JPH0664571B2 - Character processing method - Google Patents

Description

Description: TECHNICAL FIELD The present invention relates to a character processing method capable of extremely efficiently performing voice input processing in a character processing device such as an electronic typewriter and a word processor.

[Prior Art] Conventionally, there is known a character processing device such as a Japanese word processor which is provided with an input system based on an operator's voice in addition to a character input system from a keyboard device.

The advantage of the voice input method is that you can input a document instead of the keyboard input operation that requires skill,
Enter the punctuation at the end of the phrase using the keyboard or
Say "ten" or "maru" by voice, and then ","
Another input operation of the character string that you originally wanted to input, such as converting to or ".", Was required.

In addition, a case particle or the like, which is written as “ha” but is written as “wa” as a voice, needs to be re-input by the keyboard after the voice recognition.

[Problems to be Solved by the Invention] In a character processing method to which a conventional voice input method is applied, for example, a punctuation mark is added to the end of a phrase, and a pronunciation corresponding to a case particle (for example, "wa") corresponds to the case particle It has been difficult to perform operations such as converting to characters (for example, "ha") that are performed using only voice input.

[Object] To solve the above problems, an object of the present invention is to recognize a break in a character string from an input voice and determine the type of the character string before the break in the recognized character string. , By converting the code information of the recognition result according to its type, it is possible to obtain more information suitable for the notation of the document than the information obtained by recognizing the input voice. It is to provide a processing method.

EXAMPLES The present invention will be described in detail below with reference to the drawings.

FIG. 1 shows an example of the configuration of the character processing device of the present invention. Here, 1 is, for example, a microprocessor type processing unit (MPU), 2 is an address bus for specifying the address of each input / output memory, 3 is a bidirectional data bus used for transferring various data, and 4 is This is a control bus used to read, write, accept interrupts from each memory, and transmit data set timing. MPU1 has a built-in timer.

Reference numeral 5 is a keyboard provided with various keys such as an alphabet key, a hiragana key, a numeric key, a recall key, an input, edit, save, and print key. Reference numeral 6 denotes a keyboard controller for controlling the keyboard 5.
The input data is encoded and the interrupt signal is sent to the control bus 4. In addition,
The decoded key data is temporarily stored in the key buffer 6A, encoded by the controller 6 according to the access of the MPU1, and read into the MPU1.

Reference numeral 7 denotes a CRT controller, which controls the CRT display device 8, the display controller 9, the character generator 10, the refresh memory 11, and the refresh memory controller 12. The display controller 9 processes the pattern signal from the character generator 10 according to display control signals such as brightness, inversion, underline, and blink stored in the refresh memory 11, and outputs the processed output to the CRT display device 8. Supply and output the display pattern.

The character generator 10 sequentially outputs a character pattern for a line designated by a code input such as a character code and a line address of the character pattern among the stored character patterns.

The refresh memory 11 stores a character code and a display control code to be displayed on the CRT display device 8,
In response to a display instruction from the CRT controller 7, the character code and the display control code are repeatedly output. The refresh memory controller 12 is the CRT controller 7
In response to an instruction from, the above-mentioned refresh cycle for CRT display and the writing and reading of data from MPU1 are controlled.

A read-only memory (ROM) 13 stores a control procedure described later and various processing procedures such as character processing. 14 is a controller for controlling the ROM 13.

Reference numeral 31 is a microphone for the operator to input a voice, and the voice recognition device 30 performs processing such as frequency analysis, feature extraction, and pattern matching on the voice input through the microphone 31 to recognize a single syllable. Then, the voice signal is converted into a character code and taken into the main body of the apparatus. In the frequency analysis, the voice recognition device 30 sets an accent flag on the output character code when the frequency band of a certain single syllable is higher than the frequency bands of a plurality of consecutive single syllables input so far. Output.

15A is a random access memory (RAM), which temporarily stores various data and develops a dictionary for Kana-Kanji conversion in a predetermined area. Further, 15B is a RAM for storing the created document. Reference numeral 15C is an RA having a key buffer 6A, a storage area (conversion buffer) for data sent from the voice recognition unit 30, a storage area for routines for grammar processing, editing processing, and tabulation processing described later and a work area.
M and 16 are output buffers for temporarily storing the outputs of RAM15A, RAM15B and RAM15C. 17 is RAM15A, RAM15B and RAM15
It is a C controller.

18 is a floppy disk for storing dictionary information, grammar processing routines, created documents, etc. provided as necessary, 19 is a floppy disk drive, 20 is a controller of the floppy disk drive 19, 21 is a printer, 22 is a controller of the printer 21.

The character processing apparatus of the present invention having the above-described configuration operates by input from the keyboard 5, an OCR (optical reading means) (not shown) or the microphone 31. That is, when an input is made from the keyboard 5 or the microphone 31, an interrupt signal is supplied from the keyboard controller 6 or the voice recognition unit 30 to the MPU 1 via the control bus 4, and the ROM 1
Performs various control operations required for character processing in accordance with the control procedure stored in 3. Although the present invention is characterized in that character processing is performed by voice input as will be described later, of course work that cannot be processed by voice input is performed by input work from the keyboard 5, for example.

FIG. 2 shows an example of the key arrangement of the keyboard 5. Here, 5A is a group of input keys for letters and numbers, and 5B is a group of keys for selecting a processing mode for a document. For example, when a "print" key is pressed, a printing process using the printer 21 is performed. To be done. Further, when the "voice input" key is pressed, the input mode switch 32 is closed (turned on) so that the output of the voice recognition device 30 is connected to the data bus. On / off of the input mode switch 32 is controlled by the MPU 1. 5C is a space key, and 5d is a cursor movement key group on the CRT screen.

FIG. 3 shows an example of a voice input operation of the character processing device of the present invention. As shown in FIG. 3, in step S1, it is determined whether or not there is a voice input. For example, pronunciation "wa""ta"
When "shi""wa" is entered, the kana character code string (Iwa) corresponding to the input is output from the voice recognition device 30,
Proceed to step S2. Similarly, for example, pronunciation “wa” “na”
In the case of "o", the character code string (still) is output, and the process proceeds to step S2.

In step S2, the character code from the voice recognition device 30
It is input to the conversion buffer area in RAM15C. This buffer area has a capacity of, for example, 8 characters. Then step
Go to S3 and start the timer (retrigavable type) in MPU1. Next, in step S4, it is determined whether or not the character code is input to all the addresses in the conversion buffer. If character codes have been entered in all of the buffer, proceed to step S5. If not, return to step S1. In the above example, since character codes have not been input to all addresses in the conversion buffer, the process returns to step S1. If there is no next input, proceed to step S6,
Determine if the timer has timed out. The time from the start of the timer to the time-out is, for example, 1 to 2
It can be on the order of seconds. If it has not timed out in step S6, it returns to step S1, and if it has timed out, it proceeds to step S7.

In step S7, a timeout code follows the last character code in the conversion buffer. Enter, and proceed to step S5.

In step S5, the character code string in the conversion buffer is extracted, and the kana-kanji conversion is performed by referring to the dictionary in the RAM 15A. For example, a character code string (Iwa) is converted into a kanji / Kana mixed sentence (Iwa), and similarly, a character code string (Hana) is converted into a Kanji / Kana mixed sentence (Hanao).

Next, in step S8, it is determined whether the kanji / kana mixed sentence obtained in step S5 has an accent flag. Step if accent flag is attached
Proceed to S9, and if not, proceed to step S10. For example, continuous pronunciation of a single syllable means “wa” “ta” “shi” “de”
If you type in "su""ka" and then there is no voice input until the timer times out (timeout code added), and if you add a higher accent to the last "ka" pronunciation (such as In this case, the pronunciation of "ka" is usually interrupted), and in step S5 a kanji / kana mixed sentence (is it me) is obtained, and then step
Proceed from S8 to step S9, where the timeout code after the kanji-kana mixed sentence (I am) Is deleted and the mark “?” Is added after the kana-kana mixed sentence (that is, for example, (Is it?)), And the process proceeds to step S10.

In step S10, it is determined whether the end of the kanji and kana mixture is "wa,""o," or "e."
(Yes) proceeds to step S11, and (No) proceeds to step S12. For example, a kanji-kana mixed sentence (I),
In the case of (Hanao), proceed to step S11, and in the case of (Is it?) Proceed to step S12.

In step S11, the time-out code is after the kanji-kana mixed sentence. Alternatively, it is determined whether the mark is "?", And if it is applicable, the process proceeds to step S13, and if not, the process proceeds to step S14.

For example, kanji and kana mixed sentences (I) + timeout code In the case of No, since it corresponds, proceed to step S13.

In step S13, the last characters "O", "Wa", and "E" of the kanji-kana mixed sentence are the characters "o",
Convert to "ha" and "to" and proceed to step S14.
For example, a kanji-kana mixed sentence (I) is converted into (I), and a (Hanao) is converted into (Hana).

In step S12, the time-out code follows the kana-kana mixed sentence. If yes, proceed to (Yes) step S15,
Otherwise (No), proceed to Step S14. For example, in the case of the above (Is it?), No, so step S14
I will proceed. Also, for example, high voice input (pronunciation) is "Yu"
In the case of "or""to", if the input is interrupted until it times out after that, the timeout code is added to the character code. Is added (step S7), kana-kanji conversion is performed (step S5), and kanji-kana mixed sentences are added. Then, after step S10, proceed to step S12, where the kanji and kana mixture Is determined to proceed to step S15.

In step S15, the timeout code It is judged whether the sentence before is a noun, a case particle, a connecting particle or an adverb, and if so (Yes), it proceeds to step S16, and if not, it proceeds to (No) step S17. For example, kanji and kana mixed sentences In the case of, since (ni) is a case particle, it is Yes and the process proceeds to step S16.

In step S16, the time-out code after the kanji-kana mixed sentence Is converted into a reading mark "," and the process proceeds to step S14. For example, kanji and kana mixed sentences In the case of, a kanji-kana mixed sentence (on the floor) is obtained.

In step S17, the time-out code after the kanji-kana mixed sentence Is converted to a punctuation mark ".", And the process proceeds to step S14. For example, when the voice input (pronunciation) is "O" or "Ku",
After that, if there is no input until the timer times out, the timeout code is added to the character code. Is added (step S7, kana-kanji conversion is performed (step S5), kanji-kana mixed sentence Is obtained, and then proceeds to step S15 through step S10, where It is determined that the kanji-kana mixed sentence in front of is not a noun, case particle, connecting particle or sub particle, and the procedure proceeds to step S17.
In step S17, Is judged to be the final form, Is converted to a punctuation mark ".", And a kanji-kana mixed sentence (put.)
Is obtained.

In step S14, the kanji-kana mixed sentence obtained as described above is output to the output buffer 16, and output processing such as display on the CRT is performed, and the process returns to step S1. Through the above work, for example, "wa" and "ta" as voice input (pronunciation)
"Shi""wa" / timeout of timer / "wa"
"Na""o" / time out of timer / "yu"
"A" or "to" / Timer time out / "O"
You can get a kanji-kana mixed sentence (I put flowers on the floor) just by typing "ku" (see Fig. 5). In addition, pronunciation “wa” “ta” “shi” “de” “su” “ka” (however,
If you input "ka" with an accent), you will get a kanji-kana mixed sentence (I am?) (See Figure 6).

FIG. 4 shows another example of the operation of the character processing device of the present invention.
As shown in FIG. 4, in step S41, it is determined whether the input mode switch is on. If it is on, the process proceeds to step S42, and if it is off, the process returns to the original state. In step S2, the timer built into MPU1 is started, and in step S43
I will proceed. In step S43, the voice input process described in FIG. 3 is performed (steps S1 to S1).
4), and then proceed to step S44. In step S44, it is determined whether or not there is a voice input. If yes, the process proceeds to step S45 to reset the timer, and the process returns to step S42. If there is no voice input, the process proceeds to step S46, and it is determined whether the elapsed time t ₁ from the timer start is longer than a predetermined time T ₁ (for example, 30 seconds to 1 minute). If it is larger, the process proceeds to step S47 and the input mode switch is turned off. If not larger, the process returns to step S44.

As described above, the voice input mode can be automatically turned off when the voice input has not been performed for a certain period of time or more, and therefore the possibility of erroneous input due to microphone ambient noise or the like can be effectively avoided.

Note that, in the above-described automatic off operation of the input mode, for example, it is determined whether the voice input information is the end type, and if so, the timer is started, and the voice input information of the next end type must be input for a certain period of time. For example, the input mode switch can be turned off.

[Effect] As described above, according to the present invention, the break of a character string is recognized from the input voice, the type of the character string before the break of the recognized character string is determined, and the type is determined. Depending on the character processing method that converts the code information of the recognition result according to
There is an effect that information more suitable than the information obtained by recognizing the input voice can be obtained for the notation of the document, and the voice input can be performed extremely efficiently.

[Brief description of drawings]

FIG. 1 is a diagram showing an example of the configuration of a character processing device of the present invention, FIG. 2 is a diagram showing an arrangement configuration of a keyboard, FIG. 3 is a diagram showing an example of a voice input operation of the character processing device of the present invention, and FIG. The figure shows another example of the voice input operation of the character processing device of the present invention, and FIGS. 5 and 6 are diagrams showing the voice input work mode according to the present invention. 1 ... MPU, 13 ... ROM, 15A, 15B, 15C ... RAM, 30 ... Speech recognition device.

Claims (4)

[Claims] 1. A character string and a break in a character string are recognized from input voice information, and the recognized character string and code information representing the break in the character string are stored, and the break in the stored character string is stored. A character processing method, comprising: determining a type of a character string represented by code information stored before the represented code information, and converting the stored code information according to the determination. 2. When the type of the character string is determined to be a predetermined character string, the character string is converted into code information representing another predetermined character string corresponding thereto. The character processing method according to claim 1. 3. The character processing method according to claim 1, wherein the code information representing a break in the character string is converted into code information representing a punctuation mark according to the type of the character string. . 4. The character processing method according to claim 1, wherein code information representing a break in the character string is converted into code information representing a reading point according to the type of the character string. .