JPS59211095A - System of collating contents of voice registration - Google Patents

Abstract

(57) [Summary] This bulletin contains application data before electronic filing, so abstract data is not recorded.

Description

Detailed Description of the Invention (1) Technical Field of the Invention The present invention provides a system in which, when a user wants to refer to registered items in a voice recognition device, he or she can request this by voice input and obtain a response in full voice. (2) Prior art and problems Speech recognition devices aimed at unspecified speakers (unspecified people who input voice input) must identify letters and words from the voices of unspecified people, which have different differences. As a result, speech recognition dictionaries are complicated, and as a result, only those that recognize numbers and extremely simple words are put into practical use.

Furthermore, it is extremely difficult to change all the words to be recognized once they have been added.

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; It is relatively easy to create a dictionary for phonetic recognition because only the dictionary is required, and it is also relatively easy to add and change words to be recognized.

When using a speech recognition device that targets unspecified speakers, for the reasons mentioned above, it is necessary to specify the corresponding thing by inputting a numerical code, etc.
The user must memorize the correspondence between the name of the product, etc. and the numerical code, etc. However, the number of registered words is large,
In addition, when changes are made frequently, it is not only troublesome to manage them, but also leads to errors such as entering the wrong numerical code and making mistakes. .

On the other hand, when using all registration-type voice recognition devices for specific speakers, it is possible to register voices using abbreviations, etc., so compared to the method of inputting numbers using numerical codes, etc., the actual product Although it is easy to remember the correspondence between products, etc., 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. get it.

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. there were.

(3) Purpose of the invention In view of the above-mentioned conventional problems, the present invention aims to provide a system that allows users to easily refer to all code numbers and abbreviations of products registered at a center. .

(4) Structure and purpose of the invention According to the present invention, the present invention provides a speech recognition unit for unspecified speakers and a registered speech recognition unit for specific speakers. In a voice recognition response system having a speech recognition dictionary and a speech synthesis section that performs total speech synthesis from character strings, the system has the same type of items as the items registered by the user in a speech recognition dictionary for a specific speaker, and , Regarding the registration contents of the voice to be recognized in this item, characters that can be input data to the voice synthesis unit to create a word with the same meaning as the voice input by the user and send it as voice. A file is created to store the table registered as such, and in response to a user's request via the speech recognition unit for unspecified speakers, the file is searched and the necessary information is synthesized and output. This is achieved by a voice registration content inquiry method characterized by the following.

(5) Embodiment of the Invention FIG. 1 (, ) 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 a specific block diagram. 5 is a dictionary memory; 6 is a speech recognition unit that targets unspecified speakers;
7 is a control unit, 8 is a display device, 9 is a dictionary for unspecified speakers, 10 is a dictionary for specific speakers, 11 is a correspondence table file, 12 is a connection terminal with the host computer, 13 is a center, and 14 is a・ 4 ・ Communication line, 15 represents a telephone set.

FIG. 1(b) 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. 1 (,), the correspondence table file 11 contains the first
A correspondence table as shown in Figure (b) is stored. The correspondence table is inputted from the display device 8 at the center in response to a user's request, and the contents are all registered, and the abbreviations etc. registered by the user by voice are held as voice synthesis data as shown in the figure.

In this system, the user normally dials the telephone 15, connects to the center, and inputs 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, on the center side, the necessary data is transferred to the voice synthesis section 5.
Since the system employs a system in which the results are synthesized and sent back to the user as voice, the user receives the results in voice form through the handset of the telephone set 15.

Figure 2 is a diagram showing an example of the user's voice input and the voice output from the center in the case of a registered content inquiry, where (a) shows the case of registration confirmation and (b) shows the case of reading inquiry. ing.

Referring to FIG. 2(a), when the user inputs the message ``Turtle registration confirmation'' by voice, the center outputs the message ``Please enter the product code.''

Here, the code number of the product the user wants to check (1 in the figure)
23) When you input the code number by voice, it is checked whether the code number exists in the above-mentioned correspondence table and it is registered. If the user wishes to check for other code numbers, he or she can continue to input them, but the registration confirmation will be completed if he or she inputs N end shi　 as shown in the figure. Although not shown in the diagram, the user identification number is confirmed prior to registration confirmation, and the confidentiality of the registered contents for each user is protected by controls that allow only the corresponding correspondence table to be referenced. has been done.

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

(6) Effects of the Invention As explained in detail above, the method of the present invention corresponds to the content registered in the dictionary for a specific speaker, and
The voice registration item corresponding section provides a file of a correspondence table stored in characters etc. that can drive the voice synthesis section, and refers to this file as necessary to output all voices as a result, Users can easily confirm all registered contents by voice input through the voice recognition unit, which targets unspecified speakers, thereby preventing problems caused by entering the wrong numerical code or forgetting registered contents. Therefore, the effect is large.

[Brief explanation of drawings]

FIG. 1(a) is a block diagram showing one embodiment of the present invention, FIG. 1(b) is a diagram showing an example of a correspondence table, and FIG. 2 is a block diagram showing an example of a correspondence table, and FIG. FIG. 1... Line control unit, 2... Song switching unit, 5... Speech synthesis unit, 4... Speech recognition unit for song specific speaker, 5... Listening unit. Dictionary memory, 6. Speech recognition unit for song-independent speakers, 7. Volume control unit, 8.
... Display device, 9. Dictionary for specific speakers, 10... Dictionary for specific speakers, 11. Song correspondence table 7 isles, 12... Connection terminal with host computer, 15... Center, 14...
...Communication line, 15...Matsu telephone agent, patent attorney
Hiroshi Oka

Claims (1)

[Claims] In a voice recognition response system that includes a voice recognition unit for all unspecified speakers, a registered voice recognition unit for specific speakers, and a voice synthesis unit for synthesizing voice from character strings, the user It has the same type of items as the items registered in the speech recognition dictionary for a specific speaker, and the registered contents of the voice to be recognized in the item include words with the same meaning as the voice input by the user. A speech recognition section targeted at unspecified speakers is provided with a file that stores a table registered as characters, etc. that can be used with the input data of the speech synthesis section to be created by the speech synthesis section and sent out as speech. A voice registration content inquiry method characterized in that, in response to a user's request via a user's request, the file is searched and necessary information is output as a synthesized sound composition.