

  According to the above invention, the utterance to the call destination is output to both the call destination telephone and the utterance training apparatus. When the utterance training apparatus accepts an utterance, the utterance training apparatus extracts an extraction target phrase from the utterance and presents the extracted extraction target phrase.

  Therefore, the user can confirm whether or not the extraction target phrase has been uttered by looking at the utterance training device while responding to the customer. Therefore, the user can improve the utterance technology while performing normal work, without having to confirm his / her utterance, for example, by listening to a pseudo question and responding to the question separately from the normal work. Is possible. Therefore, it is possible to reduce the burden on the user for improving the speech technology.

Also, speech training apparatus of the present invention is an utterance training device connected to the voice distribution unit, the voice dividing device includes an acquisition unit for acquiring speech to the call destination phone, the acquisition unit acquires A branching unit that outputs the uttered speech to both the telephone and the speech training device, wherein the speech training device extracts the identifier that identifies the user and the speech of the user identified by the identifier. A target storage unit that stores a word to be extracted in association with each identifier, an input unit that receives the identifier, and an utterance output from the speech branching device, and converts the received utterance into character information seen including a recognition unit, an extraction unit wherein the recognition unit extracts the extraction target phrase stored in the target storage unit from the converted text information, and a presentation unit that presents the extracted target phrase extracted by the extracting unit, the ,in front Extraction unit, from the character information which the recognition unit is converted, and extracts the extraction target phrase the input unit is associated with the identifier and the object storage unit accepted.

Further, the speech training support method of the present invention includes a target storage unit that stores an identifier for identifying a user and an extraction target phrase to be extracted from the user's speech specified by the identifier in association with each identifier. wherein, speech a speech branching unit and connected speech training support method speech training device performs the voice branch unit, an acquisition unit for acquiring speech to call to phone, acquired by the acquiring unit A branching unit that outputs to both the telephone and the utterance training device, an input step for receiving the identifier, a reception step for receiving the utterance output by the voice branching device, and the received utterance a conversion step of converting the text information from the converted text information, disconnect of extracting the extraction target phrase associated with said accepted identifier in the object storage unit Comprising a step, and a presentation step of presenting the extracted target phrase the extracted.

  According to the above invention, the extraction target phrase is extracted from the utterance to the telephone output from the voice branching device, and the extraction target phrase is presented.

  For this reason, even if a pseudo question is not memorized, it becomes possible for the user to confirm whether or not the user has spoken the extraction target phrase.

  In the method of extracting a specific term from the user's answer to the pseudo question, the specific term is registered in association with the pseudo question. For this reason, there is a possibility that the specific term cannot be set for each user.

  According to the above invention, when the identifier is accepted, the extraction target phrase associated with the identifier is extracted from the character information.

  For this reason, it becomes possible to present the extraction object phrase suitable for each user according to a user's mochi etc., for example.

The program of the present invention, an identifier for identifying the user, a target storage unit and extracting target phrase to be extracted from the utterance of the user, and stores in association with each of the identifiers specified by said identifier, voice A branching device, a program connected to a computer for executing speech training support processing, wherein the voice branching device acquires an utterance to a destination telephone, and the utterance acquired by the acquisition unit A branch unit that outputs to both the telephone and the computer, the process for receiving the identifier, the reception process for receiving the utterance output by the voice branching device, and converting the received utterance into character information a conversion processing for, from the converted character information, the process of extracting the extraction target phrase associated with said accepted identifier in the object storage unit, A presentation processing of presenting the extracted target phrase serial extracted, the speech training aid processing including causing the computer to perform.

  According to said invention, it becomes possible to make a computer perform the said speech training assistance method.

  ADVANTAGE OF THE INVENTION According to this invention, it becomes possible to reduce the burden of the user concerning the improvement of the technique of speech.

  Embodiments of the present invention will be described below with reference to the drawings.

  FIG. 1 is a block diagram showing an utterance training system according to an embodiment of the present invention. In FIG. 1, the speech training system includes a voice input device 1 and a speech training device 10. The voice input device 1 is connected to the telephone network 100 and the speech training device 10. Furthermore, the telephone network 100 is connected to the telephone 101 of the call destination, and the voice input device 1 is connected to the telephone 101 via the telephone network 100.

  The voice input device 1 acquires an utterance to the telephone set 101. The voice input unit 1 includes a voice acquisition unit 11 and a voice branching unit 12.

  The voice acquisition unit 11 is a microphone, for example, and acquires an utterance to the telephone set 101. Specifically, the voice acquisition unit 11 acquires the voice uttered by the user on the telephone 101. Note that the sound acquisition unit 11 may acquire the sound as a digital signal or an analog signal.

  When the voice acquisition unit 11 acquires voice, the voice branching unit 12 outputs the voice to both the telephone network 100 and the speech training device 10.

  When the utterance training apparatus 10 accepts an utterance, the utterance training apparatus 10 extracts an extraction target phrase from the utterance and outputs the extracted extraction target phrase. Moreover, the speech training apparatus 10 presents the extracted extraction target phrases.

  The utterance training apparatus 10 includes a storage unit 2, an instruction input unit 3, a call detection unit 4, an output unit 5, and a data processing unit 6.

  The storage unit 2 includes an extracted term storage unit 21, an extraction result storage unit 22, and a memory 23.

  The extracted term storage unit 21 stores target information. In the target information, an extraction target phrase to be extracted from the voice uttered by the user and the type of the extraction target phrase are associated for each extraction target phrase. The type of the extraction target phrase is, for example, the phrase that the extraction target phrase is to be uttered (for example, the name of the user), or the phrase that the extraction target phrase is to suppress the utterance (for example, “Ano”) One of the two).

  FIG. 2 is a diagram illustrating an example of the target information. In FIG. 2, the target information 200 includes a phrase 201 and a type 202. The phrase 201 indicates an extraction target phrase to be extracted from the utterance. The type 202 is a type of the word / phrase 201 and indicates “required” or “suppressed”. “Mandatory” indicates that the phrase 202 is a phrase to be uttered, while “suppression” indicates that the phrase 202 is a phrase whose utterance is to be suppressed.

  The detection result storage unit 22 stores extraction information. The extracted information includes at least the extracted phrase extracted from the speech by the term extracting unit 62. In the extraction information, the extracted word and the number of extractions of the extracted word may be associated for each extracted word. In the extracted information, the extracted word and the type of the extracted word may be associated with each extracted word. In the extracted information, the extracted word and the number of times the user uttered the extracted word per unit time (hereinafter referred to as utterance frequency) may be associated with each extracted word.

  Furthermore, the extracted information includes an identifier for identifying the user (for example, the user's name), the date of reception when the user made a call with the customer, the start time when the user started a call with the customer, and the user An end time when the call is ended may be included.

  FIG. 3 is a diagram showing an example of the extracted information. Specifically, an example of extracted information in a call with a customer made by a user whose name is “Takahashi” from 14:55:05 to 15:02:38 on October 10, 2005 was shown. FIG.

  In FIG. 3, the extracted information 300 includes a responder 301, a response date 302, a start time 303, an end time 304, a word / phrase 305, a type 306, and a frequency 307.

  The corresponding person 301 is an identifier and indicates the name of the user. The response date 302 indicates the date when the user made a call with the customer. The start time 303 includes a start time when the user starts a call with the customer. The end time 304 indicates the end time when the user ended the call with the customer. A phrase 305 indicates an extracted phrase extracted from the utterance. A type 306 indicates the type of the word / phrase 305. The number of times 307 indicates the number of times the extracted word / phrase indicated by the word / phrase 305 has been extracted.

  The memory 23 is a recording medium readable by the CPU, and records a program that defines the operation of the data processing unit 6. Moreover, the memory 23 records an identifier that identifies a user who uses the speech training apparatus.

  The instruction input unit 3 is, for example, a keyboard or a mouse, and accepts a search instruction for searching for extracted information and an update instruction for changing target information.

  The call detection unit 4 detects the start and end of a call between a user and a customer.

  The output unit 5 presents various information to the user. The output unit 5 is a display, for example.

  The data processing unit 6 is, for example, a CPU, reads a program recorded in the memory 23, executes the read program, a voice recognition unit 61, a term extraction unit 62, a clock unit 63, and an extraction result processing unit. 64, an instruction management unit 65, an extraction result search unit 66, and a setting unit 67.

  The voice recognition unit 61 receives the voice output from the voice branching unit 12 and converts the voice into character information.

  The term extraction unit 62 confirms the target information stored in the extracted term storage unit 21 when the speech recognition unit 61 converts speech into character information. Further, the term extraction unit 62 checks whether or not the extraction target phrase included in the target information is included in the character information.

  When the extraction target word / phrase is included in the character information, the term extraction unit 62 extracts the extraction target word / phrase from the character information as an extraction word / phrase. On the other hand, if the extraction target word / phrase is not included in the character information, the term extraction unit 62 waits until the extracted speech recognition unit 61 next converts the speech into character information.

  The term extraction unit 62 outputs the extracted extracted phrase to the extraction result processing unit 5. The term extraction unit 62 associates the extracted phrase, the extraction target phrase extracted as the extracted phrase and the type associated with the extracted term storage unit 21, and outputs the extracted phrase to the extraction result processing unit 5. You may do it.

  The clock unit 63 measures the current date and time.

  When the call detection unit 4 detects the start of a call, the extraction result processing unit 64 acquires the current date from the clock unit 63 as the response date and the current time as the start time. Further, when the call detection unit 4 detects the end of the call, the extraction result processing unit 64 acquires the current time from the clock unit 63 as the end time.

  When the term extraction unit 62 extracts an extracted word / phrase, in other words, the extraction result processing unit 64 receives the extracted word / phrase output by the term extraction unit 62, the extraction result processing unit 64 causes the output unit 5 to output the extracted word / phrase.

  For example, when the output unit 5 is a display device, the extraction result processing unit 64 causes the output unit 5 to display all the extracted words / phrases in the order of extraction.

  Further, the extraction result processing unit 64 may display the same extracted word / phrase on the output unit 5 only once. In this case, when the term extraction unit 62 extracts the extracted words / phrases, the extraction result processing unit 64 checks whether the extracted words / phrases are already displayed on the output unit 5. If the extracted word / phrase is not displayed, the extraction result processing unit 64 causes the output unit 5 to display the extracted word / phrase. On the other hand, when the extracted word / phrase is displayed, the extraction result processing unit 64 stands by until the next extracted word / phrase is extracted.

  Further, the detection result processing unit 64 may display the extracted word and the utterance frequency of the extracted word on the output unit 5 in association with each extracted word.

  In this case, the detection result processing unit 64 calculates the number of extractions of the extracted phrase extracted by the term extracting unit 62 for each extracted phrase. After that, when obtaining the end time from the clock unit 63, the detection result processing unit 64 calculates a call time during which the user talks with the customer based on the end time and the start time. The detection result processing unit 64 calculates the number of utterances per predetermined time (for example, 1 minute or 10 minutes) as the utterance frequency based on the number of extractions and the call time. The detection result processing unit 64 causes the output unit 5 to display the extracted word / phrase, the utterance frequency of the extracted word / phrase, and the extracted word / phrase in association with each other.

  When the extraction result processing unit 64 causes the output unit 5 to output the extracted word / phrase, the extraction result processing unit 64 creates extraction information including the extracted word / phrase.

  For example, when creating extraction information including the utterance frequency, the extraction result processing unit 64 creates extraction information after obtaining end information.

  In this case, first, when the extraction result processing unit 64 acquires the end information, the extraction result processing unit 64 calculates the number of extraction times, the call time, and the utterance frequency. When the result processing unit 64 has already calculated the number of extractions, the call time, and the utterance frequency, such as when the utterance frequency is displayed on the output unit 5, it is necessary to calculate the number of extractions, the call time, and the utterance frequency again. Absent.

  Subsequently, the detection result processing unit 64 reads the identifier from the memory 23. The extraction result processing unit 64 creates extraction information including the identifier, response date, start time, end time, call time, extracted word / phrase, type of extracted word / phrase, number of utterances, and utterance frequency. In the extracted information, the extracted phrase, the type of the extracted phrase, the number of extractions of the extracted phrase, and the utterance frequency of the extracted phrase are associated with each extracted phrase.

  When the detection result processing unit 64 creates the extraction information, the detection result processing unit 64 stores the extraction information in the extraction result storage unit 22.

  When the instruction input unit 3 receives a search instruction, the instruction management unit 65 outputs the search instruction to the extraction result search unit 66. In addition, when the instruction input unit 3 receives an update instruction, the instruction management unit 65 outputs the update instruction to the setting unit 67.

  When the instruction input unit 3 receives a search instruction, in other words, when the instruction input unit 3 receives a search instruction from the instruction management unit 65, the extraction result search unit 66 searches the extraction result storage unit 22 for extraction information corresponding to the search instruction. .

  For example, when the instruction input unit 3 receives an identifier as a search instruction, the extraction result search unit 66 searches the extraction result storage unit 22 for extraction information including the identifier. When the instruction input unit 3 receives a date as a search instruction, the extraction result search unit 66 searches the extraction result storage unit 22 for extraction information including the date.

  When the extraction result retrieval unit 66 retrieves the extracted information, the extraction result retrieval unit 66 causes the output unit 5 to output the retrieved extraction information. Note that the extraction result search unit 66 may cause the output unit 5 to output only a part of the extracted extraction information (for example, an extracted word and the number of utterances thereof).

  When the output unit 5 is a display device, the extraction result search unit 66 displays the extracted extraction information on the output unit 5. The extraction result search unit 66 may process the extracted extraction information into information that allows the user to easily understand the content of the extraction information, and display the processed extraction information on the output unit 5.

  For example, when the instruction input unit 3 accepts a date range as a search instruction, the extraction result search unit 66 searches for extracted information including the response date included in the date range, and the extracted information The extracted words and utterance frequencies (or the number of extractions) included in are processed into a table or graph format. The extraction result search unit 66 displays the processed utterance frequency on the output unit 5.

  FIG. 4 is a diagram showing an example of the utterance frequency of the extracted word “ano” processed into a table format. In FIG. 4, the vertical axis indicates the utterance frequency of “Ano”, and the horizontal axis indicates the date.

  When the instruction input unit 3 receives an update instruction, in other words, when the instruction input unit 3 receives an update instruction from the instruction management unit 65, the setting unit 67 changes the target information stored in the extracted term storage unit 21 according to the update instruction.

  For example, when the instruction input unit 3 receives a phrase to be added to the extraction target phrase and the type of the phrase as an update instruction, the setting unit 67 stores the phrase and the type in association with each other in the extracted term storage unit 21. When the instruction input unit 3 receives a word to be deleted from the extraction target word as an update instruction, the setting unit 67 deletes the word and the type associated with the word from the extracted word storage unit 21.

  Next, the operation of the speech training system will be described.

  FIG. 5 is a flowchart for explaining the operation from when the user's utterance is acquired until the extracted phrase is stored.

  When the voice acquisition unit 11 acquires the voice uttered by the user, the voice acquisition unit 11 executes Step S501.

  In step S <b> 501, the voice acquisition unit 11 outputs the voice to the voice branching unit 12. When the voice branching unit 12 accepts the voice, the voice branching unit 12 executes Step S502.

  In step S502, the voice branching unit 12 outputs the voice to both the telephone network 7 and the voice recognition unit 61. When receiving the voice, the voice recognition unit 61 executes Step S503.

  In step S503, the voice recognition unit 61 converts the voice into character information and outputs the character information to the term extraction unit 62. When the term extraction unit 62 receives character information from the speech recognition unit 61, the term extraction unit 62 executes step S504.

  In step S <b> 504, the term extraction unit 62 confirms the target information stored in the extracted term storage unit 21. The term extraction unit 62 checks whether or not the extraction target phrase included in the target information is included in the character information. When the extraction target word / phrase is included in the character information, the term extraction unit 62 executes step S505. When the extraction target word / phrase is not included in the character information, the term extraction unit 62 ends the operation.

  In step S505, the term extraction unit 62 extracts the extraction target phrase as an extraction phrase from the character information, and outputs the extracted phrase to the extraction result processing unit 64. The extraction result processing unit 64 causes the output unit 5 to present the extracted words / phrases. After completing step S505, the extraction result processing unit 64 executes step S506.

  In step S506, the extraction result processing unit 64 creates extraction information, and stores the extraction information in the extraction result storage unit 22.

  Next, an operation for searching for an extracted word / phrase stored in the extraction result storage unit will be described.

  When receiving the search instruction, the instruction input unit 3 outputs the search instruction to the instruction management unit 65. When receiving the search instruction, the instruction management unit 65 outputs the search instruction to the extraction result search unit 66.

  When receiving the search instruction, the extraction result search unit 66 searches the extraction result storage unit 22 for extraction information according to the search instruction. The extraction result search unit 66 causes the output unit 5 to output the extracted extraction information.

  Next, the operation of changing the extraction target phrase will be described.

  When receiving an update instruction, the instruction input unit 3 outputs the update instruction to the instruction management unit 65. When receiving an update instruction, the instruction management unit 65 outputs the update instruction to the setting unit 67.

  When receiving the update instruction, the setting unit 67 changes the extraction target phrase stored in the extracted term storage unit 21 in accordance with the update instruction.

  According to the present embodiment, the voice acquisition unit 11 acquires an utterance to the telephone set 101. The voice branching unit 12 outputs the utterance acquired by the voice acquisition unit 11 to both the telephone set 101 and the utterance training apparatus 100.

  In this case, the utterance to the call destination is output to both the telephone set 101 and the utterance training apparatus 10. When the utterance training apparatus 10 accepts an utterance, the utterance training apparatus 10 extracts an extraction target phrase from the utterance and presents the extracted extraction target phrase.

  Therefore, the user can confirm whether or not the extraction target phrase has been uttered by looking at the utterance training device 10 while responding to the customer. Therefore, the user can improve the utterance technology while performing normal work, without having to confirm his / her utterance, for example, by listening to a pseudo question and responding to the question separately from the normal work. Is possible. Therefore, it is possible to reduce the burden on the user for improving the speech technology.

  The voice recognition unit 61 converts the voice output from the voice branching unit 12 into character information. The term extraction unit 62 extracts the extraction target phrases stored in the extracted term storage unit 21 from the character information. The output unit 5 presents the extraction target phrase extracted by the term extraction unit 62.

  In this case, the utterance to the telephone output from the voice input device 1 is converted into character information. Further, the extraction target phrase is extracted from the character information, and the extraction target phrase is presented.

  For this reason, even if a pseudo question is not memorized, it becomes possible for the user to confirm whether or not the user has spoken the extraction target phrase.

  Next, an utterance training system capable of setting an extraction target phrase for each user will be described.

  In this case, in the extraction information stored in the extracted term storage unit 21, the identifier, the extraction target phrase to be extracted from the user's utterance specified by the identifier, and the type of the extraction target phrase are associated for each identifier. It has been.

  FIG. 6 is a flowchart for explaining an operation from acquiring a user's utterance to storing the extracted word / phrase in the utterance training system in which an extraction target word / phrase can be set for each user. In FIG. 6, the same processes as those in FIG. 5 are denoted by the same reference numerals.

  When the term extraction unit 62 receives character information from the speech recognition unit 61 in step S503, the term extraction unit 62 executes step S601.

  In step S <b> 601, the term extraction unit 62 confirms the identifier recorded in the memory 62, and confirms the extraction target phrase associated with the identifier in the extracted term storage unit 21. The term extraction unit 62 checks whether the extraction target phrase is included in the character information. When the extraction target word / phrase is included in the character information, the term extraction unit 62 executes step S505. When the extraction target word / phrase is not included in the character information, the term extraction unit 62 ends the operation.

  Thereafter, steps S505 and S506 are executed. In step S506, the extraction result processing unit 64 creates extraction information in which at least the identifier and the extraction target phrase are associated with each other.

  Next, the operation until the memory 23 records the identifier will be described.

  When the instruction input unit 3 receives an identifier for specifying a user who uses the speech training apparatus 10, the instruction input unit 3 outputs the received identifier to the instruction management unit 65.

 When receiving an identifier from the instruction input unit 3, the instruction management unit 65 outputs the identifier to the setting unit 67.

  When the setting unit 67 receives an identifier from the instruction management unit 65, the setting unit 67 records the identifier in the memory 23. When the memory 23 has already recorded the identification, the setting unit 67 overwrites the identifier received by the instruction input unit 3 with the identifier recorded in the memory 23.

  Next, the operation | movement which changes an extraction object phrase in the speech training system which can store an extraction object phrase for every user is demonstrated.

  When the instruction input unit 3 receives the identifier and a change instruction to change the extraction target phrase to be extracted from the user's utterance specified by the identifier, the instruction input unit 3 sends the identifier and the change instruction to the instruction management unit 65. Output.

  When receiving the identifier and the change instruction, the instruction management unit 65 outputs the identifier and the change instruction to the setting unit 67.

  When receiving the identifier and the change instruction, the setting unit 67 changes the extraction target phrase associated with the identifier and the extracted term storage unit 21 in accordance with the update instruction.

  In this embodiment, the instruction input unit 3 receives an identifier. The term extraction unit 62 extracts, from the character information, an extraction target phrase associated with the identifier received by the instruction input unit 3 and the extracted term storage unit 21.

  In this case, for example, it is possible to set an extraction target word / phrase suitable for each user, such as changing the extraction target word / phrase according to the user's motivation.

  In the embodiment described above, the illustrated configuration is merely an example, and the present invention is not limited to the configuration.

It is the block diagram which showed one Example of the speech training system. It is the figure which showed an example of object information. It is the figure which showed an example of extraction information. It is the figure which showed an example of the utterance frequency processed into the format of a table | surface. It is a flowchart for demonstrating the operation example of a speech training system. It is a flowchart for demonstrating the other operation example of an utterance training system.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 Voice input device 2 Memory | storage part 3 Instruction input part 4 Call detection part 5 Output part 6 Data processing part 10 Speech training apparatus 11 Voice acquisition part 12 Voice branch part 21 Extraction term storage part 22 Extraction result storage part 23 Memory 61 Voice recognition part 62 Term Extraction Unit 63 Clock Unit 64 Extraction Result Processing Unit 65 Instruction Management Unit 66 Extraction Result Search Unit 67 Setting Unit 100 Telephone Network 101 Telephone 200 Target Information 201 Phrase 202 Type 300 Extraction Information 301 Corresponding Person 302 Corresponding Date 303 Start Time 304 End time 305 Word / phrase 306 Type 307 Number of times

Claims (9)

A speech training device that is connected to the voice branching unit,
The voice branching device includes an acquisition unit that acquires an utterance to a telephone of a call destination, and a branching unit that outputs the utterance acquired by the acquisition unit to both the telephone and the utterance training device,
The utterance training device comprises:
A target storage unit that stores an identifier for identifying a user and an extraction target phrase to be extracted from the user's utterance specified by the identifier, in association with each identifier ;
An input unit for receiving the identifier;
A recognition unit that accepts an utterance output from the voice branching device, and converts the accepted utterance into character information;
An extraction unit that extracts an extraction target phrase stored in the target storage unit from the character information converted by the recognition unit;
See containing and a presentation unit that presents the extracted target phrase extracted by the extracting unit,
The utterance training apparatus , wherein the extraction unit extracts, from the character information converted by the recognition unit, an extraction target phrase associated with the identifier received by the input unit and the target storage unit . In the utterance training apparatus according to claim 1,
The input unit further receives a change instruction from the user to change the extraction target phrase,
The utterance training apparatus includes a setting unit that changes an extraction target phrase stored in the target storage unit according to the change instruction,
The said presentation part is an utterance training apparatus which presents the said extraction object phrase to the said user. An utterance training system comprising: an utterance training device; and a speech branching device connected to the utterance training device and a telephone of a communication destination,
The audio branching device is:
An acquisition unit for acquiring an utterance to the telephone;
A branching unit that outputs the utterance acquired by the acquisition unit to both the telephone and the utterance training device;
The utterance training device comprises:
A target storage unit that stores an identifier for identifying a user and an extraction target phrase to be extracted from the user's utterance specified by the identifier, in association with each identifier ;
An input unit for receiving the identifier;
A recognition unit that accepts an utterance output from the voice branching device, and converts the accepted utterance into character information;
An extraction unit that extracts an extraction target phrase stored in the target storage unit from the character information converted by the recognition unit;
See containing and a presentation unit that presents the extracted target phrase extracted by the extracting unit,
The utterance training system , wherein the extraction unit extracts an extraction target word / phrase associated with an identifier received by the input unit and the target storage unit from character information converted by the recognition unit . In the speech training system according to claim 3,
The input unit further receives a change instruction from the user to change the extraction target phrase,
The utterance training apparatus includes a setting unit that changes an extraction target phrase stored in the target storage unit according to the change instruction,
The said presentation part is an utterance training system which presents the said extraction object phrase to the said user. It includes an identifier for identifying the user, and extracts target phrase to be extracted from the utterance of the user identified by the identifier, the object storage unit for storing in association with each of the identifier, which is connected to the voice distribution unit utterance A speech training support method performed by a training device,
The voice branching device includes an acquisition unit that acquires an utterance to a telephone of a call destination, and a branching unit that outputs the utterance acquired by the acquisition unit to both the telephone and the utterance training device,
An input step for receiving the identifier;
An accepting step of accepting an utterance output from the voice branching device;
A conversion step of converting the received utterance into character information;
An extraction step for extracting the extracted identifier associated with the received identifier and the target storage unit from the converted character information;
A presentation step of presenting the extracted extraction target phrase, and an utterance training support method. The speech training support method according to claim 5,
Further receiving from the user a change instruction to change the extraction target phrase;
Changing the extraction target words stored in the target storage unit according to the change instruction,
In the presenting step, the speech training support method for presenting the extraction target phrase to the user. An utterance training apparatus including an object storage unit that stores an identifier for identifying a user and an extraction target phrase to be extracted from the user's utterance identified by the identifier in association with each identifier; and the utterance training apparatus; An utterance training support method performed by an utterance training system including a voice branching device connected to a telephone of a destination,
The utterance training apparatus, an input step for receiving the identifier;
The voice branching device acquires an utterance to the telephone, an utterance acquisition step;
An output step in which the voice branching device outputs the acquired utterance to both the telephone and the utterance training device;
The utterance training device, the utterance reception step of receiving the output utterance;
A recognition step in which the utterance training device converts the received utterance into character information;
The utterance training device extracts an extraction target word / phrase associated with the received identifier and the target storage unit from the converted character information ;
An utterance training support method, wherein the utterance training apparatus includes a phrase presentation step of presenting the extracted extraction target phrase. An identifier for identifying the user, which is connected an extraction target phrase to be extracted from the utterance of the user identified by the identifier, and a target storage unit for storing in association with each of the identifiers, and the voice branching unit, the A program for causing a computer to execute speech training support processing,
The voice branching device includes an acquisition unit that acquires an utterance to a destination telephone, and a branching unit that outputs the utterance acquired by the acquisition unit to both the telephone and the computer,
Processing to accept the identifier;
An accepting process for accepting an utterance output by the voice branching device;
A conversion process for converting the accepted utterance into character information;
A process of extracting the extracted identifier associated with the received identifier and the target storage unit from the converted character information ;
A program for causing the computer to execute an utterance training support process including a presentation process for presenting the extracted extraction target phrase. The program according to claim 8, wherein
The speech training support process is:
A process of further receiving a change instruction to change the extraction target phrase from the user;
Processing to change the extraction target phrase stored in the target storage unit according to the change instruction,
In the presenting process, a program for presenting the extraction target phrase to the user.

Images