JP4581290B2 - Speech recognition apparatus and speech recognition method - Google Patents

Description

[0001]
BACKGROUND OF THE INVENTION
The present invention relates to a voice recognition apparatus that performs input by voice to an information terminal.
[0002]
[Prior art]
For example, in a navigation device, a speech recognition device that recognizes a destination or a nearby place name by voice input has a large number of place names that are subject to speech recognition, which causes an increase in recognition time and a decrease in recognition rate. Met. As a method of coping with this problem, a technique is known in which a dictionary of recognition target words is grouped in advance according to a region, and the priority order of each group is determined according to the distance from the current location of a car (Japanese Patent Laid-Open No. 11-325945) (Patent Publication No. 11-183190), a technique for narrowing down words to be recognized according to the current location and traveling direction, the type of road being traveled, and the like.
[0003]
[Problems to be solved by the invention]
Conventionally, a voice recognition device in a car navigation device has been used for purposes such as destination setting and route guidance command input. Therefore, as a technique for narrowing down the recognition target vocabulary, a method of selecting a subset of the recognition target dictionary provided in the speech recognition apparatus based on the current position and route information to the destination has been used. In this method, only registered words are recognized, and there is a problem that more detailed information and the latest information that the speech recognition apparatus does not have cannot be used.
[0004]
In addition, when considering using a car navigation device as an in-vehicle information terminal, the speech recognition device needs to recognize words that do not depend on the distance from the current position, such as information retrieval, news browsing, and schedule confirmation. However, with these technologies, the recognition target vocabulary cannot be efficiently set for these items.
[0005]
The present invention has been made in view of the above problems, and an object of the present invention is to improve the ability to narrow down recognition target words and improve the performance of a speech recognition apparatus.
[0006]
[Means for Solving the Problems]
In order to solve the above problems, a first aspect of the present invention, a speech recognition dictionary speech recognition target words are registered, using the voice recognition dictionary to recognize the speech recognition unit the request of the user as a word A history database that records the words recognized by the speech recognition unit, the recognized time, and the frequency of each recognized word associated with the time, and the frequency that has been recorded in the history database in the past The situation determination unit that determines whether the frequency corresponding to the recognized word and the recognized time is equal to or higher than a threshold, and searches for a web page corresponding to the recognized word through the Internet, An information acquisition unit that acquires a noun recorded in the searched web page, and a noun acquired by the information acquisition unit when the frequency is determined to be greater than or equal to a threshold value Serial includes a dictionary management unit to be registered in the speech recognition dictionary, a.
[0007]
The second aspect of the present invention, Ru-vehicle or portable information terminal der.
[0009]
Moreover , the 3rd aspect of this invention switches the method in which the said dictionary management part registers the noun which the said information acquisition part acquired into the said speech recognition dictionary according to the frequency currently recorded on the said log | history database.
[0010]
DETAILED DESCRIPTION OF THE INVENTION
Embodiments of the present invention will be described below with reference to the drawings.
[0011]
The speech recognition apparatus according to the embodiment of the present invention is based on the user's past behavior and habits on the premise that the car navigation function is set in a voice interactive format or is used on an in-vehicle information terminal that refers to a homepage on the Internet. Based on this, the recognition target vocabulary of the speech recognition dictionary is updated. Since it is desirable to constantly update to a new vocabulary, the speech recognition apparatus of the present invention acquires the recognition target vocabulary from an external database via the Internet.
[0012]
FIG. 1 is a configuration diagram of a speech recognition dictionary update unit of the speech recognition apparatus according to the embodiment of the present invention. In FIG. 1, 101 is a database such as a homepage existing on the Internet, 102 is a history database that holds a user's past history in the form of a schedule, and 103 is a history database 102 based on the user's past history. A history management unit 104 that sequentially updates, a situation determination unit 104 that predicts a user's behavior based on information in the history database 102 and the current date and time, and transmits prediction contents and prediction reliability, and 105 is transmitted from the situation determination unit 104. An information acquisition unit 107 that accesses the Internet based on the user's behavior prediction and extracts keywords on the home page, 107 is a keyword transmitted from the information acquisition unit 105, and a reliability of the schedule transmitted from the situation determination unit 104 Based on the above, the dictionary management unit 106 that updates the recognition target vocabulary of the speech recognition dictionary, Speech recognition dictionary to be used for voice recognition, 108 extracts a user request for vehicle information terminal, a user request extracting unit for transmitting the history management unit 103. FIG. 5 shows a more detailed configuration of the user request extraction unit 108 in FIG. In FIG. 5, reference numeral 110 denotes a button input unit that inputs a user request with a button such as a remote controller, 109 denotes a voice recognition unit that inputs a user request by voice, and 111 denotes a button input unit 110 and the voice recognition unit 109. A navigation control unit 112 that monitors the output and controls the in-vehicle information terminal 112 is a display unit that displays a homepage requested by the user. The database 101, the history database 102, and the history management unit 103 are the same as the components in FIG.
[0013]
FIG. 2 shows an example of history data held in the history database when the user refers to the homepage of the N newspaper. As shown in the figure, items stored in the history database 102 are “date or day of week”, “time zone”, “referenced homepage address”, and “frequency of accessing the same time zone of the same day in the past”. is there. The situation determination unit 104 determines “history reliability” from the frequency information.
[0014]
The operation of the first embodiment of the present invention configured as described above will be described below with reference to FIGS. The operation of the in-vehicle speech recognition apparatus according to the first embodiment of the present invention can be roughly divided into the following two operations. One is an operation of registering / updating the user history database 102, and the other is an operation of updating the speech recognition dictionary 106 using the user history database 102.
[0015]
First, the operation of registering / updating the user history database 102 will be described with reference to FIG.
[0016]
Hereinafter, a case where a user issues a request “read news from the homepage of N newspaper” to the in-vehicle information terminal will be described as an example.
[0017]
First, the user utters “I want to see the homepage of the N newspaper”, the voice recognition unit 109 recognizes the utterance of the user, and the navigation control unit 111 responds to the user's request “I want to see the homepage of the N newspaper”. Send to.
[0018]
The navigation control unit 111 receives the information “I want to see the homepage of the N newspaper” from the voice recognition unit 109, converts it to the operation code “Read Nshinbun.com” of the in-vehicle information terminal, and executes it. That is, the homepage of the N newspaper in the database 101 is referred to, and the corresponding homepage is provided to the user using the display unit 112. The above is a part of normal operation of the in-vehicle information terminal.
[0019]
If the name “N newspaper” is not registered in the speech recognition dictionary 106, the name is input using the button input unit 110.
[0020]
Next, the history management unit 103 receives a user request “Read Nshinbun.com” from the navigation control unit 111, and sequentially stores the history database 102 based on the “user request” and the current date and time. Register / update. In the following, for example, the history of “I referred to Nshinbun.com at 9:23 on Tuesday” is registered / updated.
[0021]
(1) If there is no history in the history database 102 where the address is “Nshinbun.com” and the day of the week / time zone includes “Tuesday 9:23”, new registration is performed. Items to be registered are the homepage address “Nshinbun.com”, the accessed time zone “9:00 to 10:00”, the day of the week “Tuesday”, and the access frequency “once”. Here, the access time is given as one hour from 9:00 to 10:00 which is set in advance.
[0022]
{Circle around (2)} If there is a history in the history database 102 where the address is “Nshinbun.com” and the day / time zone includes “Tuesday 9:23”, the corresponding history is updated. That is, if there is a history in which the home page address is the same and the current time is included in the access time zone, the update operation is performed.
[0023]
The item to be changed is the access frequency, and is updated to “frequency twice”.
[0024]
In the above description, since the history database 102 is automatically registered / updated from the user's terminal operation, the user need not be aware of registration in the history database 102. Of course, it is also possible for the user to register his / her action schedule and daily habits directly in the history database 102.
[0025]
The history database 102 is registered / updated every time a user request is input. However, the history database 102 may be controlled to be updated periodically, or the registered / updated history may be further regularly organized. good.
[0026]
Next, an operation for updating the speech recognition dictionary 106 will be described with reference to FIG. The situation determination unit 104 predicts the user's behavior based on the history in the history database 102 whose access frequency is equal to or higher than the set threshold, the input date, and the current time. The situation determination unit 104 transmits the homepage address described in the history database 102 to the information acquisition unit 105. Further, the reliability of prediction based on the access frequency is transmitted to the dictionary management unit 107. When the access frequency is higher than a preset threshold value, it is possible to determine that it is a habit.
[0027]
The information acquisition unit 105 accesses the homepage of the designated address on the database 101, extracts “noun with a link” from the data as a keyword, and transmits it to the dictionary management unit 107. FIG. 3 shows an example of a noun that is linked on the homepage of N newspaper as an example, and FIG. 4 shows an example of a keyword extracted by the information acquisition unit 105.
[0028]
The dictionary management unit 107 receives the keyword and changes the speech recognition dictionary 106, but switches the change method in three stages according to the prediction reliability transmitted from the situation determination unit 104, and narrows down recognition target words. The change method is shown below in descending order of reliability.
[0029]
(1) Only the keyword received from the information acquisition unit 105 is set in the speech recognition dictionary 106 as a recognition target vocabulary.
[0030]
(2) The keyword received from the information acquisition unit 105 and the words necessary for the operation of the in-vehicle information terminal are set in the speech recognition dictionary 106 as the recognition target vocabulary.
[0031]
(3) The keyword received from the information acquisition unit 105 is added to the speech recognition dictionary 106.
[0032]
When the time period described in the history database 102 ends, the situation determination unit 104 transmits an end to the dictionary management unit 107, and the dictionary management unit 107 updates the speech recognition dictionary 106 to a normal dictionary.
[0033]
As described above, according to the embodiment of the present invention, the speech recognition apparatus automatically updates the speech recognition dictionary based on the past operation history of the user, so that the speech is made prior to the user's request. Since words that are likely to be stored can be held in the speech recognition dictionary, the recognition performance of the speech recognition apparatus can be effectively improved and the recognition time can be shortened.
[0034]
Note that the information acquisition unit 105 can extract keywords even in a hierarchically described home page by accessing the home page of the link destination.
[0035]
In the above embodiment, an in-vehicle information terminal is taken as an example, but this can also be implemented in a portable information terminal.
[0036]
In the above embodiment, the voice recognition dictionary is limited or added. However, the same effect can be expected even if priority is given to the voice recognition dictionary.
[0037]
【The invention's effect】
As described above, according to the present invention, the voice recognition dictionary of the voice recognition apparatus in the information terminal can be uttered ahead of the user's request by dynamically changing the user's behavior by predicting the user's behavior. Since high words can be held in the speech recognition dictionary, it is possible to effectively improve the recognition performance of the speech recognition apparatus and shorten the recognition time.
[Brief description of the drawings]
FIG. 1 is a configuration diagram of a speech recognition dictionary changing unit according to an embodiment of the present invention. FIG. 2 is a diagram showing an example of a history in a history database. FIG. 3 is a diagram showing an example of a home page. Diagram showing examples of keywords [Fig. 5] Configuration diagram of history database registration / update [Explanation of symbols]
101 Database 102 History Database 103 History Management Unit 104 Situation Determination Unit 105 Information Acquisition Unit 106 Speech Recognition Dictionary 107 Dictionary Management Unit 108 User Request Extraction Unit

Claims (7)

  A speech recognition dictionary in which words for speech recognition are registered,
  Using the speech recognition dictionary, a speech recognition unit that recognizes a user's request as a word;
  A history database that records the words recognized by the voice recognition unit, the recognized time, and the frequency of each recognized word associated with the time;
  A situation determination unit that refers to the frequency recorded in the past in the history database and determines whether the frequency corresponding to the recognized word and the recognized time is equal to or greater than a threshold;
  An information acquisition unit that searches a web page corresponding to the recognized word through the Internet and acquires a noun recorded in the searched web page;
  When it is determined that the frequency is equal to or higher than a threshold, a dictionary management unit that registers the noun acquired by the information acquisition unit in the speech recognition dictionary;
A speech recognition apparatus comprising: Automotive or portable information terminal der Ru請 Motomeko 1 speech recognition apparatus according.   According to the frequency recorded in the history database, the dictionary management unit,
Switching the method of registering the noun acquired by the information acquisition unit in the speech recognition dictionary
The speech recognition apparatus according to claim 1.   When it is determined that the frequency is equal to or higher than a threshold value, the dictionary management unit records only the noun acquired by the information acquisition unit in the speech recognition dictionary when the frequency is higher than a certain level.
The speech recognition apparatus according to claim 1.   The dictionary management unit adds the noun acquired by the information acquisition unit to the speech recognition dictionary when the frequency is determined to be equal to or higher than a threshold and the frequency is lower than a certain level.
The speech recognition apparatus according to claim 1.   A speech recognition device for use in an in-vehicle information terminal,
  In the case where the dictionary management unit determines that the frequency is equal to or higher than a threshold value,
If the frequency is greater than or equal to a first threshold, only the nouns acquired by the information acquisition unit are recorded in the speech recognition dictionary,
When the frequency is lower than the first threshold value and greater than or equal to the second threshold value, the noun acquired by the information acquisition unit and the words necessary for the operation of the in-vehicle information terminal are recorded in the voice recognition dictionary. And
When the frequency is lower than the second threshold, the noun acquired by the information acquisition unit is added to the speech recognition dictionary
The speech recognition apparatus according to claim 1.   A speech recognition step of recognizing a user request as a word using a speech recognition dictionary in which speech recognition target words are registered;
  A recording step of recording in a history database that records the word recognized in the speech recognition step, the recognized time, and the frequency of each recognized word associated with the time;
  A situation determination step of referring to the frequency recorded in the history database in the past and determining whether the frequency corresponding to the recognized word and the recognized time is equal to or greater than a threshold;
  An information acquisition step of searching a web page corresponding to the recognized word through the Internet and acquiring a noun recorded in the searched web page;
  When it is determined that the frequency is equal to or higher than a threshold, a dictionary management step of registering the noun acquired in the information acquisition step in the speech recognition dictionary;
A speech recognition method comprising:

Images