JP3139679B2 - Voice input device and voice input method - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a voice input for recognizing an input voice as a word and selecting an input item for inputting the word from a plurality of input items predetermined from the meaning of the word. The present invention relates to a device and a voice input method.

[0002]

2. Description of the Related Art Conventionally, a voice input device for performing input from a screen by voice has been developed. This apparatus inputs data by voice so that even a user who is unfamiliar with the operation of a computer or the like can easily use it.

[0003] A voice input device recognizes a voice or the like uttered by a person as a word and performs processing according to the word. When entering data, it is necessary to specify the input items. As such a voice input device, Japanese Patent Laid-Open No. 8-1
Japanese Patent No. 29476 discloses a voice input device which is specified by uttering an input item of the data together with the data.

FIG. 3 is a flowchart for explaining the operation of the conventional speech recognition. The voice input device described in the official gazette is applied to a pathological examination support system.

First, a user instructs a pathological finding or an input item thereof by voice while looking at an object to be inspected by a microscope or the like (step S21). The voice generated at this time is collected by a microphone or the like and input to the voice recognition / synthesis device (step S22). The input voice is output to a voice recognition unit.

The voice recognition unit recognizes a word associated with the voice from the input voice (step S23). The recognized word is transmitted to the OS such as Windows in a code format (step S25). In addition, in order to allow the user to confirm whether the input voice matches the recognized word, the recognized sound is output from headphones or the like (step S26).

The recognized word is output from the OS to the pathological examination support system, which is an application program (step S27). Pathological examination support system
If the input word is related to a pathological finding, the word is displayed in a desired item on the screen of the display device (step S28). On the other hand, if the input word relates to an input item, a desired word can be input to the input item (step S29).

In such a conventional technique, data input can be performed only by voice. Further, since the recognized word is answered by voice, it is possible to confirm whether the input voice matches the recognized voice without looking at the screen of the display device. Therefore, when inputting voice, it is not necessary to interrupt the work such as observing the inspection object.

[0009]

However, the above-mentioned prior art has the following problems.

First, when a conventional voice input device is used, a user inputs a data input item by voice or the like. Therefore, in order to use the voice input device, the user needs to know what input items are provided as a prerequisite.

Therefore, if the user does not know the input items provided, he or she must look at the screen of the display device and check the input items displayed on the display screen before giving an instruction.

Further, the conventional voice input device has a function of recognizing an input voice as a word as described above. However, it does not have a function of converting a word related to input data into another word of the same concept. Therefore, even if words with the same concept are input, the words may be treated as different words in the apparatus.

[0013] In order to solve the above problems, it is an object of the present invention to specify a word related to a voice and an input item thereof from a voice uttered by a user.

[0014]

In order to solve the above problems BRIEF SUMMARY OF THE INVENTION The present invention includes an input unit for inputting a voice, the input
A recognition unit for recognizing a word from the voice waveform of the voice input by the unit, and a plurality of input items for inputting the word in advance are provided.
Which words are recognized by the recognition means
A selector for selecting belongs to force entry, selected by the selection unit
Enter the word recognized by the recognition unit in the selected input item
A voice input device having a display unit for displaying in a state where the
The word recognized by the recognition unit and the selection unit
Outputs the input item selected by using synthesized speech
Responding to user responses based on output results.
The word recognized by the recognition unit is selected by the selection unit.
Displayed in the input field or prompt the user to re-enter the voice
It is characterized by the following .

Further, the present invention inputs a voice, to recognize words from voice speech waveform the input, pre-word
A plurality of input items to be input are provided and the
To select which input item the word belongs to ,
State words were allowed to input to the input item is the selected
In a speech input method for display, the recognized allowed a single
Output words and selected input items by synthesized speech
And select the recognized word based on the output result
Display the input items as they are entered or re-enter the audio.
It is characterized by performing . In addition, the present invention provides for audio input.
An input unit for inputting, and a voice input by the input unit.
Recognition unit for recognizing words from speech waveforms and pre-input words
There are a plurality of input items to be
Select which input item the recognized word belongs to
A selection unit and the input item selected by the selection unit
A display unit that displays the words recognized by the unit as they are input
A method of using a voice input device comprising:
The force device is configured to control the word recognized by the recognition unit and the selection.
The input item selected by the selector
A response unit for outputting, based on an output result of the response unit.
The word recognized by the recognition unit is selected by the selection unit
Display or re-input the voice
It is characterized by the following.

[0016]

Embodiments of the present invention will be described below with reference to the drawings. In this embodiment,
The voice input device is applied, for example, to software for a household account book.

FIG. 1 is a block diagram showing a voice input device according to this embodiment. In FIG. 1, a voice input device includes an input unit 11 for inputting a voice uttered by a user, a recognition unit 12 for recognizing a voice input from the input unit 11 as a word,
A recognition dictionary 13 which is connected to the recognition unit 12 and in which information for recognizing voice is registered.

A selecting unit 14 for selecting an input item of the word recognized by the recognizing unit 12, a selection dictionary 15 connected to the selecting unit 14 and registering information for selecting the input item of the word; A response unit 16 that responds by voice between the recognized word and the selected input item, an output unit 17 that outputs data of the recognized word and the selected input item, and an output unit 17 that outputs the recognized word and the selected input item. It has a storage unit 18 for storing input items, and a display unit 19 for displaying words and their input items.

FIG. 2 is a diagram showing a household account book table displayed on the display unit 19. In FIG. 2, the first line of the household account book table includes columns of “Breakdown”, “Product Name”, “Amount”, and “Remarks”. For example, as a breakdown of food, columns such as dairy products, staple food, meat / fish, and so on are provided. Further, in the breakdown of clothing, columns such as clothing, shoes, and so on are provided.

FIG. 3 is a flowchart showing the operation of the voice input device of this embodiment. Here, for example,
An example in which the input data is milk and its price is 200 yen will be described.

First, when the user utters "milk, 200 yen", the input unit 11 such as a microphone inputs the voice (step S1). The input voice is recognized by the recognition unit 12.
Is output to The recognizing unit 12 removes unnecessary noise and the like for recognizing the voice by using a filter (not shown) for the voice. A / D converter (not shown) converts the A / D
The audio waveform is digitized by D-conversion.

The digitized speech waveform is stored in the recognition dictionary 13
Is compared with the waveform of the word registered in. The comparison is H
This is performed using a technique such as MM (Hidden Markov Model). The input voice is regarded as a word having a waveform closest to the voice waveform among words registered in the recognition dictionary 13. Thus, the input voice "milk, 200 yen" is recognized as two words "milk" and "200 yen" (step S2).

Next, the recognized word is input to the selection unit 14. The selection unit 14 extracts selection information registered in the selection dictionary 15. In the selection dictionary 15, words related to items such as "food" and "clothing" are registered collectively. Then, from the selection information, the input item to which the recognized word should be output is selected (step S3). The selection information will be described later in detail.

When the input item of "milk" is selected as "food" by the selection unit 14, "milk" is data to be input into the breakdown of "food", and "200 yen" is "
Is output from the response section 16 as synthesized speech from the response section 16 (step S4). For example, the synthesized speech is “food, milk” and “amount, 20 yen”. This allows the user to say “milk, 20
It can be confirmed that the voice "0 yen" has been correctly recognized by the voice input device (step S5).

When the voice uttered by the user matches the synthesized voice uttered from the response unit 16, the user utters the voice to that effect to select the recognized word. An input item can be specified (step S7).

On the other hand, if the voice uttered by the user does not match the synthesized voice uttered from the response unit 16, that is, if a synthesized voice other than "milk" or "¥ 200" is output from the response unit 16, The user can correct an error by issuing a corrected word or the like (step S6).

When the word and the input item thereof are specified in this way, the output unit 17 inputs the word "milk" in the "food" column and the word "200 yen" in the "amount" column. Is output to the storage unit 18 and the display unit 19. The storage unit 18 stores the input word and its input items. The display unit 19 displays the input word on the specific input item (step S8).

Next, the selection information registered in the selection dictionary 15 will be described. The selection information is a summary of various types of information as described below. That is, the first information of the selection information is for selecting an input item corresponding to the input word. The recognition unit 12 selects which of the predetermined items the word recognized by the recognition unit 12 is related to.

The second is information such as a character format to be displayed on the display screen of the display unit 19. For example, it is information such as a character type such as kanji, hiragana, and katakana, and a character size such as half-width or full-width.

The third is speech information of the synthesized speech output from the response unit 16. For example, in the selection dictionary 15,
If voice information such as "Which input item do you want to input data?" Is registered and there are a plurality of candidates to be data input items, the response unit 16 outputs as synthesized voice based on this voice information. Is done. This information will be described in a second embodiment.

Fourth, the selection information is information for converting a word uttered by the user into a word having the same concept.
Therefore, for example, even when the user utters “milk”, the word “milk” can be converted to the word “milk” and displayed on the display unit 19.

(Embodiment 2) FIG. 4 shows a household account book displayed on a display unit 19 of this embodiment. FIG.
, The first row of the household account book contains "breakdown", "article name",
It consists of "Amount" and "Remarks" columns. For example, the income is provided with fields such as husband, wife,... In the "article name" column, a salary,...

FIG. 5 is a flowchart showing the operation of the voice input device of this embodiment. Here, for example,
The case where the input data is “salary, 200,000 yen” and the user is a husband will be described as an example.

First, the husband who is the user says, "Salary, 200,000
When saying "circle", the input unit 11 such as a microphone inputs the voice (step S11). The same processing as in the first embodiment is performed on the input voice. In the recognizing unit 12, an A / D converter (not shown) performs A / D conversion to digitize the audio waveform.

The digitized voice waveform is compared with the word waveform registered in the recognition dictionary 13, and the input voice of "salary, 200,000 yen" is converted to "salary", "200,0
00 yen "(step S
12).

Next, the recognized word is input to the selection unit 14. The selection unit 14 extracts selection information registered in the selection dictionary 15. Then, from the selection information, the input item to which the recognized word should be output is selected (step S13).

When the input item of "salary" is selected as "article name", "salary" is data to be entered in the breakdown of "article name", and "200,000 yen" is entered in the "amount" column. Is output from the response unit 16 as a synthesized voice (step S14). As a result, the user can confirm that the voice "salary, 200,000 yen" is correctly recognized.

As in the first embodiment, the user can specify and correct the recognized word and the selected input item (steps S15 to S17).

Here, there may be a case where there are a plurality of word input item candidates. In the software of the household account book of this embodiment, input items of husband and wife are provided as “breakdown” in the income column. In such a case, the user may be asked "
200,000 yen ", it is not possible to determine which is the input item.

Therefore, in this embodiment, the response unit 16
Output a synthesized voice saying "Is the input item a husband? Or a wife?" (Step S18). Thus, the user can specify an input item by saying “husband” (step S19).

When the word and the input item thereof are specified in this way, the output unit 17 inputs that the word "salary" is entered in the "article name" column and "200,000 yen" is entered in the "amount" column. Is output to the storage unit 18 and the display unit 19. The storage unit 18 stores the input word and its input items. The display unit 19 displays the input word on the specific input item (step S2).
0).

In each of the above embodiments, the case where the voice input device is applied to household account book software has been described as an example. However, these voice input devices are also applicable to a pathological examination support system or the like. Can be.

[0043]

According to the present invention, when a user utters a voice, a word relating to the voice and an input item thereof are selected. Then, the selection result is output as a synthesized voice and the user is specified. Therefore, the user does not need to look at the display screen to confirm the input item every time the user speaks. Therefore, it is possible to concentrate on reading out a receipt or a form.

Further, according to the present invention, even when there are a plurality of candidates for the input item of the voice uttered by the user, it is possible to output the effect to the user by the synthesized voice and to specify the input item. Therefore, similarly to the above, it is not necessary to look at the display screen to confirm the word input items.

Further, the voice input device of the present invention has information for converting a voice uttered by the user into a word having the same concept. For this reason, if the sounds uttered by the user are of the same concept, the storage and display of the input sounds will not be mixed.

[Brief description of the drawings]

FIG. 1 is a configuration diagram of a voice input device according to a first embodiment.

FIG. 2 is a diagram illustrating a household account book table displayed on a display unit according to the first embodiment.

FIG. 3 is a flowchart illustrating an operation of the voice input device according to the first embodiment.

FIG. 4 is a diagram illustrating a household account book table displayed on a display unit according to a second embodiment.

FIG. 5 is a flowchart illustrating an operation of the voice input device according to the second embodiment.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 11 Input part 12 Recognition part 13 Recognition dictionary 14 Selection part 15 Selection dictionary 16 Response part 17 Output part 18 Storage part 19 Display part

Claims (6)

(57) [Claims] 1. An input unit for inputting a voice, and the input unit
Therefore, a recognition unit that recognizes a word from the voice waveform of the input voice and a plurality of input items for inputting the word in advance are provided.
Which input item is the word recognized by the recognizing means.
A selector for selecting belongs eye, it is selected by the selection unit
Input the word recognized by the recognition unit to the input item
A voice input device having a display unit for displaying in a state.
The word recognized by the recognition unit and the selection unit
Output the selected input item with synthesized speech
Answer part, and respond to the user's response based on the output result
The word recognized by the recognizing unit is input to the word selected by the selecting unit.
A voice input device which is displayed on a force item or prompts the user to re-input a voice. 2. The method according to claim 1, wherein the selecting unit is connected to a selection dictionary in which words controlled by the plurality of input items are registered collectively, and the selection unit extracts the input items from the words registered in the selection dictionary. The voice input device according to claim 1, wherein the voice input device is selected. 3. The speech input device according to claim 1, wherein, when there are a plurality of candidates for the input item of the recognized word, the response unit outputs the fact by a synthesized speech. 4. The voice input device according to claim 2, wherein the selection dictionary includes information for converting the recognized word into a word having the same concept. 5. Enter the voice, to recognize words from voice speech waveform the input, the input terms advance word is input
Which input words eyes were and the recognition provided with a plurality
A voice for allowing the user to select whether the item belongs to an item and displaying the recognized word in a state where the word is input to the selected input item.
An input method, wherein the recognized word and the selected input item are combined.
Output by voice and recognize based on output results
Words are displayed as they are entered in the selected input item.
Is a voice input method characterized by re-inputting voice. 6. An input unit for inputting voice, and said input unit
Therefore, a word is recognized from the speech waveform of the input speech.
Multiple input items to input words in advance Provided
Which input item is the word recognized by the recognizing means.
A selection unit for selecting whether the object belongs to the eyes,
Input the word recognized by the recognition unit to the input item
How to use a voice input device with a display unit for displaying in a state
Law, The voice input device is a unit that is recognized by the recognition unit.
Synthesized sound between word and input item selected by the selection unit
Equipped with a response unit that outputs by voice, Recognized by the recognition unit based on an output result of the response unit
Words in the input items selected in the selection section.
Voice input,
How to use force device.