JPS62222359A - Voice input word processor - Google Patents

Abstract

PURPOSE:To increase the speed of the inscription of words by inscribing recognition results of a word and not performing Kana (Japanese syllabary)-Kanji (Chinese character) conversion. CONSTITUTION:The voice inputted from a microphone 1 is recognized as a word or a single syllable by a word recognizing part 2 and a single syllable recognizing part 3, and the single syllable or the word is discriminated by a discriminating part 4. A word number code of the word or a single syllable code of the single syllable is sent to a word table 7 and a Kana-Kanji converting part 8 through a word recognition result 5 or a single syllable recognition result 6 respectively. In case of the word, a word retrieved from the word table 7 by the word number is sent to a control processing part 9 and is outputted to a character string output part 10 as it is. By this processing, the conversion processing in the Kana-Kanji converting part 8 is unnecessary for the word and only retrieval of the word table 7 is performed, and the processing is performed more quickly.

Description

DETAILED DESCRIPTION OF THE INVENTION TECHNICAL FIELD The present invention relates to voice input word processors, and more particularly to
This invention relates to a speech input word processor that recognizes the speech of a sentence in units of words and monosyllables and exchanges it into character strings.

A method has already been developed in which the speech of a sentence uttered in words and monosyllable units is recognized, the results are converted into a kana character string, and then the kana and kanji are converted into a character string. Therefore, in this method, even though it is known in advance what characters the word recognition results correspond to, homophone processing is possible by converting the word back to a kana character string and then converting it again using kana-kanji conversion. Therefore, it takes time to convert, which is a bottleneck when using it.

Purpose The present invention was made in view of the above-mentioned circumstances.
In particular, this was done with the aim of speeding up the notation of words in voice input word processors that input speech uttered in units of words and monosyllables and convert them into character strings.

Composition In order to achieve the above-mentioned object, the present invention includes a word recognition unit that recognizes voice input in tongo and a monosyllable recognition unit that recognizes voice input in units of monosyllables, and a kana character string that converts the result into a character string. This voice input word processor includes a kanji conversion section and is characterized in that the word recognition results are expressed as they are, and no kana-to-kanji conversion is performed. Hereinafter, the present invention will be explained based on examples.

FIG. 1 is an electrical block diagram for explaining one embodiment of the present invention, and in the figure, 1 is a microphone.

2 is a word recognition unit, 3 is a monosyllable recognition unit, 4 is a word and monosyllable discrimination unit, 5 is a word recognition result, 6 is a monosyllable recognition result, 7
is a word table, 8 is a kana-kanji conversion unit, 9 is a control processing unit, and 10 is a character string output. /masu/, these input voices are recognized as words or monosyllables by the word recognition unit 2 and monosyllable recognition unit 3, and the discrimination unit 4 recognizes /wa/no/ based on the respective recognition results.
It is determined that shi/ is a single syllable and the others are words.

The word receives its word number code through the word recognition result 5.
For monosyllables, monosyllable codes are sent to the word table 7 and kana-kanji converter 8 via the monosyllable recognition result 6, respectively. In the case of a word, the word table 7 is searched by the word number, and the corresponding word is directly sent to the control processing section 9 and outputted as it is to the character string output section 10. In the case of a single syllable, the kana-kanji conversion unit 8 performs conversion/non-conversion processing as before, and the control processing unit 9 outputs the kanji or kana as is to the character string output 10.

With this process, the conversion process in the kana-kanji converter 8 is not necessary for words, and only the search in the word table 7 is required, so that the process becomes extremely fast.

Table 1 (one example of a word table 7, which stores word notation and kana notation, and usually outputs word notation.

Table 1 Figure 2 is an electrical block diagram for explaining an example of outputting kana notation. In the figure, 11 is a control signal for calling kana notation, 12 is a kana notation output signal, and others. Parts which function similarly to the embodiment shown in FIG. 1 are given the same reference numerals as in FIG. In this embodiment, the control processing unit 9 sends the kana notation output signal 12 of the word table 7 to the kana-kanji conversion unit 8 by the control signal 11 that calls the kana notation, and converts the kana to kanji to obtain homophones. You can choose the word. Through this process, the word number 2 "cooperation" shown in Table 1 can be expressed as "strong" through a secondary conversion.

Effects As is clear from the above explanation, according to the present invention, the conversion process when inputting words by voice can be realized by searching the word table, so there is no need to perform the kana-kanji conversion process, and the processing time is shortened. can. Furthermore, by storing kana notation in the word table, homophones can be selectively written.

[Brief explanation of drawings]

1 and 2 are electrical block diagrams for explaining the present invention in detail, respectively. 1...Microphone, 2...Word recognition unit, 3...
Monosyllable recognition unit, 4... Word and monosyllable discrimination unit, 5...
- Word recognition results, 6... Monosyllable recognition results, 7... Word table. 8... Kana-kanji conversion unit, 9... Control processing unit, 10
...Character string output, 11...Control signal, 12...Output signal.

Claims (4)

[Claims] (1) In a voice input word processor consisting of a word recognition unit that recognizes voice input in units of words, a monosyllable recognition unit that recognizes voice input in units of monosyllables, and a kana-kanji conversion unit that converts the results into character strings, word A voice input word processor characterized in that the recognition results are expressed as they are, and no kana-kanji conversion is performed. (2) The voice input word processor according to claim (1), further comprising a word table that indicates the content of word recognition results in parallel with the kana-kanji converter. (3) The voice input word processor according to claim (2), wherein the word table includes at least word notation. (4) The voice input word processor according to claim (3), further comprising means for selecting homophones using the kana character notation.