JPH0465391B2 - - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION [Technical Field of the Invention] The present invention relates to a speech recognition system for improving the recognition rate in speech recognition.

[Technical background of the invention]

As a conventional speech recognition system, there is one shown in FIG. 2, for example. As shown in the figure, this voice recognition system includes a microphone 1 that collects voice, a voice recognition means 3 that registers and recognizes the voice collected by the microphone 1, and a recognition output that does not cause false recognition. It consists of a recognition confirmation means 5 for confirming that the information is correct, a central processing unit (CPU) 7, a read-only memory (ROM) 9, a memory (RAM) 11 that can be read and written at any time, and a data bus 13.

In such a voice recognition system, the voice recognition means 3 is set in the registration mode and the voice is registered in the voice recognition means 3 via the microphone 1, and then the voice recognition means 3 is set in the recognition mode and the voice input from the microphone 1 is recognized. I do.

[Problems with background technology]

In such a speech recognition system, each word registered by the speech recognition means 3 is complicatedly composed of several types of vowels and consonants, and when recognizing the word, in order to obtain the best recognition rate, It is necessary to recognize each word using different word-specific registration and recognition levels.

However, in conventional voice recognition systems, when registering voices, all voices are registered at the same level, so there was a problem that the required recognition rate could not be obtained.

[Purpose of the invention]

The present invention has been made in view of the above problems, and
The purpose is to provide a speech recognition system that can eliminate bias in the recognition rate of each recognized word and make the overall recognition rate the best value.

[Summary of the invention]

In order to achieve the above object, the present invention provides a microphone for collecting sound, a voice recognition means for registering and recognizing the voice input from the microphone, and a variable registration level for the voice registered at the time of voice registration. and level setting means for setting the level.

[Embodiments of the Invention] Examples of the present invention will be described in detail below. FIG. 1 is a block diagram showing the configuration of main parts of a speech recognition system according to an embodiment of the present invention, and elements that perform the same functions as the conventional speech recognition system shown in FIG. 2 are given the same numbers. , and their explanations will be omitted to avoid duplication.

In this embodiment, a level setting means 15 is provided, and the voice collected from the microphone 1 is inputted to the voice recognition means 3 via this level setting means 15. or
The CPU 7 sends level setting information for the voice to be registered to the level setting means 15. Also ROM
9 stores in advance level setting information for the voice to be registered.

Next, the operation of this system will be explained.

First, when performing voice registration, the mode of the voice recognition means 3 is automatically switched to the registration mode by the CPU 7.

Next, the registration address of the registered voice is set, but at this time it is assumed that the registration address of each word is determined in advance for each word. For example “1”
For the word (one), in order to set its registration level to "1", which is the optimal level, a dedicated registration address is set for the word "1" in the voice recognition means 3, and the CPU 7 registers it. When address data is received, the level setting information is stored in the ROM in order to set the level corresponding to that data.
9 to the read level setting means 15.
This level setting information is estimated in advance from experimental results, etc., and stored in the ROM 9. The level setting means 15 that receives this level setting information performs level setting corresponding to the level setting information. As a result, the optimum level setting for the word "1" is completed. That is, if the voice "1" is input from the microphone 1, the registration level in this case will be optimal.

In this way, by registering each registered word in a predetermined location, each word can be registered at an optimal registration level. When recognizing a voice, the voice recognition means 3 is switched to recognition mode by the CPU 7, and the best recognition result can be obtained by inputting voice from the microphone 1.

Therefore, by such an operation, it is possible to obtain uniform recognition characteristics as a whole without seeing a decrease in the recognition rate for specific words, and it is possible to obtain the result that the overall recognition rate is also inevitably improved. be.

〔Effect of the invention〕

As explained in detail above, according to the present invention, the best recognition rate can be obtained by setting the optimal level for each registered word that can be estimated from experimental results etc. in advance, and the user using the system can However, the number of misrecognitions is reduced, the reliability of the system is increased, and the system can be used with peace of mind.

[Brief explanation of the drawing]

FIG. 1 is a block diagram of a speech recognition system according to an embodiment of the present invention, and FIG. 2 is a block diagram of a conventional speech recognition system. 1... Microphone, 3... Voice recognition means, 7
...CPU, 9...ROM, 15...Level setting means.

Claims (1)

[Claims] 1. A microphone that collects sound, a voice recognition means that registers and recognizes the voice input from the microphone, and a level setting means that varies and sets a registration level for each voice for the voice registered at the time of voice registration. A voice recognition system comprising: