JPH0376760B2 - - Google Patents

Description

[Detailed Description of the Invention] (1) Technical Field of the Invention The present invention provides a voice recognition system that allows a user to make a voice input request when he or she wants to refer to registered items in a voice recognition device, and to obtain a voice response. This invention relates to a recognition response device.

(2) Prior art and problems Speech recognition devices that target unspecified speakers (unspecified people who input voice) must identify letters and words from the voices of unspecified people, which have different differences. This makes the dictionary for speech recognition complicated,
For this reason, only devices that recognize numbers and extremely simple words are usually put into practical use. Furthermore, it is extremely difficult to add or change words to be recognized. However, when using it, it is convenient in that the user does not have to register the voice in advance.

On the other hand, registration-type speech recognition devices that target a specific speaker (a specific person who inputs voice) have the trouble of having to register the voice in advance; The dictionary for speech recognition is simple, and it is relatively easy to add or change words to be recognized.

When using a voice recognition device targeted at unspecified speakers, for the reasons mentioned above, it is necessary to specify the relevant thing by inputting a numerical code, etc., so registered words (names of products, etc.) The user must memorize the correspondence between the information and the numerical code. However, if the number of registered words is large and changes are made frequently, managing them is not only troublesome, but also leads to frequent mistakes such as entering the wrong numerical code. I had a problem.

On the other hand, when using a registration-type voice recognition device for a specific speaker, you can register the voice using an abbreviation, etc., so it is easier to register the voice than the actual product, etc. compared to the method of inputting using a numerical code etc. mentioned above. Although it is easy to memorize the correspondence, if there are a large number of products, etc., and there are many changes, it may become difficult to know which products are registered under what abbreviations. .

In response to these problems, there has been a problem in the past that even if a user wanted to check the code number or voice-registered abbreviation of a product registered at the center, they could not easily check it. It was hot.

(3) Purpose of the Invention In view of the above conventional problems, the present invention aims to provide a voice recognition response device that allows a user to easily refer to the code number or abbreviation of a product registered at a center. The purpose is

(4) Structure of the Invention According to the present invention, this object is as described in the claims, which includes a speech recognition unit targeted at unspecified speakers, a speech recognition unit targeted at specific speakers,
A speech recognition unit that targets a plurality of specific speakers and has a registered speech recognition dictionary corresponding to each of the plurality of specific speakers, and a voice that synthesizes and outputs speech from information such as input characters. The synthesis unit has the same type of items as the items registered by the user in the speech recognition dictionary for a specific speaker, and the registration contents of the voice to be recognized in the item are the same as the voice input by the user. A file for each user that stores a table registered as characters, etc. that can be input information to the speech synthesis section for creating words with the same meaning by the speech synthesis section and transmitting them as speech, and means for searching the file for the user in response to a user's request via the target speech recognition unit, extracting information such as relevant characters, and transmitting the extracted information to the speech synthesis unit; This is achieved by the provided voice recognition response device with voice registered content inquiry function.

(5) Embodiment of the Invention Figure 1a is a block diagram showing an embodiment of the present invention, in which 1 is a line control unit, 2 is a switching section, 3 is a speech synthesis section, and 4 is intended for a specific speaker. 5 is a dictionary memory, 6 is a speech recognition unit for unspecified speakers, 7 is a control unit, 8 is a display device, 9 is a dictionary for unspecified speakers, and 10 is for specific speakers. 11 is a correspondence table file, 12 is a connection terminal with the host computer, 13 is a center,
14 represents a communication line, and 15 represents a telephone set.

Figure 1a is a diagram showing an example of a correspondence table, in which one part of the correspondence table for the user whose user identification number is 1234 and one part of the correspondence table for the user whose user identification number is 2345.
It represents the department.

In FIG. 1a, a correspondence table as shown in FIG. 1b is stored in the correspondence table file 11. In FIG.
The correspondence table is registered by inputting the contents from the display device 8 at the center upon application from the user, and the abbreviations registered by the user by voice are retained as data for speech synthesis as shown in the figure. There is.

In this system, the user normally dials the telephone 15 to connect to the center and enters data by voice to perform processing. By instructing the system via the voice recognition unit 6 to refer to the correspondence table, it can check whether the code number of the registered product is registered or not, and the product name or abbreviation that corresponds to the code. I can know. At this time, the center uses a system in which the necessary data is synthesized by the voice synthesis section 3 and sent back to the user as voice, so the user receives the result in voice through the receiver of the telephone 15.

FIG. 2 is a diagram showing an example of the user's voice input and the voice output from the center in the case of registered content inquiry,
"a" indicates the case of registration confirmation, and "b" indicates the case of reading inquiry.

Referring to FIG. 2a, when the user inputs voice input of "registration confirmation", the center outputs "product code please". Here, when the user enters the code number of the product they want to check (123 in the diagram) by voice, it will be checked to see if the code number exists in the above-mentioned correspondence table, and the message ``Registered'' reads ``Big Beer''. If the user wants to find out about other code numbers, they can continue to input them, but as shown in the figure, if they input the voice command "End," the registration confirmation ends. . Although it is not shown in the figure, the user identification number is confirmed prior to registration confirmation, and the registered contents for each user are controlled so that only the corresponding correspondence table can be referenced. kept secret.

In the case of FIG. 2b, the operation and control are the same as in the case of a, except that the content of inquiry is different.

(6) Effects of the Invention As explained in detail above, according to the present invention, the content corresponding to the content registered in the dictionary for a specific speaker is stored in the center, and the voice registration item corresponding section is connected to the voice synthesis section. A file with a correspondence table stored in drivable characters, etc. is provided, and the results are outputted as voice by referring to this file as needed. The registered contents can be easily confirmed through voice input through the system, which is very effective as it prevents troubles such as inputting the wrong numerical code or forgetting the registered contents.

[Brief explanation of drawings]

Figure 1a is a block diagram showing one embodiment of the present invention, Figure 1b is a diagram showing an example of a correspondence table, and Figure 2 is a diagram showing the user's voice input and voice output from the center when inquiring registered contents. It is a figure showing an example. 1...Line control unit, 2...Switching unit, 3...
Speech synthesis section, 4... Speech recognition section targeted at a specific speaker, 5... Dictionary memory, 6... Speech recognition section targeted at unspecified speakers, 7... Control section, 8... Display device, 9... Unspecified Dictionary for speaker, 10... Dictionary for specific speaker, 11... Correspondence table file, 12... Connection terminal with host computer, 13... Center, 14...
Communication line, 15...telephone.

Claims (1)

[Scope of Claims] 1. A speech recognition unit that targets unspecified speakers, and a voice that targets a plurality of specific speakers and is provided with a registered speech recognition dictionary corresponding to each of the plurality of specific speakers. A recognition unit, a speech synthesis unit that synthesizes and outputs speech from information such as input characters, and a speech synthesis unit that has the same type of items as the items registered in the speech recognition dictionary for a specific speaker by the user, and Regarding the registration contents of the speech to be recognized in the item, the speech synthesis section creates words with the same meaning as the speech input by the user, and characters that can be input information to the speech synthesis section to be sent as speech. A file for each user that stores tables registered as such, and a file for that user in response to a user's request via a speech recognition unit targeted at unspecified speakers, 1. A voice recognition and response device with voice registration content inquiry function, characterized in that it is provided with means for extracting information such as corresponding characters and transmitting the extracted information to the voice synthesis section.