JPH11194795A - Speech recognition operating device - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide the speech recognition operating device which reduces effect of noise without using any high-performance expensive component and prevents the cost from increasing. SOLUTION: Of a portable telephone set 10 equipped with a voice input part 12 which inputs a speaker's voice and outputs an electric voice signal, a speech recognition part 18 which inputs the voice signal and recognizes the speaker's voice, an information transmission part 30 which transmits information to the user, and a control part 20 which controls registration, recognition, and other specific processing operation according to the voice signal and the signal from the voice recognition part 18, the control part 20 controls the information transmission part 30 to sends a predetermined alarm to the user at registration when judging that noise exceeding a reference level is mixed in addition to human voices according to the voice signal.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a voice recognition operating device that recognizes voice input by a human and operates appropriately based on the voice.

[0002]

2. Description of the Related Art As a conventional voice recognition operating device, for example, a telephone subscriber's name (name) corresponding to a dialed number is directed to a microphone as described in Japanese Patent Application Laid-Open No. 61-144157. By speaking,
2. Description of the Related Art There is a telephone that can recognize a name by a voice recognition unit of the telephone and automatically call a corresponding dial number.

[0003] Such a telephone can be used for an automobile telephone or a portable telephone, and it is possible to make a telephone call without manually operating a dial or a numeric keypad, and even if the hand is busy for other work. You can make a phone call.

However, in such a conventional voice recognition operating device, when a nominal name corresponding to a dialed number is registered, a voice input is performed in a loud place.
In addition to human speech, noise is also input from the microphone, and this noise causes noise to be included in the registered nominal name pattern, hindering speech recognition and making phone calls smoothly Becomes difficult.

In order to solve such a problem,
A speech recognition operating device (Japanese Patent Publication No. 2-15897) that eliminates the speaker's ambient noise and does not recognize the noise so that only the original speech data can be reliably recognized, or from input speech data. 2. Description of the Related Art A speech recognition operation device (Japanese Patent Laid-Open Publication No. 3-3540) has been proposed which removes ambient noise to obtain speech data with very little noise.

[0006]

However, in the conventional voice recognition operating device disclosed in the above publication, a large amount of arithmetic processing is required to eliminate (remove) ambient noise from original voice data. Is required, for example, a DSP (digital soundfield) used for speech recognition.
For example, a high-performance or high-priced device such as a d processor is required, resulting in an increase in the cost of the apparatus.

Accordingly, the present invention has been made in view of the above problems, and provides a voice recognition operating device capable of reducing the influence of noise without using high-performance and expensive parts and preventing an increase in cost. The task is to do so.

[0008]

In order to solve the above-mentioned problems, a voice recognition operating device according to the present invention comprises a voice input section for inputting a voice of a speaker and outputting an electrical voice signal; , A voice recognition unit that recognizes the speaker's voice by inputting, an information transmission unit that transmits information to the user, and registration based on the voice signal and a signal from the voice recognition unit.
In a voice recognition operation device including a recognition unit and a control unit for controlling other predetermined processing operations, the control unit is configured to include a noise equal to or higher than a reference level in addition to human voice based on the voice signal at the time of registration. When it is determined that the information is present, a predetermined warning is transmitted to the user by the information transmitting unit.

According to the voice recognition operating device having such a configuration, when the control unit determines at the time of registration that noise of a reference level or more is mixed in addition to the human voice based on the voice signal, the information transmission unit. By transmitting a predetermined warning to the user, it is possible to prevent the registered voice to be verified from being registered with noise at the time of voice input at the time of calling, and at the time of voice input at the time of calling. It is possible to prevent difficulty in recognizing voice.

[0010]

Embodiments of the present invention will be specifically described below with reference to the drawings. 1 and 2
FIG. 1 is a diagram referred to for describing a portable telephone 10 according to a first embodiment of a voice recognition operating device according to the present invention.

In the portable telephone 10 shown in FIG.
Voice input unit 1 to which voice is input when talking with the other party
2 (microphone), a voice signal is sent to the communication circuit 34, and the voice signal is further subjected to noise removal by the noise removing unit 23 included in the communication circuit 34, and then sent to the high frequency unit 24 to be directed from the antenna 24a to the other party. Radio is transmitted to the nearest base station.

When a response from the other party is received by the antenna 24a from the base station by radio, the received signal is
4 is sent to the speech circuit 34, the noise is removed by the noise removing unit 23, and then the speech is sent from the speech circuit 34 to the audio output unit 26 (speaker). The person can hear the voice of the response.

When calling the other party, the operation unit 36
The call can be made by pressing the dial number with the key of (1), but when the other party's nominal name is registered in the nominal name pattern storage section (registering section) 22, the user inputs the nominal name into the voice input section 12 , It is possible to automatically call the other party by the dial circuit included in the communication circuit 34.

In this case, the audio signal from the audio input unit 12 is amplified by the amplifier unit 14, converted from an analog signal to a digital signal by the A / D unit 16 (analog / digital converter), and the digital signal is input. The recognized voice recognition unit 18 recognizes the input voice and compares and selects the same name from the nominal names registered in the nominal name pattern storage unit 22.

When the voice recognition unit 18 outputs a signal related to the nominal name to the control unit 20, the control unit 20 selects one of the dial numbers registered in the dial number storage unit 38 in advance so as to correspond to the nominal name. By calling the associated dial number and sending it to the dial circuit of the call circuit 34, a call is automatically made.

The portable telephone 10 has a recording / reproducing section 40 composed of an IC memory or the like, and can record a voice during a call or in an answering machine mode, and reproduce it later. The portable telephone 10 has a temperature measurement sensor 42. The temperature measurement sensor 42 measures the temperature of the environment and notifies the control unit 20 of the temperature. Can be controlled so as not to deteriorate.

The portable telephone 10 has a timer 44.
Is provided, and a clock signal is always output to the control unit 20 to advance the date and time displayed on the display unit 30.

In FIG. 2, the sound input to the sound input unit 12 is converted into an electric sound signal and sent to the amplifier unit 14. This audio signal is amplified by a predetermined gain (amplification rate) in the amplifier unit 14, and is amplified by the A / D unit 16.
(Analog / digital converter). Further, the audio signal is converted from an analog signal to a digital signal in the A / D unit 16, and the digital signal of the audio signal is sent to each of the audio recognition unit 18 and the control unit 20.

If there is no noise in the voice input to the voice input unit 12, the digital signal from the A / D unit 16 is correctly recognized by the voice recognition unit 18 at the time of calling, and the correct partner station is recognized. Is called, and the dialing device can call the correct partner station. At the time of registration, it is possible to register a clear voice name.

However, if the input voice at the time of registration contains noise of a predetermined reference level or more in addition to the human voice, the voice from the A / D unit 16 is compared with the internal memory value. When the determination is made by the control unit 20, the control unit 20 outputs a control signal to the display unit 30 (information transmission unit) and controls the display unit 30 to display a predetermined display for transmitting a predetermined warning to the user.

For example, the display unit 30 displays "the current voice was not a voice in a good environment for registration". By displaying such a message, the user is informed that there is a possibility that the voice may not be correctly recognized at the time of calling even if the current registration is performed, and the user is confirmed whether the registration is still performed.

[0022] By doing so, it is possible to prevent registration of voice that may not be able to recognize voice correctly at the time of calling. As another display, it may be displayed that "registration is not possible because voice may not be correctly recognized at the time of calling", and control may be performed to prohibit registration at the same time.

FIG. 3 is a diagram showing a main part of a portable telephone according to a second embodiment of the present invention. In the second embodiment, an audio signal from an audio input unit 12 is input to an audio recognition unit 18 via an amplifier unit 14 and an A / D unit 16 and a recording / reproducing unit 32 composed of an IC memory or the like. The audio signals from the audio recognition unit 18 and the recording / reproducing unit 32 are input to the control unit 20.

In the control unit 20, information on the length of the feature amount recognized by the voice recognition unit 18, for example, the LPC cepstrum (Transactions of the Institute of Electronics and Communication Engineers, Vol. J65-DN)
o. 5 "Evaluation of LPC Spectrum Matching Scale in Word Speech Recognition" May 1982)
Comparing the data length of the voice section in the recording / reproducing unit 32, the ratio of each data length falls within a certain range. On the other hand, when the noise of the voice input to the voice input unit 12 is equal to or higher than a predetermined reference level, the ratio of the lengths of the noises often differs greatly.

For this reason, when the ratios of the two lengths are largely different from each other, the control unit 20 outputs a control signal to the display unit 30 and provides a predetermined signal to the user, as in the first embodiment. Is controlled so as to display a predetermined display for transmitting the warning, or to prohibit registration.

FIG. 4 is a diagram showing a main part of a portable telephone according to a third embodiment of the present invention. In the third embodiment, an input from the audio input unit 12 and the amplification unit 14, A /
The nominal name of the audio signal that has passed through the D section 16 is registered in the nominal name pattern storage section 22 after being input and recognized by the audio recognition section 18, and the signal relating to this nominal name is input to the control section 20. Then, the control unit 20 registers the dial number corresponding to the nominal name in the dial number storage unit 38 in association with the signal related to the nominal name.

When the speaker inputs voice again from the voice input unit 12 to confirm the registration, the control unit 2
0 indicates that the speech recognition unit 18 compares information on the length of the recognition feature amount, for example, information on the length of the LPC cepstrum, with the data length of the nominal name registered in advance in the nominal name pattern storage unit 22, and If the ratios of the two lengths are significantly different from each other, it is determined that the noise is equal to or higher than the predetermined reference level, and the control unit 20 outputs a control signal to the display unit 30 as in the first embodiment, and Is controlled so as to display a predetermined warning for transmitting a predetermined warning to the user, or to prohibit registration.

In the above embodiment, the case where the present invention is applied to the portable telephone 10 has been described.
The present invention need not be limited to the mobile phone 10, but may be other types of phones, such as car phones, or facsimile,
The present invention can be applied to any other device or device such as a personal computer.

In the above embodiment, the display unit 30 is used as the information transmitting unit. However, a voice warning message may be output using the audio output unit 26 as the information transmitting unit.

Although the embodiments of the present invention have been specifically described above, the present invention is not limited to the above-described embodiments, and based on the technical idea of the present invention, various other embodiments are possible. Can be changed.

[0031]

As described above, according to the voice recognition operating device of the present invention, the control unit determines at the time of registration that a noise equal to or higher than the reference level is mixed in addition to the human voice based on the voice signal. In this case, a predetermined warning is transmitted to the user by the information transmitting unit, so that it is possible to prevent the registered voice to be collated when the voice of the call is input from being registered with noise. It is possible to prevent difficulty in recognizing a voice at the time of voice input. For this reason, the influence of noise can be reduced without using high-performance or expensive components, and the cost of the apparatus can be prevented from increasing.

[Brief description of the drawings]

FIG. 1 is a block circuit diagram showing a configuration of a portable telephone 10 according to a first embodiment of a voice recognition operation device according to the present invention.

FIG. 2 is a block circuit diagram showing a main part of the portable telephone 10 in FIG.

FIG. 3 is a block circuit diagram showing a main part of a portable telephone according to a second embodiment of the voice recognition operating device according to the present invention.

FIG. 4 is a block circuit diagram showing a main part of a portable telephone according to a third embodiment of the voice recognition operating device according to the present invention.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 10 Mobile telephone 12 Audio input part 14 Amplifier part 16 A / D part 18 Voice recognition part 20 Control part 22 Nominal name pattern storage part 23 Noise removal part 24 High frequency part 24a Antenna 26 Audio output part 30 Display part 32 Recording / reproducing part 34 Communication circuit 36 Operation unit 38 Dial number storage unit 40 Recording and playback unit 42 Temperature measurement sensor 44 Timer

──────────────────────────────────────────────────の Continued on the front page (51) Int.Cl. ⁶ Identification code FI H04M 1/27 H04B 7/26 109Q

Claims (3)

[Claims] A voice input unit for inputting a voice of a speaker and outputting an electrical voice signal; a voice recognition unit for inputting the voice signal and recognizing the voice of the speaker; And a control unit for controlling registration, recognition, and other predetermined processing operations based on the voice signal and the signal from the voice recognition unit. The unit, when determining at the time of registration based on the voice signal that the noise of the reference level or more in addition to the human voice is mixed, a predetermined warning is transmitted to the user by the information transmission unit. Characteristic speech recognition operating device. 2. A voice input unit for inputting a voice of a speaker and outputting an electrical voice signal; a voice recognition unit for receiving the voice signal and recognizing the voice of the speaker; A recording / reproducing unit capable of recording a voice and reproducing the voice; an information transmitting unit for transmitting information to a user; a registration / recognition / other predetermined based on the voice signal or a signal from the voice recognition unit. A control unit that controls a processing operation, comprising: a control unit that controls a length of a feature amount recognized by the voice recognition unit based on the voice signal and a voice recorded in the recording / playback unit. Wherein the data transmission unit transmits a predetermined warning to the user when the ratio of the two lengths is significantly different from each other. 3. A voice input unit for inputting a voice of a speaker and outputting an electrical voice signal; a voice recognition unit for inputting the voice signal and recognizing the voice of the speaker; A registration unit for registering a name in association with a dial number; an information transmission unit for transmitting information to a user; registration, recognition, and other predetermined processing based on the voice signal or a signal from the voice recognition unit A voice recognition operation device comprising: a control unit configured to control an operation; and wherein the control unit is configured to store data of a length of a feature amount recognized by the voice recognition unit based on the voice signal and a name registered in the registration unit. A voice recognition operating device, wherein a predetermined warning is transmitted to the user by the information transmitting unit when the ratio of the two lengths is significantly different from each other.