JP2012155272A - Speech recognition dictionary extension device, system, method and program - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a speech recognition dictionary extension device which can utilize a term which is registered in a user dictionary created by each user as a term to be used for speech recognition.SOLUTION: A speech recognition dictionary storage means 81 stores a speech recognition dictionary including terms which are objects of the speech recognition. An unregistered term extraction means 82 extracts a term which is not present in the speech recognition dictionary from a term list that is a list created by associating terms which are speech recognition objects with the readings of the terms and adds the extracted term to the speech recognition dictionary. A speech recognition means 83 performs speech recognition based on the speech recognition dictionary.

Description

  The present invention relates to a speech recognition dictionary expansion device, a speech recognition dictionary expansion system, a speech recognition dictionary expansion method, and a speech recognition dictionary expansion program for expanding a speech recognition dictionary used when a device provided on a communication network performs speech recognition. .

  Patent Literature 1 describes a dictionary distribution method that makes information registered in a user's individual dictionary available to other users. In the method described in Patent Literature 1, when the user dictionary of each user is registered as a public dictionary, other users select a dictionary that the user wants to use from the user dictionaries registered in the public dictionary, Import the user dictionary as needed.

Japanese Patent No. 3648881

  Currently, business applications provided in cloud computing are widespread. In addition, information terminals such as smart phones are widely used. However, in order to construct a business application using such information terminals, there is a problem that it is troublesome to input information. Thus, efforts are being made to input information by utilizing the voice recognition function of the information terminal.

  When performing speech recognition, it is necessary to register many terms in order to improve recognition accuracy. However, it is difficult for the operation side alone to cover technical terms, and the recognition accuracy cannot be improved. Therefore, in order to have the user register such terms, a mechanism that promotes many uses by the user is necessary.

  According to the dictionary distribution method described in Patent Document 1, each user can handle a dictionary selected as a dictionary when inputting characters.

  However, the dictionary distribution method described in Patent Document 1 cannot be applied to a processing model in which speech recognition for a plurality of users is performed using a single dictionary (speech recognition dictionary). Further, the dictionary distribution method described in Patent Document 1 requires a mechanism for controlling a plurality of dictionaries.

  As described above, in order to increase the accuracy of speech recognition, it is necessary to register many terms in a single speech recognition dictionary. For that purpose, it is desirable to be able to utilize a user dictionary created by each user.

  Therefore, the present invention provides a speech recognition dictionary expansion device, a speech recognition dictionary expansion system, a speech recognition dictionary expansion method, and a speech recognition dictionary expansion program that can use terms registered in a user dictionary created by each user as terms used for speech recognition. The purpose is to provide.

  A speech recognition dictionary expansion device according to the present invention is created by associating speech recognition dictionary storage means for storing a speech recognition dictionary including a term to be speech-recognized, and a term to be speech-recognized and a reading of the term. A term that is not included in the speech recognition dictionary, and an unregistered term extraction unit that adds the extracted term to the speech recognition dictionary, and speech recognition that performs speech recognition based on the speech recognition dictionary Means.

  A speech recognition dictionary expansion system according to the present invention includes a speech recognition server that performs speech recognition, and a request terminal that transmits speech to the speech recognition server and requests speech recognition, and the speech recognition server is a target of speech recognition. A speech recognition dictionary storage means for storing a speech recognition dictionary including terms, and a term that does not exist in the speech recognition dictionary from a term list that is a list created by associating a term that is a target of speech recognition with a reading of the term An unregistered term extracting unit that extracts and adds the extracted term to the speech recognition dictionary, and a speech recognition unit that performs speech recognition of the speech received from the requesting terminal based on the speech recognition dictionary.

  A speech recognition dictionary expansion method according to the present invention converts a term that is a target of speech recognition into a speech recognition dictionary that includes a term that is a target of speech recognition, from a term list that is a list that is created by associating a term that is subject to speech recognition with reading of the term A term that does not exist is extracted, the extracted term is added to a speech recognition dictionary, and speech received from a requesting terminal that requests speech recognition is recognized based on the speech recognition dictionary.

  A speech recognition dictionary expansion program according to the present invention is a speech recognition dictionary expansion program applied to a computer provided with speech recognition dictionary storage means for storing a speech recognition dictionary including a term to be a target of speech recognition. A term that does not exist in the speech recognition dictionary is extracted from the term list, which is a list that is created by associating the terms that are subject to speech recognition and the readings of the terms, and the extracted terms are not added to the speech recognition dictionary. A registered term extraction process and a speech recognition process for performing speech recognition based on a speech recognition dictionary are executed.

  According to the present invention, terms registered in a user dictionary created by each user can be utilized as terms used for speech recognition.

It is a block diagram which shows one Embodiment of the speech recognition dictionary expansion system by this invention. It is explanatory drawing which shows the example of a term list. It is explanatory drawing which shows the example of the contribution word list | wrist containing a contribution degree point. It is a flowchart which shows the example of operation | movement of the speech recognition dictionary expansion system in this embodiment. It is a block diagram which shows the example of the minimum structure of the speech recognition dictionary expansion apparatus by this invention. It is a block diagram which shows the example of the minimum structure of the speech recognition dictionary expansion system by this invention.

  Hereinafter, embodiments of the present invention will be described with reference to the drawings.

  FIG. 1 is a block diagram showing an embodiment of a speech recognition dictionary expansion system according to the present invention. The speech recognition dictionary expansion system according to the present invention includes a service providing server 10, an information terminal 20, and a registration terminal 30. The service providing server 10, the information terminal 20, and the registration terminal 30 operate under program control and are connected to each other via a communication network 100 such as the Internet.

  The information terminal 20 is an information processing apparatus such as a small personal computer. The information terminal 20 includes a voice transmission / reception unit 21. Further, a microphone 22 and a display unit 23 are connected to the information terminal 20. The voice transmitting / receiving unit 21 requests voice recognition by transmitting the voice recorded by the microphone 22 to the service providing server 10 via the communication network 100. In addition, when the voice transmitting / receiving unit 21 receives the voice recognition result from the service providing server 10, the voice transmitting / receiving unit 21 displays the result on the display unit 23.

  The service providing server 10 is an information processing apparatus such as a workstation server installed on the Internet. The service providing server 10 includes an ID management unit 11, a business service providing unit 12, a speech recognition dictionary management unit 13, a speech recognition unit 14, and a contribution management unit 15.

  The ID management unit 11 stores user information for identifying a user of the service providing server 10. Further, the business service providing unit 12 performs various business services provided to the user. For example, when the result of speech recognition is transmitted to the user by electronic mail, the business service providing unit 12 edits the result of speech recognition into a format that can be transmitted by electronic mail. Also, for example, when the voice recognition result is shared within the group to which the user belongs, the business service providing unit 12 edits the voice recognition result into a format that can be used by the groupware used by the members of the group. Also good. However, the business service provided by the business service providing unit 12 is not limited to the above content.

  The voice recognition dictionary management unit 13 includes a dictionary control unit 16 and a dictionary storage unit 17.

  The dictionary storage unit 17 stores a speech recognition dictionary (hereinafter also referred to as an existing dictionary) used when the speech recognition unit 14 described later performs speech recognition. Specifically, the existing dictionary is a dictionary of terms used for deriving a speech recognition result from speech data in a speech recognition service. The speech recognition dictionary includes terms that are subject to speech recognition and reading of the terms. However, the information included in the speech recognition dictionary is not limited to terms and reading of terms.

  When the dictionary control unit 16 receives a term list from the registration terminal 30 described later, the dictionary control unit 16 extracts terms that are not registered in the existing dictionary (speech recognition dictionary) from the terms included in the received term list, and the extracted terms Add to dictionary (voice recognition dictionary). Here, the term list is a list in which a term to be subjected to speech recognition and a reading of the term are paired. That is, the term list corresponds to the user dictionary. FIG. 2 is an explanatory diagram illustrating an example of a term list. In the example shown in FIG. 2, for example, the term “securities” corresponds to the reading of the term “Shoken”. The dictionary control unit 16 may add all the terms included in the received term list to the speech recognition dictionary.

  The contribution management unit 15 includes a contribution control unit 18 and a contribution storage unit 19.

  The contribution storage unit 19 stores a list of unregistered terms in the existing dictionary extracted by the dictionary control unit 16. Since this list is a list of terms that contribute to speech recognition, in the following description, this list is referred to as a contribution word list. In other words, the contribution word list means that a user registered a term in order to make a determination that "the user who registered the term contributed to dictionary expansion by registering a term that does not exist in the existing dictionary in the existing dictionary". It can be said that it is a term list including the information. The contribution word list may include a creation source user of a term list from which unregistered terms are extracted.

  The contribution storage unit 19 stores predetermined points given to each user. Specifically, the contribution storage unit 19 stores points given when terms registered from the term list of each user among terms registered in the speech recognition dictionary are used for speech recognition. Since the points given in this way are points given when contributing to voice recognition, in the following description, these points are referred to as contribution points.

  FIG. 3 is an explanatory diagram illustrating an example of a contribution word list including contribution degree points. In the example illustrated in FIG. 3, for example, the term “investment trust” is registered from the term list of the user “Mr. A”, and the term is given 20 contribution points.

  The contribution degree control unit 18 generates a contribution word list from terms not registered in the existing dictionary extracted by the speech recognition dictionary management unit 13. For example, as illustrated in FIG. 3, the contribution degree control unit 18 may generate a contribution word list by associating a term with a user who created the term list.

  Further, the contribution control unit 18 determines whether or not a term included in the speech recognized by the speech recognition unit 14 described later is a term extracted from the term list, and is determined to be a term extracted from the term list. In that case, contribution points are given to the user. Specifically, when the term used for speech recognition is a word existing in the contribution word list, the contribution degree control unit 18 gives a contribution point to the user who registered the word. The contribution control unit 18 may give a fixed value as the contribution point, or may give a point changed based on other factors.

  The contribution control unit 18 may total the terms included in the expanded speech recognition dictionary for the terms included in the speech recognized by the speech recognition unit 14. Then, the contribution level control unit 18 may collate the total result with the contribution word list and total the number of terms used in the contribution word list from the total result. The administrator can give the user a discount on the usage fee for the business service and a privilege according to the number of usages and the contribution points thus collected.

  When the contribution control unit 18 receives a contribution point inquiry request from the registration terminal 30, the contribution control unit 18 extracts the contribution point of the designated user from the contribution storage unit 19 and transmits the extraction result to the registration terminal 30. .

  The voice recognition unit 14 performs voice recognition of the voice transmitted from the information terminal 20 based on the existing dictionary, and returns a recognition result to the information terminal 20. The method for speech recognition is not particularly limited, and any of currently known methods may be used.

  The speech recognition unit 14, the dictionary control unit 16, and the contribution control unit 18 are realized by a CPU of a computer that operates according to a program (speech recognition dictionary expansion program). For example, the program is stored in a storage unit (not shown) of the service providing server 10, and the CPU reads the program and operates as the voice recognition unit 14, the dictionary control unit 16, and the contribution control unit 18 according to the program. May be. Further, each of the speech recognition unit 14, the dictionary control unit 16, and the contribution control unit 18 may be realized by dedicated hardware.

  The dictionary storage unit 17 and the contribution degree storage unit 19 are realized by, for example, a magnetic disk.

  The registration terminal 30 is an information processing apparatus such as a small personal computer. The registration terminal 30 transmits the term list created by the user to the service providing server 10 via the communication network 100. In addition, the registration terminal 30 displays the data received from the service providing server 10 on a display device (not shown).

  In the above description, the case where the term list is registered from the registration terminal 30 is illustrated, but the registration source of the term list is not limited to the registration terminal 30 and may be another device. For example, registration means (not shown) provided in the information terminal 20 may register the term list in the voice recognition server 10. Moreover, the user who registers a term list may be one person, and plural may be sufficient as it. In the following description, it is assumed that the creator of the term list and the registrant are the same user.

  Next, the operation will be described. FIG. 4 is a flowchart showing an example of the operation of the speech recognition dictionary expansion system according to this embodiment. In the following description, a case where the user “Mr. B” uses a term described in the term list registered by the user “Mr. A” will be described as an example.

  First, when Mr. A registers a term list in the registration terminal 30, the registration terminal 30 transmits the registered term list to the service providing server 10 (step A1). As described above, the term list is a list in which terms and readings are paired as illustrated in FIG.

  Next, when the service providing server 10 receives the term list, the speech recognition dictionary management unit 13 (more specifically, the dictionary control unit 16) determines whether each term described in the term list exists in the existing dictionary. Is determined (step A2). If a term that does not exist in the existing dictionary is included in the term list, the contribution management unit 15 (more specifically, the contribution control unit 18) extracts the term into the contribution word list (step A3). Further, the dictionary control unit 16 takes the terms included in the term list into the existing dictionary and expands the existing dictionary (step A4).

  Here, a method for determining whether or not each term described in the term list exists in the existing dictionary will be specifically described. Terms included in the term list include those with the same reading and different notation (for example, kanji), and those with the same notation and different readings. Examples of terms that have the same reading but different notation include, for example, a case where the reading is “Hashi” and the notation is “bridge” or “chopsticks”. Examples of terms that have the same notation but different readings include, for example, the case where the notation is “Kobe” and the reading is “Kobe”, “Kando”, or “Godo”.

  Therefore, when the service providing server 10 receives the term list, the dictionary control unit 16 searches the existing dictionary and determines whether or not a portion (for example, kanji) corresponding to the notation of the term list is included. If the existing dictionary does not include the corresponding notation, the dictionary control unit 16 determines that the term is a new term and imports the term into the existing dictionary.

  On the other hand, when a notation corresponding to the existing dictionary is included, the dictionary control unit 16 determines whether the reading of the term list corresponding to the notation matches the reading of the existing dictionary. If the readings do not match, the dictionary control unit 16 determines that the term is a new term, and imports the term into the existing dictionary. On the other hand, if both readings match, the dictionary control unit 16 determines that the term is an existing term and does not import the existing dictionary.

  As described above, the dictionary control unit 16 does not match the term notation included in the term list with the term notation in the existing dictionary, or does not match the term reading in the term list with the term reading in the existing dictionary. A term is extracted as a term (new term) that does not exist in the existing dictionary.

  Next, when an operation indicating the start of service use by Mr. B is performed, the information terminal 20 accesses the service providing server 10. At this time, Mr. B is assumed to use voice recognition for the input of the business service (step A5). Specifically, when Mr. B inputs sound to the information terminal 20 via the microphone 22, the sound transmitting / receiving unit 21 transmits the input sound data to the service providing server 10.

  When the service providing server 10 receives the voice data, the voice recognition unit 14 executes voice recognition and transmits a voice recognition result to the information terminal 20 (step A6). The information terminal 20 displays the received voice recognition result on the display unit 23 (step A7).

  Next, in the service providing server 10, the contribution degree control unit 18 determines whether or not the term used for speech recognition is present in the contribution word list (step A9). When the term used for speech recognition is present in the contribution word list, the contribution degree control unit 18 causes the contribution degree storage unit 19 to store contribution degree points. Specifically, a contribution point is given to the user who created the term list from which the term is extracted. In this way, contribution points are accumulated each time a term registered in an existing dictionary is used, a ranking display showing the degree of use, a service that discounts the usage fee for business services when a certain number is reached, etc. To be used.

  Thereafter, when an operation for confirming his / her own contribution point is performed by Mr. A, the registration terminal 30 requests the contribution point from the service providing server 10 in accordance with the operation (step A10). Then, the contribution control unit 18 notifies the registration terminal 30 of Mr. A's contribution points extracted from the contribution storage unit 19 (step A11).

  As described above, according to this embodiment, the dictionary control unit 16 extracts terms that do not exist in the speech recognition dictionary from the term list, and adds the extracted terms to the speech recognition dictionary. Then, the voice recognition unit 14 performs voice recognition of the voice received from the information terminal 20 based on the voice recognition dictionary. With the configuration as described above, the terms registered in the term list (user dictionary) created by each user can be utilized as terms used for speech recognition.

  In addition, when the term included in the speech that has been speech-recognized is a term extracted from the term list, the contribution level control unit 18 may give a contribution point to the user who created the term list. By including such a configuration, a user dictionary can be actively provided to each user.

  Specifically, the dictionary control unit 16 extracts a difference between the registered term list and the existing dictionary, and the contribution degree control unit 18 generates a contribution word list. Each time the speech recognition unit 14 performs speech recognition, the contribution control unit 18 determines whether or not a term included in the speech that has been speech-recognized exists in the contribution word list. The contribution degree control unit 18 gives contribution degree points to the user who created the term list. Therefore, it is possible to motivate the user to register the term.

  Further, the dictionary control unit 16 creates an expanded dictionary by taking the registered term list into the existing dictionary. As a result, the dictionary can be expanded without making the user aware of it, and the accuracy of voice recognition can be improved without the effort of the operator.

  Further, for example, in the dictionary distribution method described in Patent Document 1, when a user dictionary registered by another user is used, it is considered that the user is charged for using the user dictionary. However, if the existing dictionary can be unified and the terms registered by other users can be used within the normal service provided (ie, without special charge), the use of the system is facilitated and the voice Further improvement of the accuracy of the recognition dictionary can be expected. Furthermore, by using the present invention to provide a cloud service that charges so that the more the registered terms are used, the cheaper the fee, the registration of terms is promoted and the accuracy of speech recognition is improved. Can do.

  Next, the minimum configuration of the present invention will be described. FIG. 5 is a block diagram showing an example of the minimum configuration of the speech recognition dictionary expansion apparatus according to the present invention. The speech recognition dictionary expansion apparatus according to the present invention includes a speech recognition dictionary storage unit 81 (for example, dictionary storage unit 17) that stores a speech recognition dictionary (for example, an existing dictionary) that includes a term that is a target of speech recognition, and a speech recognition dictionary. Unregistered term extraction means 82 for extracting a term that does not exist in the speech recognition dictionary from a term list that is a list created by associating the target term with the reading of the term, and adding the extracted term to the speech recognition dictionary (For example, dictionary control unit 16) and speech recognition means 83 (for example, speech recognition unit 14) for performing speech recognition based on the speech recognition dictionary.

  With the above configuration, terms registered in a user dictionary (term list) created by each user can be used as terms used for speech recognition.

  In addition, when the term included in the speech recognized by the speech recognition unit 83 is a term extracted from the term list, a predetermined point (for example, contribution point) is given to the user who created the term list Point giving means (for example, a contribution degree control unit 18) may be provided. By including such a configuration, it is possible to actively provide a user dictionary (term list) to each user.

  The point granting unit creates a term extraction list (for example, a contribution word list) that is a list in which the term extracted by the unregistered term extracting unit 82 is associated with the user who created the term, and the voice recognition unit 83. When the term included in the speech subjected to speech recognition is included in the term extraction list, a predetermined point may be given to the creation source user corresponding to the term.

  FIG. 6 is a block diagram showing an example of the minimum configuration of the speech recognition dictionary expansion system according to the present invention. The speech recognition dictionary expansion system according to the present invention includes a speech recognition server 70 (for example, service providing server 10) that performs speech recognition, and a request terminal 90 (for example, information that requests speech recognition by transmitting speech to the speech recognition server 70). Terminal 20).

  The voice recognition server 70 includes a voice recognition dictionary storage means 71, an unregistered term extraction means 72, and a voice recognition means 73.

  The voice recognition unit 73 (for example, the voice recognition unit 14) performs voice recognition of the voice received from the request terminal 90 based on the voice recognition dictionary. The speech recognition dictionary storage unit 71 and the unregistered term extraction unit 72 are the same as the voice recognition dictionary storage unit 81 and the unregistered term extraction unit 82 illustrated in FIG.

  Even with such a configuration, terms registered in a user dictionary (term list) created by each user can be used as terms used for speech recognition.

  The speech recognition server 70 gives a predetermined point to the user who created the term list when the term included in the speech that the speech recognition unit 73 performed speech recognition is a term extracted from the term list. Point granting means (for example, contribution degree control unit 18) may be included.

  In addition, the point granting unit creates a term extraction list that is a list in which the term extracted by the unregistered term extracting unit 72 is associated with the user who created the term, and the voice recognition unit 73 performs the voice recognition. When a term included in the term extraction list is included in the term extraction list, a predetermined point may be given to the creation source user corresponding to the term.

  Further, the speech recognition dictionary expansion system may include a registration unit (for example, registration unit 30 or registration unit of the information terminal 20) that registers a term list. Then, the unregistered term extraction unit 72 may extract a term that does not exist in the speech recognition dictionary from the term list registered by the registration unit, and add the extracted term to the speech recognition dictionary.

  A part or all of the above-described embodiment can be described as in the following supplementary notes, but is not limited thereto.

(Supplementary Note 1) Extracting a term that is not present in the speech recognition dictionary including a term that is a target of speech recognition from a term list that is a list that is created by associating a term that is a target of speech recognition with a reading of the term A speech recognition dictionary expansion method comprising: adding an extracted term to the speech recognition dictionary and performing speech recognition of a speech received from a requesting terminal that requests speech recognition based on the speech recognition dictionary.

(Supplementary note 2) The speech recognition dictionary expansion method according to supplementary note 1, wherein, when a term included in speech-recognized speech is a term extracted from the term list, a predetermined point is given to a user who creates the term list .

(Supplementary Note 3) Terms included in speech that has been speech-recognized by creating a term extraction list that is a list in which terms extracted from the term list as terms that do not exist in the speech recognition dictionary are associated with the user who created the term The speech recognition dictionary expansion method according to appendix 2, wherein a predetermined point is given to a creation source user corresponding to the term when the term extraction list is included in the term extraction list.

(Supplementary note 4) A speech recognition dictionary expansion program applied to a computer having speech recognition dictionary storage means for storing a speech recognition dictionary including a term to be speech recognition target. Unregistered term extraction processing for extracting a term that does not exist in the speech recognition dictionary from a term list that is a list created by associating a term to be read and a reading of the term, and adding the extracted term to the speech recognition dictionary And a speech recognition dictionary expansion program for executing speech recognition processing for performing speech recognition based on the speech recognition dictionary.

(Additional remark 5) When the term contained in the audio | voice with which speech recognition was performed by the speech recognition process is a term extracted from the said term list to a computer, a predetermined point is given to the creation origin user of the said term list The speech recognition dictionary expansion program according to appendix 4, which executes a point giving process.

(Additional remark 6) Let a computer create the term extraction list | wrist which is the list | wrist which matched the term extracted from the unregistered term extraction process, and the creation origin user of the said term by point provision processing, and performs speech recognition by speech recognition processing. The speech recognition dictionary expansion program according to appendix 5, which gives a predetermined point to a creation source user corresponding to the term when the term included in the speech is included in the term extraction list.

  The present invention is preferably applied to a speech recognition dictionary expansion device that expands a speech recognition dictionary used when a device installed on a communication network performs speech recognition. The present invention can also be applied to information sharing services such as groupware and business information management services.

DESCRIPTION OF SYMBOLS 10 Service provision server 11 ID management part 12 Business service provision part 13 Speech recognition dictionary management part 14 Speech recognition part 15 Contribution management part 16 Dictionary control part 17 Dictionary storage part 18 Contribution control part 19 Contribution storage part 20 Information terminal 21 Audio transmission / reception unit 22 Microphone 23 Display unit 30 Registration terminal 100 Communication network

Claims (10)

Speech recognition dictionary storage means for storing a speech recognition dictionary including terms to be subjected to speech recognition;
Extract a term that does not exist in the speech recognition dictionary from a term list, which is a list that is created by associating a term that is the target of speech recognition with the reading of the term, and add the extracted term to the speech recognition dictionary Term extraction means;
A speech recognition dictionary expansion device comprising speech recognition means for performing speech recognition based on the speech recognition dictionary. The point granting means for giving a predetermined point to a user who created the term list when the term included in the speech subjected to the speech recognition by the voice recognition unit is a term extracted from the term list. The speech recognition dictionary expansion device described. The point granting unit creates a term extraction list that is a list in which the term extracted by the unregistered term extracting unit and the user who created the term are associated with each other, and the term included in the speech that the speech recognition unit has performed speech recognition 3. The speech recognition dictionary expansion device according to claim 2, wherein when a word is included in the term extraction list, a predetermined point is given to a creation user corresponding to the term. Unregistered term extraction means that terms that are included in the term list do not match the terms in the speech recognition dictionary or terms that do not match the terms in the term list and terms in the speech recognition dictionary The speech recognition dictionary expansion device according to any one of claims 1 to 3, wherein: is extracted as a term that does not exist in the speech recognition dictionary. A voice recognition server for voice recognition;
A requesting terminal that transmits voice to the voice recognition server and requests voice recognition;
The voice recognition server
Speech recognition dictionary storage means for storing a speech recognition dictionary including terms to be subjected to speech recognition;
A term that does not exist in the speech recognition dictionary is extracted from a term list that is a list that is created by associating a term that is a target of speech recognition with a reading of the term, and the extracted term is not added to the speech recognition dictionary. Registered term extraction means;
A speech recognition dictionary expansion system comprising speech recognition means for performing speech recognition of speech received from the requesting terminal based on the speech recognition dictionary. 6. A point granting unit for granting a predetermined point to a user who created the term list when the term included in the speech that has been voice-recognized by the voice recognition unit is a term extracted from the term list. The speech recognition dictionary expansion system described. The point granting unit creates a term extraction list that is a list in which the term extracted by the unregistered term extracting unit and the user who created the term are associated with each other, and the term included in the speech that the speech recognition unit has performed speech recognition 7. The speech recognition dictionary expansion system according to claim 6, wherein a predetermined point is given to a creation source user corresponding to the term when the is included in the term extraction list. A registration means for registering a term list;
The unregistered term extraction unit extracts a term that does not exist in the speech recognition dictionary from the term list registered by the registration unit, and adds the extracted term to the speech recognition dictionary. The speech recognition dictionary expansion system according to item 1. Extract a term that does not exist in the speech recognition dictionary that includes the term that is subject to speech recognition from the term list that is a list that is created by associating the term that is subject to speech recognition and the reading of the term,
Adding the extracted terms to the speech recognition dictionary;
A speech recognition dictionary expansion method, wherein speech recognition of speech received from a requesting terminal that requests speech recognition is performed based on the speech recognition dictionary. A speech recognition dictionary expansion program applied to a computer provided with speech recognition dictionary storage means for storing a speech recognition dictionary including a term to be a target of speech recognition,
In the computer,
A term that does not exist in the speech recognition dictionary is extracted from a term list that is a list that is created by associating a term that is a target of speech recognition with a reading of the term, and the extracted term is added to the speech recognition dictionary. Unregistered term extraction process, and
A speech recognition dictionary expansion program for executing speech recognition processing for performing speech recognition based on the speech recognition dictionary.

Images