JPH0570868B2 - - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION (A) Field of Industrial Application The present invention relates to a voice input text creation method for displaying sentences created using voice input.

(b) Conventional technology Conventional word processors create sentences by converting characters entered using a keyboard, etc. into kana and kanji at the appropriate time, but there is a method for conveying information that is more natural for humans. A voice input word processor and the like as disclosed in Japanese Patent Laid-open No. 114042/1983 have been developed for the purpose of creating sentences by voice input as a medium. However, when displaying the input sentence, for example, "Kiya"
If you enter the word by voice, it will be divided into [ki] and [ya] and displayed continuously on the display device, and the cursor used to indicate the input position is only the size of one character, so it will be displayed continuously on the display device. Therefore, it was practically difficult to treat "Kiya" inputted by voice as a single phonetic character. In other words, if you try to delete the above-mentioned display of "Kiya", for example, move the cursor to the position of "Ki", delete the character "Ki", and then press [Kiya] ] had the disadvantage of requiring a complicated operation of deleting the characters.

(C) Problems to be Solved by the Invention The present invention has been made in view of the above-mentioned circumstances, and allows editing processing such as deletion of a set of characters displayed in response to voice input to be performed in one step. The purpose is to make it easy to operate.

(d) Means for solving the problem The speech input text creation method of the present invention targets speech expressed by a written character string of two or more characters, and recognizes the input speech with a speech recognition means. In the voice input sentence creation method of creating a sentence while displaying the recognition result on the display means, the display means 2
When displaying the recognition result of a voice represented by a written string of characters or more, a cursor for specifying the recognition result of the voice displays the entire written string. In other words, the display range of the cursor is variably set in accordance with the number of characters indicating the phonetic unit.

(e) Effect: According to the voice input text creation method of the present invention, a series of written character strings of voice units specified by the cursor on the display means can be simultaneously changed,
It performs editing processing such as deletion and deletion.

(F) Embodiment FIG. 1 shows an embodiment for realizing the speech input sentence creation method of the present invention using a monosyllabic speech recognition system. Reference numeral 1 denotes a microphone, which converts the uttered voice into an electrical signal. 2 is the feature extraction part,
The input speech waveform is frequency-analyzed by, for example, emphasizing the high frequencies and passing through a 16-channel bandpass filter, and segmentation is performed using power information, spectral changes, etc. to create characteristic patterns for each single syllable. . 3 is a standard pattern memory in which characteristic patterns extracted for each single syllable are registered as standard patterns. Reference numeral 4 denotes an identification section that compares the characteristic pattern created in the feature extraction section 2 with the standard pattern of each monosyllable in the standard pattern memory 3, and outputs character information according to the result. 5 is an identification data memory section, and 6 is a data pointer, which stores the character information output from the identification section 4 at the position pointed to by the data pointer 6 in the identification data memory section 5. 7
is an operation section for changing the pointer value of the data pointer 6 and operating character information in the identification data memory section 5 pointed to by the data pointer 6. A display section 8 displays characters based on the character information in the identification data memory section 5 and displays a cursor to clearly indicate the displayed character based on the character information pointed to by the data pointer 6.

FIG. 2 is a block diagram showing the display unit 8 in detail. According to the figure, character information sent from the identification data memory unit 5 is processed by the operation of the character generator and video ram of the screen display unit 81. , the corresponding characters are displayed on the CRT82. On the other hand, based on the character information in the identification data memory section 5 pointed to by the data pointer 6, the cursor control section 83 determines the cursor display range and its display position, and the screen display section 81 A cursor is displayed on the CRT 82 based on the display range and display position information, that is, cursor information.

Figure 3 schematically shows the characters and cursor displayed on the CRT 82. In the figure, a indicates the words "to", "ku", "chiyo", and "u" recognized in monosyllable units.
This is an example in which the ``ku'' part of the character ``ha'' is displayed with the cursor 9. From this state, if the cursor 9 is moved to the right by changing the value of the data pointer 6 using the operation section 7, the result will be as shown in b in the figure. That is, the cursor control section 83
After moving the cursor 9 to the right, it detects that the recognition unit "chiyo" to be pointed to consists of two characters "chi" and "yo", and the display range of the cursor 9 is two characters. cursor information indicating the new position of the cursor as well as the new position of the cursor is displayed on the screen display section 8.
1, and the cursor 9 displayed on the CRT 82 points to the entire "Chiyo". FIG. 3b shows an example in which "Chiyo" pointed to by the cursor is deleted using the operation section 7 in the state a of the same figure [the same state as b of FIG. 3]. The cursor control unit 83 detects that the recognition unit "U" to be newly pointed consists of one character, and informs that the display range of the cursor 9 is one character, as well as cursor information indicating the new position of the cursor. is sent to the screen display section 81. In this state, the operation section 7 deletes the data corresponding to "Chiyo" pointed to by the data pointer 6 in the identification data memory section 5, and displays the information on the screen display section 8.
1, and the results are displayed on the CRT82. Further, editing processes such as changes and deletion other than the above-mentioned deletion can be executed in the same manner.

Note that although the above explanation has been given for the case of recognition in units of monosyllables, the present invention can also be implemented in recognition of units of words. Further, although the cursor is illustrated and explained using an underline method, the same applies to a method using a luminance inversion method.

(G) Effects of the Invention As explained above, when the pronunciation of a monosyllabic sound such as a sound, a half sound, a long sound, a long sound, etc. is displayed using two or more characters, the displayed characters are By handling monosyllable units, that is, pronunciation units, rather than units, and by handling display characters in pronunciation units even in the case of word recognition consisting of multiple characters, it is possible to edit text input by voice extremely efficiently. It can be carried out.

[Brief explanation of the drawing]

Fig. 1 is a block diagram of an embodiment of the voice input text creation method of the present invention, Fig. 2 is a block diagram of an embodiment of the display unit used in the method of the present invention, and Fig. 3 and 3 illustrate display examples of the method of the present invention, respectively. FIG. 1...Microphone, 4...Identification section, 5...Identification data memory, 6...Data pointer, 7...Operation section, 8...Display section, 9...Cursor.

Claims (1)

[Claims] 1. A speech input text creation method that targets speech expressed by a written character string of two or more characters, recognizes the input speech with a speech recognition means, and creates a text while displaying the recognition result on a display means. When displaying the recognition result of a voice represented by a written string of two or more characters on the display means, the cursor for specifying the recognition result of the voice will display the entire written string, and the cursor will At the same time, the written character string of the identified phonetic unit may be changed, deleted, or
A voice input text creation method characterized by performing editing processing such as deletion.