JP2000122686A - Speech recognizer, and electronic equipment using same - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a speech recognizer capable of reducing lowering of a speech recognition rate compared with a standard voice signal, even relating to a voice signal different from a standard in its pronunciation and word. SOLUTION: This speech recognizer 16 is provided with a voice input part 12 input with a sound signal, a voice dictionary storing part 18, a voice dictionary selecting part 26, and a speech recognizing part 38. The voice dictionary storing part 18 stores plural dictionaries 22 classified by attributes, which are different by the attributes such as an area, a race, a social stratum, an individual, and in which voice informations are stored. The voice dictionary selecting part 26 selects the dictionary 22 corresponding to the sound signal based on a feature of the sound signal. The speech recognizing part 38 speech-recognizes the sound signal using the dictionary 22 selected by the selecting part 26.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a voice recognition device and an electronic device using the same.

[0002]

2. Description of the Related Art Conventionally, there has been known a speech recognition apparatus having a speech dictionary containing phoneme data and word information, and recognizing and processing human speech input through a microphone or the like using the speech dictionary. By mounting such a voice recognition device, an electronic device that can recognize a voice when a user utters a predetermined voice and execute a corresponding operation has been developed. This makes it possible for the user to use the electronic device without performing complicated panel operations or the like.

[0003]

However, since the dictionary provided in the above-described speech recognition apparatus is prepared only for a speech dictionary corresponding to a speech signal composed of standard pronunciations, it differs depending on attributes such as area and occupation. There is a problem that the recognition rate is reduced in recognizing a speech signal composed of phonemes or words different from the generated standard.

In an electronic device equipped with a voice recognition device, a predetermined voice message may be answered in response to a voice input from a user. The device will return a standard voice response message,
There was a problem of lack of moisture.

The present invention has been made in view of the above-described problems, and has as its object to input a speech signal different from the standard pronunciation due to a difference in attributes of a person who performs speech input. Even in such a case, an object of the present invention is to provide a speech recognition device capable of performing speech recognition with high accuracy.

Another object of the present invention is to provide an electronic device capable of returning a response message in accordance with attributes such as the area and occupation to which the person who performed the voice input belongs.

[0007]

In order to achieve the above object, a speech recognition apparatus according to the first aspect of the present invention comprises a speech input unit to which a speech signal is input and attributes such as area, ethnicity, hierarchy, and individual. A different voice dictionary storage unit that stores a plurality of attribute-based dictionaries containing voice information, and a voice dictionary selection unit that selects the attribute-based dictionary corresponding to the voice signal based on features in the voice signal. And a voice recognition unit that recognizes the voice signal using the attribute dictionary selected by the voice dictionary selection unit.

[0008] According to the voice recognition device of the first aspect,
The voice dictionary selecting unit selects an attribute dictionary containing voice information of the attribute corresponding to the input voice signal, and the voice recognition unit recognizes the voice signal using the voice dictionary. Therefore, the voice recognition device of the present invention can recognize a voice signal having an attribute different from the standard attribute at a voice recognition rate substantially equal to that of the voice signal corresponding to the standard attribute.

According to a second aspect of the present invention, there is provided an electronic apparatus comprising: a voice recognition device according to the first aspect; and an attribute-specific message storage unit storing a spoken message of each attribute for each attribute corresponding to the attribute dictionary. A message selecting unit for selecting, from the attribute-specific message storage unit, a message in a spoken language having an attribute corresponding to the attribute-specific dictionary selected by the voice dictionary selecting unit; and outputting the message selected by the message selecting unit as a voice signal. And an output unit.

[0010] According to the speech recognition apparatus of the second aspect,
The message selector selects and outputs a message in the spoken language of the attribute corresponding to the voice dictionary selected by the voice dictionary selector in accordance with the attribute of the input voice signal from the attribute-specific message storage unit in which the attribute-specific message is stored. The unit outputs the audio signal. Therefore, the electronic device of the present invention can output a message having an attribute corresponding to the attribute of the input audio signal.

According to a third aspect of the present invention, there is provided an electronic apparatus, wherein: a voice input unit to which a voice signal is input; and an attribute-based message in which a spoken language of each attribute is stored for each attribute such as area, ethnicity, hierarchy, and individual A message storage unit, a message selection unit that selects the attribute-specific message corresponding to the audio signal based on a feature in the audio signal, and an output unit that outputs the message selected by the message selection unit as an audio signal. It is characterized by having.

According to the electronic device of the third aspect, a message having an attribute corresponding to the input audio signal is converted from a message stored in the attribute-specific message storage unit based on a feature in the input audio signal. The selection section selects,
The output unit outputs the message as an audio signal. Therefore, the electronic device of the present invention can output a message having an attribute corresponding to the input signal as a voice signal.

According to a fourth aspect of the present invention, there is provided an electronic apparatus, wherein a number information recognition unit for recognizing telephone number information of a communication partner, a voice input unit for inputting a voice signal, and voice information different depending on the telephone number information are stored. A voice dictionary storage unit that stores a voice dictionary including the obtained number information dictionary, and a corresponding number information dictionary from the voice dictionary storage unit based on the telephone number information recognized by the number information recognition unit. A recognition dictionary extraction unit for extracting a recognition dictionary including the recognition dictionary; a speech recognition unit for recognizing the speech signal based on the recognition dictionary extracted by the recognition dictionary extraction unit; A message storage unit for each number information in which a message in a spoken language is stored for each number information corresponding to the message, and a message in a spoken language corresponding to the dictionary for each number information selected by the voice dictionary selecting unit. Characterized in that it has a message selector for selecting from the number information by the message storage unit, and an output unit for outputting as a voice signal messages the message selector has selected.

According to the electronic device of the fourth aspect, the recognition dictionary extraction unit extracts a recognition dictionary including a dictionary for each number information corresponding to the telephone number information of the communication partner recognized by the number information recognition unit, The voice recognition unit recognizes the voice signal using the recognition dictionary, the message selection unit selects a message corresponding to the telephone number information from the message storage unit for each number information, and the output unit transmits the message to the voice signal. Output as Therefore, the electronic device according to the present invention can recognize a voice signal at a high voice recognition rate using a dictionary for number information corresponding to telephone number information of a communication partner, for example, a station number. The electronic device can output a spoken message corresponding to the telephone number information of the communication partner as voice.

[0015]

DESCRIPTION OF THE PREFERRED EMBODIMENTS Preferred embodiments of the present invention will be described below more specifically with reference to the drawings.

[First Embodiment] In the first embodiment, a case where the electronic apparatus of the present invention is applied to an intercom will be described as an example.

<Configuration of Intercom> Intercom 10
1 is shown in FIG. 1 as a perspective view. Intercom 10
When the button 11 is pressed to enter a call state, the voice of the visitor or the like is received by the microphone 14 and the voice of the respondent or the like is emitted from the speaker.

FIG. 2 is a block diagram showing an electrical configuration of the intercom 10. As shown in FIG. As shown in this figure, the intercom 10 is, for example, a microphone 14
Voice input unit 12 configured to include voice
An output unit 52 configured to include a speaker 64 for outputting sound and including a speaker 64;
A ROM 17 including a program storage unit 23 and a voice dictionary storage unit 18 storing various programs executed by the PU 15 and an E including an attribute-specific message storage unit.
It is configured to include an EPROM 47 and a RAM 43 including a recognition dictionary storage unit 44 and the like.

Further, a functional outline of the intercom 10 is shown in FIG. 3 as a block diagram.

As shown in FIG. 1, functionally, the intercom 10 includes a voice input unit 12 having a microphone 14, a voice recognition device 16 for recognizing the input voice signal, and a message output having a speaker 64. Device 4
0 is formed.

The speech recognition device 16 includes a speech dictionary storage section 18, a speech dictionary selection section 26, a recognition dictionary extraction section 30, a recognition dictionary storage section 44, and a speech recognition section 38. Is done.

The voice dictionary storage unit 18 is provided in the ROM 17 and stores a voice dictionary including a standard dictionary 20 and a plurality of attribute-specific dictionaries 22. The attribute-based dictionary 22 includes information on regions, ethnic groups,
It stores audio information such as phonemes and words that vary depending on attributes such as hierarchy and individual.

The voice dictionary selection unit 26 is configured to include a voice dictionary selection program stored in the ROM 17 and the CPU 15, and corresponds to a voice signal based on features in the voice signal input to the voice input unit 12. The attribute-based dictionary 22 is selected.

The recognition dictionary extraction unit 30 includes a recognition dictionary extraction program stored in the ROM 17 and the CPU 15. The recognition dictionary extraction unit 30 includes the attribute dictionary 22 selected by the speech dictionary selection unit 26 and includes a recognition dictionary from the speech dictionary. Extract a dictionary. For example, the recognition dictionary extracting unit 30 extracts the attribute dictionary 22 selected by the speech dictionary selecting unit 26 and the words of the standard dictionary 20 excluding the portions corresponding to the respective words of the attribute dictionary 22.

The recognition dictionary storage unit 44 is provided in the RAM 43, and stores the recognition dictionary extraction unit 30 from the speech dictionary storage unit 18.
Store the extracted recognition dictionary. In addition, the recognition dictionary storage unit 44 is such that the voice dictionary extraction unit 30
Alternatively, only the code corresponding to the word extracted from may be stored.

The voice recognition unit 38 is configured to include a voice recognition program stored in the ROM 17 and the CPU 15, and converts a voice signal using a recognition dictionary including the attribute-specific dictionary 22 selected by the voice dictionary selection unit 26. Recognize voice.

The message output device 40 includes an attribute-specific message storage unit 48, a message selection unit 34, and an output unit 5.
2 is included.

The attribute-specific message storage unit 48 stores the EEPR
The OM 47 stores an attribute-specific message in spoken language of each attribute for each attribute such as region, ethnicity, hierarchy, and individual.

The message selecting section 34 includes a message selecting program stored in the ROM 17 and the CPU 15, and outputs a message in a spoken language having an attribute corresponding to the attribute dictionary 22 selected by the voice dictionary selecting section 26.
It is selected from the attribute-specific message storage unit 48. For example, the message selection unit 34 selects a message corresponding to the content and attribute of the voice signal recognized by the voice recognition unit 38 of the voice recognition device 16 from the attribute-specific message storage unit 48.

The output unit 52 includes a D / A, an amplifier (not shown), a speaker 64, and the like. The message selected by the message selection unit 34 is converted into a speaker signal as an audio signal via the D / A, the amplifier, and the like. 64.

The control unit 24 includes the above-described speech dictionary selection unit 26, recognition dictionary extraction unit 30, and message selection unit 34. The above-described units constituting the intercom 10, which is the electronic apparatus of the present embodiment, are connected to each other via the signal line 56 shown in FIGS.

The voice dictionary storage unit 18, the recognition dictionary storage unit 44, the attribute-specific message storage unit 48, and the program storage unit 23 do not necessarily need to be constituted by the above-described elements. It can also be configured by any of storage elements such as a storage device or a combination thereof.

<Operation of Intercom> FIG. 4 is a block diagram showing a basic operation of the intercom 10 as the electronic apparatus of the present embodiment. Here, an example of the operation of the intercom 10 will be described with reference to FIG.

First, as shown in step S31,
An audio signal is input via the microphone 14 of the audio input unit 12.

Then, the voice dictionary selecting unit first determines whether or not the input voice signal has the standard attribute as shown in step S32. For example, the speech signal (word, phrase, etc.) input first is first recognized by the speech recognition unit 38 using the standard dictionary 20. If the speech signal can be recognized, it is determined that the attribute is a standard attribute. If not,
Recognition processing is performed using one of the plurality of attribute-based dictionaries 22. If a recognizable attribute-based dictionary 22 is found, it is determined that the attribute is not a standard attribute. If all the attributes cannot be recognized by the attribute-based dictionary 22, the attribute may be determined as a standard attribute. If the input audio signal has the standard attribute, the process proceeds to step S33. If the input audio signal is not the standard attribute, the process proceeds to step S37.

If the input audio signal has the standard attribute, as shown in step S33, the audio dictionary selecting section 26
Selects only the standard dictionary and proceeds to step S34.

In step S34, the recognition dictionary extracting unit 3
A value of 0 extracts the standard dictionary 20 selected by the voice dictionary selection unit 26 from the voice dictionary storage unit, and stores it in the recognition dictionary storage unit 44. Note that the recognition dictionary storage unit 44 may store only codes corresponding to the words extracted by the speech dictionary extraction unit 30. The voice recognition unit 38 performs voice recognition, for example, by comparing a voice signal of a word stored in the recognition dictionary storage unit 44 with a voice signal input to the voice input unit. Then, the process proceeds to step S35.

In step S35, the message selecting section 34 selects an attribute-specific message of the standard attribute from the attribute-specific message storage section 48 based on the content of the voice signal recognized by the voice recognition section 38. That is, the attribute of the selected attribute-based message also corresponds to the attribute of the attribute-based dictionary 22 selected by the voice dictionary selecting unit 26.

In step S36, the audio signal of the attribute-specific message selected in step S35 is output from the speaker 64 of the output unit 52.

Step S37 is a step to be performed when the voice dictionary selecting unit 26 determines in step S32 that the voice signal input to the voice input unit 12 is not a standard attribute. In step S37, the voice dictionary selecting unit 2
6 is an attribute dictionary 2 for attributes corresponding to the input audio signal.
2 and the standard dictionary 20 are selected. Then, step S3
Proceed to 8.

In step S38, the recognition dictionary extraction unit 3
0 indicates, for example, the attribute dictionary 22 selected by the voice dictionary selecting unit 26 and the contents of the standard dictionary 20 excluding words corresponding to each word in the attribute dictionary 22 from the standard dictionary 20. , And stored in the recognition dictionary storage unit 44. Note that the recognition dictionary storage unit 44 may store only codes corresponding to the words extracted by the speech dictionary extraction unit 30. The voice recognition unit 38 performs voice recognition by comparing the voice signal of the word stored in the recognition dictionary storage unit 44 with the voice signal input to the voice input unit 12, for example. Then, step S39
Proceed to.

In step S39, the message selecting section 34 selects, from the attribute-specific message storage section 48, an attribute-specific message corresponding to the content of the voice signal recognized by the voice recognition section 38 in step S39 and its attribute.
Then, as in the case where the voice signal input to the voice input unit 12 has the standard attribute, the process proceeds to S36, and the voice signal of the message is output from the output unit 52.

The above is an example of the operation of the intercom 10 shown in FIG.

Therefore, the intercom 1 of the present embodiment
0, for example, as an input speech signal in step S31, the visitor's voice is input as "(hello)" in expression are frequently used in a specific geographic area, step S3
In step S2, the voice dictionary selecting unit 26 determines that the attribute of the voice signal is that of a specific area, and
In step S38, the speech recognition unit 38 performs speech recognition using the recognition dictionary including the attribute-based dictionary 22 in step S38. In S39, the message selection unit 34 selects a message corresponding to the specific area described above, and if there is a message to be responded to the input voice, the process proceeds to step S36, and the output unit outputs, for example, "( Hello. I hear your business.) "
Can be answered in the expression of the specific area. In addition, the intercom 10 of the present embodiment corresponds to the input of a voice signal (for example, a standard language) having a standard attribute,
You can answer with a message with the corresponding content of the standard attribute, or respond to the input of the audio signal of another attribute (for example, US English) and answer with the message with the corresponding content of the attribute (US English). You can do it.

As described above, the speech recognition device 16 used in the intercom 10 as the electronic device of the present embodiment.
The voice dictionary selecting unit 26 selects the attribute dictionary 22 containing voice information of the attribute corresponding to the input voice signal, and the voice recognition unit 38 recognizes the voice signal using the attribute dictionary 22. . Therefore, this speech recognition device 16
A speech signal having an attribute different from the standard attribute can be speech-recognized at a speech recognition rate substantially equal to that of the speech signal corresponding to the standard attribute.

Further, the intercom 10 as the electronic device of the present embodiment converts a message in a spoken language having an attribute corresponding to the voice dictionary selected by the voice dictionary selecting unit 26 according to the attribute of the input voice signal into a message for each attribute. Is selected by the message selecting unit 40 from the attribute-specific message storage unit 48 in which the message is stored, and the output unit 52 outputs the audio signal. Therefore, intercom 10 can output a message having an attribute corresponding to the attribute of the input audio signal.

[Second Embodiment] In the second embodiment, a case will be described in which the electronic apparatus of the present invention is applied to a telephone function of a facsimile apparatus.

<Configuration of Facsimile Apparatus> The appearance of a facsimile apparatus 70 as an electronic apparatus of the present embodiment is shown in FIG.
Is shown as a perspective view.

As shown in this figure, the facsimile machine 7
The handset 72 is attached to the side of the main body 74 of the telephone. The receiver 72 is placed on a hook (not shown) provided on the main body 74 during a non-call, and is picked up from the hook during a call. An operation panel 66 having a plurality of buttons is provided at a front portion of the upper surface of the main body 74, and a liquid crystal display panel (LCD) 66 is provided at a left rear portion of the operation panel 66. When a button provided on the operation panel 66 of the facsimile apparatus 70 is operated, the operated state, operation procedure, and the like are displayed on the LCD 67.

A document insertion port 68 is provided at the rear of the operation panel 66 and the LCD 67. After the original to be transmitted by facsimile is inserted into the original insertion slot 68 and read, the operation panel 6 is located on the front of the main body 74.
The document is discharged from a document discharge port 69 provided below the document 6.
A recording paper holder mounting portion 65 is provided at the rear of the document insertion opening 68, and a recording paper holder 63 capable of stacking and storing a plurality of recording papers is detachably attached to the recording paper holder mounting portion 65. Have been. The recording paper supplied and printed from the recording paper holder 63 is discharged from the recording paper discharge port 62 provided below the document discharge port 69.

Next, FIG. 6 is a block diagram showing an outline of an electric configuration of the facsimile apparatus 70.

The facsimile device 70 includes a CPU 61,
ROM 59, EEPROM 58, RAM 57, image memory 55, audio memory 54, audio LSI 53, network control unit (hereinafter referred to as "NCU") 51, modem 50, buffer 49, scanner 46,
An encoding unit 45, a decoding unit 42, a printer 41, an operation panel 66, an amplifier 39, and an LCD 67 are provided, and these are connected via a signal line 75.

The CPU 61 controls each unit connected to the signal line 75 in accordance with various signals transmitted and received via the NCU 51 to perform a facsimile operation, a telephone operation, and the like.

The ROM 59 includes a program storage section 32 for storing a control program executed by the facsimile machine 70 and the like, and is a non-rewritable storage element. The program shown as a flowchart in FIG. 8 is stored in the ROM 59.

The EEPROM 58 is a nonvolatile storage element that can be rewritten by electrical control.

The RAM 57 is a storage element for temporarily storing various data when each operation of the facsimile machine 70 is executed.

The image memory 55 is a storage element for storing a communication history, image data, and a bit image for printing, and is constituted by a DRAM. The received image data is temporarily stored in an image memory and printed on recording paper by the printer 41, or
After the image data is transferred to another device via the device 51, the image data is erased from the image memory 55.

The voice memory 54 is a storage element for storing a response message sent to the partner device and an incoming message sent from the partner device. Voice memory 54
Is composed of, for example, a DRAM.

The audio LSI 53 converts an analog audio signal received by the NCU 51 into a digital audio signal, and converts a digital audio signal generated inside the facsimile device 70 into an analog audio signal to convert the analog audio signal into an NCU.
51.

The modem 50 modulates and demodulates image information and various procedural signals for transmission control, and transmits and receives them.

The buffer 49 temporarily stores data including coded image information transmitted / received to / from the other party's facsimile machine.

The scanner 46 is for reading an image from a document inserted into the document insertion port 68.

The encoding section 45 encodes an image read by the scanner 46.

The decoding section 42 reads out and decodes the image data stored in the buffer 49 or the image memory 55.

The printer 41 prints the decoded image data on recording paper.

The operation panel 66 is used by an operator to operate the facsimile machine 70.

The amplifier 39 drives a speaker 99 connected to the amplifier 39 to output a ringing tone and a voice sound.

The facsimile apparatus 70 configured as described above is connected to the telephone line 60 via the NCU 51. The telephone line 60 is connected to an exchange 35, and the exchange 35 is connected to another exchange via the telephone line 33.

Next, an outline of a functional configuration of the facsimile machine 70 is shown in FIG. 7 as a block diagram.

As shown in this figure, the facsimile machine 7
0 is a communication control unit 76 functioning as an interface for exchanging information via the telephone line 60;
Number information recognition unit 78 for recognizing telephone number information of a communication partner
, A voice input unit 77 including a receiver 72, a voice recognition device 80 for recognizing the input voice signal, and a speaker 9.
9 and a message output device 92 having the same.

The speech recognition apparatus 80 includes a speech dictionary storage section 82, a speech dictionary selection section 85, a recognition dictionary extraction section 86, a recognition dictionary storage section 88, and a speech recognition section 90. Is done.

The voice dictionary storage unit 82 is provided in the ROM 59, and stores a voice dictionary including a telephone number information dictionary 84, which differs according to telephone number information, for example, a telephone number, and is a telephone number information dictionary 84 in which voice information is stored. The station number-specific dictionary 84 stores voice information such as phonemes and words that differ depending on the number information, for example, the station number.

The voice dictionary selecting section 85 includes a voice dictionary selecting program stored in the ROM 59 and the CPU 61. The voice dictionary selecting section 85 responds to voice signals based on telephone number information recognized by the number information recognizing section 78, for example, a station number. The station number dictionary 84 is selected as the number information dictionary.

The recognition dictionary extraction unit 86 includes a recognition dictionary extraction unit stored in the ROM 59 and the CPU 61, and extracts a recognition dictionary including the station number dictionary 84 selected by the speech dictionary selection unit 85. For example, the recognition dictionary extraction unit 8
6 is a dictionary 84 for each station number selected by the voice dictionary selecting section 85.
Then, the vocabulary of the standard dictionary 83 excluding a portion corresponding to the vocabulary of the dictionary 84 according to the station number is extracted.

The recognition dictionary storage unit 88 is provided in the RAM 57, and stores the recognition dictionary extraction unit 86
Store the extracted recognition dictionary. The recognition dictionary storage unit 88 stores a code such as each word constituting the recognition dictionary extracted by the recognition dictionary extraction unit 86, and extracts each word from the speech dictionary storage unit 82 according to the code. It may be.

The speech recognition section 90 includes a speech recognition program stored in the ROM 59 and the CPU 61. The speech recognition section 90 uses the recognition dictionary including the station number-specific dictionary 84 selected by the speech dictionary selection section 85 to generate a speech signal. Is recognized.

The message output unit 92 includes a station number-specific message storage unit 96 as a number information-specific message storage unit.
, A message selection unit 94, and an output unit 98.

The station number-specific message storage section 96 as the number information-specific message storage section is provided in the EEPROM 58, and stores a station number-specific message in spoken language corresponding to each station number for each station number as telephone number information.

The message selection unit 94 includes a message selection program stored in the ROM 59 and the CPU 61. The message selection unit 94 stores a message in a spoken language corresponding to the station number dictionary 84 selected by the voice dictionary selection unit 85, and stores the message by station number. Select from section 96. For example, the message selection unit 94 selects a message corresponding to a station number as telephone number information recognized by the number information recognition unit 78 from the station number-specific message storage unit 96.

The output section 98 includes the audio LSI 53, the amplifier 3
9 and a speaker 99, and outputs the message selected by the message selection unit 94 as an audio signal.

The control section 97 includes the above-described speech dictionary selecting section 85, recognition dictionary extracting section 86, and message selecting section 94.

The above-described components constituting the facsimile apparatus 70, which is an electronic apparatus according to the present embodiment, are connected to each other via signal lines 75 shown in FIGS.

The voice dictionary storage unit 82, the recognition dictionary storage unit 88, the station number-specific message storage unit 96, and the program storage unit do not necessarily need to be constituted by the above-described elements, and may include ROM, DRAM, EEPROM, magnetic storage, and the like. It can also be configured by any of the storage elements such as the device or a combination thereof.

<Operation of Facsimile Apparatus> FIG. 8 is a block diagram showing a basic operation of the facsimile apparatus 70 as the electronic apparatus according to the present embodiment. Hereinafter, an example of the operation of the facsimile machine 70 will be described with reference to FIG.

First, as shown in step S61,
The number information recognition unit 78 recognizes telephone number information of a communication partner, for example, a station number.

Next, as shown as step S62,
An audio signal is input via the receiver 72 of the audio input unit 71 or the communication control unit 76 connected to the telephone line 60.

Then, as shown in step S63, the voice dictionary selecting section 85 determines whether or not the telephone number information recognized by the number information recognizing section 78, for example, the station number is a standard station number. Proceed to S64, and if not a standard station number, proceed to step S68.

If the station number recognized by the number information recognizing section 78 is a standard station number, as shown in step S64, the voice dictionary selecting section 85 selects only the standard dictionary 83 and proceeds to step S65.

In step S65, the recognition dictionary extraction unit 8
6 extracts the standard dictionary 83 selected by the voice dictionary selection unit 85 from the voice dictionary storage unit 82 and stores it in the recognition dictionary storage unit 88. The recognition dictionary storage unit 88 may store only codes corresponding to the words extracted by the recognition dictionary extraction unit 86. Then, the voice recognition unit 90
For example, by comparing the audio signal of the word stored in the recognition dictionary storage unit 88 with the input audio signal,
Perform voice recognition. Then, the process proceeds to step S66.

In step S66, the message selecting section 94 outputs a message for each number information corresponding to the contents of the voice signal recognized by the voice recognition section 90 and the number information recognized by the number information recognition section 78, for example, the station number, for example, for each station number. A message is selected from the station number-specific message storage unit 96.

Then, in step S67, the voice signal of the message for each station number selected in step S66 is output from output section 98 and transmitted to telephone line 60 via communication control section 76.

[0092] Step S68 is a step to be performed when the speech dictionary selecting section 85 determines in step S63 that the station number is not a standard station number based on the recognition result of the number information recognizing section 78. In this step, the voice dictionary selecting unit 85 selects the station number-specific dictionary 84 and the standard dictionary 83 corresponding to the station number recognized by the number information recognizing unit 78. Then, the recognition dictionary extraction unit 86, for example, extracts the dictionary for each station number selected by the speech dictionary selection unit 85 and the contents of the standard dictionary 83 except for the words corresponding to the words in the dictionary for each station number 84 from the standard dictionary 83. Is extracted from the voice dictionary storage unit 82 and stored in the recognition dictionary storage unit 88.

Next, in step S69, the voice recognition unit 90 performs voice recognition by comparing the voice signal of the word stored in the recognition dictionary storage unit 88 with the input voice signal, for example. Then, the process proceeds to step S70.

In step S70, the message selecting section 94 outputs a message for each station number corresponding to the contents of the voice signal recognized by the voice recognition section 90 and the number information recognized by the number information recognition section, for example, the station number in step S69. Select from the different message storage unit 96. Then, as in the case where the number information recognized by the number information recognition unit, for example, the station number is standard, the process proceeds to S67, the message is output from the output unit 98, and transmitted to the telephone line via the communication control unit 76. Can be

The above is an example of the operation of the facsimile machine 70 shown in FIG.

As described above, the facsimile apparatus 70 of this embodiment receives a call from, for example, the direction of Nagoya, and the number information recognizing unit 78 recognizes that it is a specific station number in step S61. as input, the voice took telephone is input and its expression, for example, corresponding to the local station number "(hello)" speech dictionary selection unit 85 in step S63 is based on the recognition of the number information recognition section 78 Is not standard telephone number information. In that case, the process proceeds to step S68, in which the recognition dictionary extraction unit 86 extracts the recognition dictionary including the station number dictionary 84 corresponding to the station number and stores the dictionary in the recognition dictionary storage unit 88. In step S69, the speech recognition is performed. part 90 is speech recognition using the recognition dictionary, the message selector 94 selects the station number by a message of a corresponding station number with the corresponding contents in step S70, the willing output unit, for example, "(hello to step S67. Your request will be asked.) ”(In the format corresponding to the area of the area code).

As described above, according to the facsimile apparatus 70 as the electronic apparatus of the present embodiment, the number information recognition section 7
The recognition dictionary extraction unit 86 extracts a recognition dictionary including a dictionary for number information corresponding to the telephone number information of the communication partner recognized by the communication unit 8, and the speech recognition unit 90 recognizes a speech signal using the recognition dictionary. The message selection unit 94 selects a message corresponding to the recognition result and the telephone number information from the message storage unit for each number information, and the output unit 98 outputs the message as a voice signal. Therefore, the facsimile apparatus 70 of the present embodiment can recognize a voice signal at a high voice recognition rate using a telephone number information of a communication partner, for example, a dictionary for number information corresponding to a station number. The facsimile device 70 can output a spoken message corresponding to the telephone number information of the communication partner as voice.

Although the embodiments of the present invention have been described above, the present invention is not limited to the above-described embodiments.
Various modifications can be made within the scope of the present invention or within the equivalent scope of the claims.

For example, in the above embodiment, an example of an intercom or a facsimile apparatus has been described as an electronic apparatus. However, the present invention is applicable to various OA apparatuses such as a copier, a television, a radio, a lighting apparatus, an air conditioner, home appliances, The present invention can be applied to electronic devices such as industrial devices.

In the above-described embodiment, the example in which the voice dictionary storage unit, the recognition dictionary storage unit, and the station number-specific message storage unit are separately provided has been described. If stored, the voice dictionary storage unit, the recognition dictionary storage unit, and the station number-specific message storage unit may be the same storage unit.

[0101]

As described above, according to the speech recognition apparatus of the first aspect, the speech dictionary selection unit selects the attribute dictionary containing the speech information of the attribute corresponding to the input speech signal. The speech recognition unit recognizes the speech signal using the attribute-based dictionary. Therefore, the voice recognition device of the present invention can recognize a voice signal having an attribute different from the standard attribute at a voice recognition rate substantially equal to that of the voice signal corresponding to the standard attribute.

According to the speech recognition apparatus of the second aspect,
The message selector selects and outputs a message in the spoken language of the attribute corresponding to the voice dictionary selected by the voice dictionary selector in accordance with the attribute of the input voice signal from the attribute-specific message storage unit in which the attribute-specific message is stored. The unit outputs the audio signal. Therefore, the electronic device of the present invention can output a message having an attribute corresponding to the attribute of the input audio signal.

According to the electronic device of the third aspect, based on the characteristics in the input audio signal, the message of the attribute corresponding to the audio signal is converted from the message stored in the attribute-specific message storage unit. The selection section selects,
The output unit outputs the message as an audio signal. Therefore, the electronic device of the present invention can output a message having an attribute corresponding to the input signal as a voice signal.

According to the electronic device of the fourth aspect, the recognition dictionary extraction unit extracts a recognition dictionary including a dictionary for each number information corresponding to the telephone number information of the communication partner recognized by the number information recognition unit, The voice recognition unit recognizes the voice signal using the recognition dictionary, the message selection unit selects a message corresponding to the telephone number information from the message storage unit for each number information, and the output unit transmits the message to the voice signal. Output as Therefore, the electronic device according to the present invention can recognize a voice signal at a high voice recognition rate using a dictionary for number information corresponding to telephone number information of a communication partner, for example, a station number. The electronic device can output a spoken message corresponding to the telephone number information of the communication partner as voice.

[Brief description of the drawings]

FIG. 1 is a perspective view showing an intercom as an electronic apparatus according to a first embodiment.

FIG. 2 is a block diagram illustrating an electrical configuration of the intercom according to the first embodiment.

FIG. 3 is a block diagram illustrating a functional configuration of the intercom according to the first embodiment.

FIG. 4 is a flowchart showing an operation of the intercom of the first embodiment.

FIG. 5 is a perspective view showing a facsimile apparatus as an electronic apparatus according to a second embodiment.

FIG. 6 is a block diagram illustrating an electrical configuration of a facsimile apparatus according to a second embodiment.

FIG. 7 is a block diagram illustrating a functional configuration of a facsimile apparatus according to a second embodiment.

FIG. 8 is a flowchart illustrating an operation of the facsimile apparatus according to the second embodiment.

[Explanation of symbols]

Reference Signs List 10 intercom (electronic device) 12, 77 voice input unit 16, 80 voice recognition device 18, 82 voice dictionary storage unit 22 attribute-based dictionary 26, 85 voice dictionary selection unit 30, 86 recognition dictionary extraction unit 34, 94 message selection unit 38,90 Voice recognition unit 44,88 Recognition dictionary storage unit 48 Message storage unit for each attribute 52,98 Output unit 70 Facsimile device (electronic device) 78 Number information recognition unit 84 Dictionary by area number (dictionary by number information) 96 By area number Message storage (message storage by number information)

──────────────────────────────────────────────────続 き Continued on the front page (51) Int.Cl. ⁷ Identification symbol FI Theme coat ゛ (Reference) H04M 9/00 H04M 9/00 D H04N 1/32 H04N 1/32 Z F term (Reference) 5C075 BA08 FF02 FF90 5D015 GG00 HH13 KK01 KK02 KK04 LL09 LL10 5K024 BB01 BB04 CC00 CC08 EE09 5K036 AA00 AA15 BB11 DD25 JJ10 JJ16 5K038 AA06 DD16 GG04

Claims (4)

[Claims] A voice input unit to which a voice signal is input; a voice dictionary storage unit that stores a plurality of attribute-based dictionaries containing voice information, which differ according to attributes such as region, ethnicity, hierarchy, and individual; A voice dictionary selecting unit that selects the attribute dictionary corresponding to the voice signal based on a feature in the voice signal; and voice recognition of the voice signal using the attribute dictionary selected by the voice dictionary selecting unit. A voice recognition device comprising: 2. The speech recognition device according to claim 1, wherein an attribute-specific message storage unit storing a spoken message of each attribute for each attribute corresponding to the attribute dictionary; and A message selection unit for selecting a message in the spoken language of the attribute corresponding to the attribute-based dictionary from the message storage unit for attribute, and an output unit for outputting the message selected by the message selection unit as a voice signal. And electronic equipment. 3. An audio input unit to which an audio signal is input; an attribute-specific message storage unit storing an attribute-specific message in a spoken language of each attribute for each attribute such as region, ethnicity, hierarchy, and individual; An electronic device, comprising: a message selection unit that selects the attribute-specific message corresponding to the audio signal based on a feature in the message; and an output unit that outputs the message selected by the message selection unit as an audio signal. machine. 4. A number information recognizing unit for recognizing telephone number information of a communication partner, a voice input unit for inputting a voice signal, and a dictionary for number information containing voice information which differs depending on the telephone number information. A voice dictionary storage unit for storing a voice dictionary, and a recognition for extracting a recognition dictionary including a corresponding dictionary for each number information from the voice dictionary storage unit based on the telephone number information recognized by the number information recognition unit. And a voice recognition unit that recognizes the voice signal based on the recognition dictionary extracted by the recognition dictionary extraction unit. A message storage unit storing a message according to the number information, and a message in a spoken language corresponding to the dictionary according to the number information selected by the voice dictionary selecting unit. An electronic device comprising: a message selection unit that selects from a message storage unit; and an output unit that outputs a message selected by the message selection unit as an audio signal.