JPS59206938A - Voice word processor - Google Patents

Abstract

PURPOSE:To simplify the operation of a voice word processor, by automatically returning a cursor which has advanced to the expected position of the next character, when characters other than a desired one are displayed and a correction key is depressed during the period when the word processor is operated by means of voice inputs. CONSTITUTION:When a voice is inputted from a microphone 18 after the kind of character is appointed through a keyboard 14, the voice is sent to a CPU12 after it is recognized 11 and encoded. The CPU12 converts plural recognized results into characters and displays 13 the characters. An operator can appoint the kind of character (alphabetical and numerical characters) by means of the auxiliary key of the keyboard 14 before inputting the voice, but usually ''Kana'' (Japanese syllabary) character is displayed. When a ''Kanji'' (Chinese character) conversion key is depressed, a dictionary 15 is accessed and homonymous characters are successively displayed. The operator depresses a correction key until a desired character appears. At this time, a cursor has advanced to the expected position of the next character irrespectively of the intention of the operator, but, when the correction key is depressed, the cursor is automatically returned to the original position. Therefore, the operating of the operator is lightened and the operating property is improved.

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention relates to a voice word processor for inputting desired characters by voice.

In this type of voice word processor, an operator's utterances are recognized by a voice recognition device, and characters corresponding to a plurality of recognition results are generated by a processing device. The processing device first displays a first candidate character based on the first recognition result obtained from the speech recognition device. If this display is not the character intended by the operator, the operator presses the correction key.

This causes the processing device to display the next candidate character obtained by converting the second recognition result.

At this time, the cursor on the display screen is located at the position of the next character to be input, and the character located immediately before the cursor is changed to the next candidate character as the correction key is pressed. When the operator continues to press the correction key to display up to the nth candidate character, the first candidate character is displayed again.

In a conventional speech word processor, 1 to 1 converted by a processing device according to n results from a speech recognition device.
If the operator's intended result does not exist among the candidate characters, the operator's utterance is assumed to have been misrecognized by the voice recognition device, and the operator presses the key that moves the cursor back on the display screen, then speaks again and re-enters the input. was going on. Therefore, the operator's re-input procedure is complicated, and the character input speed is accordingly slow.

SUMMARY OF THE INVENTION Accordingly, an object of the present invention is to reduce the number of operations performed by an operator when re-entering characters and to increase the speed of character input.

According to the present invention, when information indicating correction is input after displaying the characters corresponding to the last recognition result from the speech recognition device, the cursor that had been moved to the next display position is automatically moved to the current display position. A voice word processor is obtained which is characterized in that it returns the voice and instructs re-voice input.

Next, the present invention will be described in detail with reference to the drawings showing one embodiment of the present invention.

In Figure 1, the voice uttered by the operator is transmitted through microphone 1.
8 is input to the speech recognition device 11. Thereby, the first to nth recognition results of the recognition device 11 are encoded and sent to the processing device 12.

The processing device 12 converts the plurality of recognition results into corresponding characters and displays them on the display device 13 starting from the first candidate character. The type of displayed character for each character (kana, alphabetic character, etc.) is specified by the operator using a group of input auxiliary keys on the keyboard 14 before voice input, but normally the mode is set to kana mode. If the operator wishes to convert the input kana into kanji, the operator presses the kanji conversion key on the keyboard 14, and the processing device 12 accesses the kanji dictionary 15 and converts the input kana to the displayed kana. Kanji characters are selected and displayed in order on the display device 13 according to instructions from an operator. The finally input character string is stored in the document storage device 16, and when printing it on paper, the operator presses the print key on the keyboard 14 to output it to the printer 17 and print it. .

Next, the operation of this embodiment will be explained with reference to the flowcharts of FIGS. 2 and 3. In this embodiment, a case where [NECJ is input] is shown.

First, the operator presses the capital letter key on the keyboard 14 and speaks "N" into the microphone 1B. When this is correctly recognized by the recognition device 11 and input to the processing device 12, the processing device 12 outputs rNJ to the display device 13 as the first candidate character. At this time, the cursor on the display screen is positioned next to rNJ. Since this is the character intended by the operator, the operator either presses the "No conversion" key on the keyboard 14 to send a confirmation signal to the processing device 12, or
Alternatively, this operation may be omitted and the next input character "E" may be input into the recognition device 11 via the microphone 18. If this is also correctly recognized by the recognition device 11, the processing device 12 causes the display device 13 to display “E”. This state is shown in Figure 2 (
Shown in a).

If the recognition device 11 is unable to correctly recognize the next operator's utterance “shi”, the processing device 12 selects the first candidate character based on, for example, the first to third recognition results sent from the recognition device 11. rEJ is displayed as (Fig. 2b). The cursor position is advanced to the next character position. When the operator discovers that the displayed character is not what he intended and presses the correction key on the keyboard 14, the processing device 12 displays the second candidate character "S" on the screen instead of "E". (Figure 2C). Since this character is also not the character intended by the operator, the operator presses the correction key on the keyboard 14 again.

As a result, the processing device 12 causes the third candidate character rUJ to be displayed on the screen instead of "S" (FIG. 2d). In this case, it is assumed that the recognition device 11 has sent the first to third recognition results to the processing device 12. When the operator presses the correction key indicating that the character rUJ is an unintended character, the processing unit 12 automatically returns the cursor to the position where rUJ was displayed (FIG. 2e). As a result, the operator knows that rCJ's voice has been misrecognized, and without pressing the back key of the cursor, the operator presses the microphone 1.
The re-input of "C" by voice can be performed via 6-8. If the crack is correctly recognized by the recognition device 11, the processing device 12 that receives this information displays "C" on the screen 13 (FIG. 2 f), so that the operator can finish the series of input operations. can.

In addition, in the state shown in FIG. 2e, if the operator presses the correction key again without performing the re-input operation, the processing device 1
2 is the first candidate character again [display is repeated in order from J.

In the present embodiment, the conversion to the next candidate was performed by pressing the correction key on the keyboard 14, but "A", "I"
Monosyllabic sounds such as ``u'' have little audio information and are easily misrecognized, but words such as ``correction'' have a large amount of information and are less likely to be misrecognized.Use this fact to use the correction key. Instead of pressing the button, it is also possible to make corrections by inputting the voice command ``Teisei''.

As described above, the present invention has the effect of reducing the number of operations performed by the operator by moving the cursor back when correcting the last candidate character selected by the processing device and allowing re-input by voice.

[Brief explanation of drawings]

FIG. 1 is a schematic diagram of a device configuration showing an embodiment of the present invention, FIG. 2 is a diagram showing display contents of a display device, and FIG. 3 is a flowchart showing the operation of an embodiment of the present invention.夷／ 圀・・）NE口（b）NE E■ （・）NF２口・・・NE■ Bear 2 Leap

Claims (1)

[Claims] a speech recognition device that outputs first to nth recognition results for input speech; and a speech recognition device that converts the first recognition result into a character code and displays the corresponding character on a display device and at the next display position. an audio word processor that moves a cursor and, when information indicating a correction to a displayed character is input, displays the character sequentially at the current display position of the display device starting with the character corresponding to the second recognition result; wherein, when the information indicating the correction is input after the processing device displays the character corresponding to the n-th recognition result, the processing device returns the cursor to the current display position and instructs re-input of voice. A voice word processor.