JP3129893B2 - Voice input word processor - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a speech input word processor , and more particularly to a speech input word processor capable of outputting input data as it is when the input data for a language cannot be recognized. Related to voice input word processor .

[0002]

2. Description of the Related Art Conventionally, various types of language recognition devices have been realized, such as a handwritten character recognition device and a speech recognition device. Some records the mark "No," and others require repeated input until they can be recognized.

In OCR, there is a system that stores an original image and compares the recognition result with the original image, thereby making it easy to correct an erroneously recognized character.

[0004]

However, in the first method of the prior art, data input when recognition was not possible is lost. Since data that could not be recognized by the recognition device can often be recognized by humans, this data should be utilized.

[0005] Further, in the second method of the prior art, the OC method is used.
It is necessary to store all image data read by R, which requires a large storage area.

Further, in the conventional method, it is difficult to judge whether a character is recognized as the closest to a character rather than repeatedly seeking an input. In some cases, it was forced to be recognized as a certain character.

An object of the present invention is to save input data itself when language input data cannot be recognized, so that when the user refers to the data later, what is input can be determined. Voice In
To provide a power word processor .

[0008]

SUMMARY OF THE INVENTION An object of the present invention is to provide a voice input means for inputting voice data, a voice recognition means for recognizing and processing voice data input from the voice input means, and a voice recognition apparatus. depending on the result of recognition means, when the speech recognition means in together when the case of recognizing the voice data the input are stored as character data, did not recognize the voice data to which the Symbol speech recognition means is the input means for storing the audio data said input, said Symbol speech recognition means
If the input voice data is not recognized, an unrecognition is displayed on the screen of the word processor, and the voice data stored in the storage unit is sounded by an instruction to the unrecognized display portion.
This is achieved by a voice input word processor having a voice output control means.

[0009]

The voice recognition means is input from the voice input means.
Recognize voice data.

[0010]

[0011]

[0012] The control means is configured so that the voice recognizing means does not recognize.
Unrecognized audio data is displayed on the word processor screen.
And an instruction to the unrecognized display portion causes the memory
The voice is output from the voice data stored in the row.

[0013]

FIG. 1 is a block diagram showing an embodiment of a language recognition apparatus related to the present invention.

The language recognition apparatus according to the present embodiment has an input means 1 for inputting data to be recognized, a recognition means 2 for recognizing a language, and recognized (coded) data. Alternatively, storage means 3 for storing unrecognized data (as input data), and a language or input data corresponding to the recognized data stored in storage means 3 being output as it is. Output means 4 and control means 5 for controlling the operation of the entire apparatus.
Storage means 3 together with the recognized and coded input data
The user can save the data later.
When referring to the data, it is possible to determine what has been input.

That is, input data input from the input means 1 is passed to the recognition means 2 by the control means 5.
Here, the input data is also passed to the output means 4 and output as required.

The recognition means 2 performs a recognition process based on the received input data. And if it is recognized,
The coded data after recognition is stored in the storage means 3. If the recognition is not possible, the input data itself which is not recognized is stored in the storage means 3.

At the time of output, according to the contents of the storage means 3, if the coded data after recognition is converted into a form at the time of inputting characters, voices, etc., it is outputted from the output means 4, but is output from the output means 4. In the case of data that could not be obtained, the input data itself is stored, so that it is output from the output means 4 as it is.

FIG. 2 is a block diagram showing a first modification of the embodiment of the language recognition apparatus.

A first modification of the embodiment of the language recognition apparatus is as follows: an input unit 11, a handwritten character recognition unit 12, and a storage unit 13.
And output means 14 and control means 15 for controlling the operation of the entire apparatus. The control means 15 is located at the center of the device,
A microprocessor or dedicated LSI is provided. The input means 11 is a tablet device provided with a light pen, a pressure-sensitive sheet, an electronic pen using electromagnetic induction or the like,
The handwriting of the written handwritten character is input to the control means 15.
The output unit 14 is configured by a display such as a CRT or a liquid crystal, and displays handwriting input from the input unit 11, characters stored in the storage unit 13, and unrecognized handwriting.

The handwritten character recognizing means 12 also includes a microprocessor or a dedicated LSI (which can be shared with the control means 15) similarly to the control means 15, and recognizes handwriting data input from the input means 11.

If the handwritten character recognition means 12 can recognize the input handwriting, the handwriting character recognition means 12 stores a flag indicating the recognition and the character code via the control means 15 into the storage means 1.
If the input handwriting cannot be recognized, a flag indicating that the handwriting could not be recognized and the input data are stored.
The data itself is stored in the storage means 13. That is, the data read from the storage means 13 is provided with a flag indicating whether the data is character code or input data itself.

FIG. 3 is a flowchart for explaining character input according to a first modification of FIG.

In FIG. 3, the handwritten character recognizing means 12 performs an input handwriting recognition process (3-1). If the character code is recognized, the flag indicating the character code and the character code of the recognized character are stored in the storage means 13 (3.
-2, 3-3). If the recognition is not possible, the storage means 13 stores a flag indicating that recognition was not possible (this can also be referred to as a flag indicating that the data is the input data itself) and the input data itself which could not be recognized. (3-2, 3-4).

Next, a process for outputting the information stored in this manner will be described.

At this time, the data is read from the storage means 13 and the value of the flag added to the data is checked. If it is a flag representing a character code, the data is regarded as a character code, and the character represented by the character code is output to the output means 14. If it is a flag indicating that it cannot be recognized, the data is regarded as input data itself (that is, handwriting at the time of input), and the input data itself is output to the output means 14. .

FIG. 4 is a flowchart for explaining the output of characters according to the first modification of FIG.

Data is read from the storage means 13 (4-
1) Check the value of the flag added to the data (4)
-2).

If the flag is a flag representing a character code, the read data is regarded as a character code, and a character corresponding to the character code is output to the output means 14 (4).
-3) On the other hand, if this flag is a flag representing the input data, the data is regarded as the input data itself (that is, the handwriting at the time of input) and is output to the output means 14 (4).
4).

Although there are various methods for adding a flag to data, a simple example will be described below.

(1) When handwriting at the time of input is stored as graphic data The first bit is set as a flag.
Then, it represents "input data". This flag is 0
In the case of, the character code itself follows. When this flag is 1, a numerical value representing the size of the graphic data representing the handwriting follows, followed by the graphic data itself representing the handwriting.

FIG. 5 is a diagram for explaining the data structure when the handwriting at the time of input is recognized, and FIG. 6 is a diagram for explaining the data structure when the handwriting at the time of input is not recognized. .

In FIG. 5 (when handwriting is recognized)
20 is a storage area for inputting the flag 0, 21 is a character code.
In the storage area to which the character itself is to be inputted, reference numeral 22 denotes graphic data representing the character "A", a flag 0 is stored in the storage area 23 of the storage means 13, and "A" is stored in the storage area 24.
Character code (0001) is stored.

In FIG. 6 (when handwriting cannot be recognized), 25 is a storage area to which the flag 1 is to be input, and 26 is the size of input data (the number of horizontal dots and the number of vertical dots). A storage area 27 to which the input data itself (bit image data) is to be input;
Numeral 8 denotes graphic data representing the character "A", a flag 1 is stored in the storage area 29 of the storage means 13, and a horizontal 1 which is the bit image size in this case is stored in the storage area 30.
Data representing 1 dot and 11 vertical dots, and 31 is a bit image data (000, 138, 31) consisting of 11 pieces of three-digit hexadecimal numbers obtained by dividing each row of the bit image by 4 dots. ... 450,000) is stored. In this example, the number of horizontal dots in the bit image is desirably a multiple of four. However, referring to the size data, the number of horizontal dots is eleven, and the number of horizontal dots is eleven. Can be determined to be dummy data, so there is no problem.

(2) When the handwriting at the time of input is stored as stroke data The first bit is set as a flag.
Represents "input data". When this flag is 1, a numerical value representing the size (stroke number) of the stroke data representing the handwriting follows, followed by the stroke data itself representing the handwriting.

FIG. 7 is a diagram for explaining another data structure when handwriting at the time of input is recognized. FIG. 8 is a diagram for explaining another data structure when handwriting at the time of input is not recognized. FIG.

In FIG. 7 (when handwriting is recognized)
35 is a storage area for inputting the flag 0, and 36 is a character code.
In the storage area to which the character itself is to be input, reference numeral 37 denotes graphic data representing the character "i". The storage area 38 of the storage means 13 has a flag 0, and the storage area 39 has "i".
Is stored.

In FIG. 8 (when handwriting cannot be recognized), reference numeral 40 denotes a storage area for inputting the flag 1, 41 denotes a storage area for inputting the number of strokes of the input data itself, and 42 denotes an input area. Of the data itself, a storage area to which vector data is to be input, 43 indicates graphic data representing the character "i", and a storage area 4 of the storage means 13
4 is a flag 1 and a storage area 45 is 2 strokes.
However, in the storage area 46, the stroke data ((2, 2,
2) (-1, 6, 6) (1, 1, 1) (9, 2, 1) (0,
1, 5)) are stored. In this case, the stroke data has the x-coordinate of the starting point, the y-coordinate of the starting point, and the number of continuous straight lines, followed by the number of continuous straight lines (x component of direction, y component of direction, length). It is expressed as something.

FIG. 9 is a block diagram showing a second modification of the embodiment of the language recognition apparatus.

In FIG. 9, reference numeral 51 denotes input means, 52 denotes voice recognition means, 53 denotes storage means, 54 denotes output means, and 55 denotes control means for controlling the operation of the entire apparatus.

A control means 55 is arranged at the center of the apparatus, and includes a microprocessor or a dedicated LSI. The input unit 51 is a voice input device provided with a microphone or the like, and inputs a pronounced voice to the control unit 55. The output means 54 is composed of a voice output device provided with a voice synthesizer, a speaker, and the like.

The voice recognition means 52 is similar to the control means 55,
A microprocessor or a dedicated LSI is provided (which can be shared with the control unit 55), and recognizes voice input from the input unit 51.

If the voice recognition means 52 can recognize the input voice, it stores a flag indicating the recognition and the character code in the storage means 53 via the control means 55. If the voice cannot be recognized, the flag indicating that the voice cannot be recognized and the input data itself are stored in the storage means 53. That is, the storage unit 5
A flag is added to the data read from No. 3 to indicate whether the data is character code or input data itself.

[0043] Figure 10 is a word according to a second modification shown in FIG. 9
It is a flowchart explaining input of a word .

In FIG. 10, the voice recognition means 52 performs a recognition process on the input voice (10-1).

Here, if the character code is recognized, a flag representing the character code and the character code are stored in the control means 55.
Through the storage means 53 (10-2, 10-
3).

If recognition is not possible, a flag indicating that recognition was not possible (this can also be referred to as a flag indicating that the data is input data itself) and the input data itself that could not be recognized Is stored in the storage means 53 (10-2, 10-4).

Next, a process for outputting the information stored as described above will be described.

First, data is read from the storage means 53 and the value of the flag added to the data is checked.

If the flag is a flag representing a character code, the data is regarded as a character code, and the voice of the character represented by the character code is output to the output means 54.

If the flag is a flag indicating that it could not be recognized, the data is regarded as input data itself (that is, voice at the time of input), and the input data itself is output. Output to

FIG. 11 is a flowchart for explaining the output of the language according to the second modification shown in FIG.

Data is read from the storage means 53 (11)
-1) Check the value of the flag added to the data (11-2). If this is a flag indicating a character code, the read data is regarded as a character code, and a sound corresponding to the character code is output to the output means 54 (11-).
3) If this flag is a flag representing input data, the data is regarded as the input data itself (that is, voice at the time of input) and output to the output means 54 (11-).
4).

There are various methods for adding a flag to data. A simple example will be described below.

The first bit is used as a flag, and if 0, it indicates "character code", and if 1, it indicates "input data". When this flag is 0, the character code itself follows.
When this flag is 1, a numerical value representing the magnitude of the PCM data representing the voice follows, followed by the PCM data itself representing the voice.

FIG. 12 is a diagram for explaining the data structure when the voice at the time of input can be recognized, and FIG. 13 is a diagram for explaining the data structure when the voice at the time of input cannot be recognized. is there.

In FIG. 12 (when speech can be recognized), reference numeral 60 denotes a storage area for inputting the flag 0, 61 denotes a storage area for inputting the character code itself, and 62 denotes a waveform of the voice "U". And the storage area 6 of the storage means 53
3 is a flag 0, and the storage area 64 is a character code "U".
(0003) is stored.

In FIG. 13 (when speech cannot be recognized), 65 is a storage area for inputting the flag 1, and 66 is a storage area for inputting the size of input data (the number of bytes of input data). , 67 are the input data itself (PCM data).
And 68, a waveform of the input voice "U", and a storage area 68 of the storage means 53.
9, the flag 1 is stored in the storage area 70, and the PCM data (853) is stored in the storage area 71.
0, 25,... 30, 96) are stored.

In the second modified example, an apparatus which can recognize and store a voice and output the voice later is described as an example, but this can be applied to a voice input word processor or the like. The present invention
When the voice cannot be recognized, a symbol indicating the fact is displayed on a screen of a word processor, and when the portion is pointed by a cursor key, an electronic pen, or the like, the voice input word processor outputs the voice data itself that cannot be recognized.

In the present embodiment, even if input data that the recognition device cannot recognize is stored, the input data itself is stored, so that what is input when the user later refers to the input data. It is possible to perform a recognition process that can determine whether or not it has been performed.

[0060]

According to the present invention, as described above, a speech input word processor is used.
, The voice data that the voice recognition means could not recognize
However, since the user can often recognize it, speech recognition
If the means cannot recognize the input voice,
-Unrecognized display on the professional screen and
Unrecognized audio data from the storage means according to the instruction
The voice data input when performing voice recognition
That can achieve efficient voice input
You.

[Brief description of the drawings]

FIG. 1 is a block diagram showing an embodiment of a language recognition device related to the present invention.

FIG. 2 is a block diagram showing a first modification of the embodiment of the language recognition device.

FIG. 3 is a flowchart for explaining character input according to a first modification of FIG. 2;

FIG. 4 is a flowchart for explaining character output according to a first modification of FIG. 2;

FIG. 5 is a diagram illustrating a data structure when handwriting at the time of input is recognized.

FIG. 6 is a view for explaining a data structure when handwriting at the time of input cannot be recognized.

FIG. 7 is a diagram illustrating another data structure when handwriting at the time of input is recognized.

FIG. 8 is a diagram for explaining another data structure when handwriting at the time of input cannot be recognized.

FIG. 9 is a block diagram showing a second modification of the embodiment of the language recognition device.

FIG. 10 is a flowchart illustrating input of a language according to a second modification of FIG. 9;

FIG. 11 is a flowchart illustrating language output according to a second modification of FIG. 9;

FIG. 12 is a diagram for explaining a data structure in a case where speech at the time of input has been recognized.

FIG. 13 is a diagram showing data when speech at the time of input cannot be recognized.
FIG. 3 is a diagram illustrating a data structure.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 1 Input means 2 Recognition means 3 Storage means 4 Output means 5 Control means

Continuation of the front page (58) Fields investigated (Int. Cl. ⁷ , DB name) G10L 15/22 G06F 17/22 503 INSPEC (DIALOG) JICST file (JOIS) WPI (DIALOG) IBM Intellectual Property Network

Claims (1)

(57) [Claims] A voice input unit for inputting voice data; a voice recognition unit for recognizing and processing voice data input from the voice input unit; together when recognizing the voice data the input is stored as character data, and means if the Symbol voice recognition unit does not recognize the voice data the input is for storing audio data said input, said The voice recognition means recognizes the input voice data.
If not, a control means for displaying unrecognition on a screen of a word processor and outputting voice data stored in the storage means in response to an instruction to the unrecognized display portion. Input word processor.