JPS6168669A - Voice response word processor - Google Patents

Abstract

PURPOSE:To improve the input speed of a sentence by voice by displaying next candidate of syllable recognition and honomyms of Kana-Chinese character conversion according to key disposition, and by making selection of them by selective use of the same function key. CONSTITUTION:Plural syllable candidates are outputted by a syllable recognition means 1 and they are displayed by display means 5 corresponding to the suitable selection key of a keyboard 7. The objective syllable candidate is selected by the first selection means 2 which is activated by the input from the keyboard. And, honomyms for the input obtained by a kana-Chinese character conversion means 3 are displayed in the same manner as the case of syllables. The objective Chinese character is selected by the second selection means 4 by the instruction from the keyboard 7.

Description

DETAILED DESCRIPTION OF THE INVENTION Field of the Invention The present invention relates to improving the operability of a word processor that uses voice input.

Configuration of conventional examples and their problems Conventionally, a speech recognition device and a Japanese food processor are combined to recognize input speech in units of syllables or similar units, such as vowel-consonant-vowel chains (hereinafter abbreviated as VCV chains). A voice input word processor has been reported that converts this into kana-kanji and outputs a sentence containing kanji and kana. However, this type of word processor has drawbacks such as selecting the next candidate for syllable recognition and selecting one homophone for kana-kanji conversion. In other words, recognition errors occur in syllable recognition. The best way to deal with this is awareness.

Also, in kana-kanji conversion, many so-called homonyms, which have the same pronunciation but different spellings, may occur. In this case, the selection method is to display the homonyms one after another each time you press the corresponding key on the keyboard, and to confirm the corresponding kanji when the next hiragana is input, or to add numbers to multiple homonyms. Some homonyms are determined by selecting the corresponding kanji using the attached numeric keypad. In the former case, it is unclear whether the desired kanji is included in the homonym candidates, and even if it is included, it is unclear in what order it will appear, which poses a problem in terms of operability. The latter has the disadvantage that it is necessary to visually confirm the position of the numeric keypad after it is displayed, resulting in movement of the line of sight, and that high-speed input is not possible until the location of the numeric keypad is memorized.

Whichever method is chosen, it is necessary to select the next syllable error candidate by key operation and then select the homophone again by another key operation, which has the disadvantage that high-speed text input is not possible.

Purpose of the Invention In view of the above drawbacks, an object of the present invention is to provide a speech word processor that can perform high-speed input by standardizing the next candidate selection method and homophone selection method in speech recognition, as well as the display method. do.

Structure of the Invention In order to achieve this object, the present invention uses a plurality of recognition candidates obtained by a syllable recognition means and a plurality of homophones obtained by a kana-kanji conversion means together with the key arrangement of a predetermined part of the keyboard. This is an audio word processor whose display corresponds to the key layout, and is configured so that syllable candidates and homophones can be selected using the same key.

DESCRIPTION OF EMBODIMENTS An embodiment of the present invention will be described below with reference to the drawings.

FIG. 1 is a block diagram of an audio word processor in one embodiment of the present invention. In the figure, reference numeral 1 denotes a syllable recognition means for recognizing input speech in units of syllables or similar units of VCV, and a known monosyllable recognition device that recognizes 68 or 101 syllables is used. Reference numeral 2 denotes a first selection means, which selects one of the plurality of recognition candidates obtained by the syllable recognition means 1 in response to an instruction from the keyboard 7. Reference numeral 3 denotes a kana-kanji conversion means, which performs kana-kanji conversion on a group of syllables sequentially sent from the syllable recognition means 1 and the first selection means 2. Reference numeral 4 denotes a second selection means which, if a homophone is found in the result of the kana-kanji conversion by the kana-kanji conversion means 3, selects the corresponding kanji by designation from the keyboard 7.

Reference numeral 5 denotes a display means that displays a plurality of recognition candidates obtained from the syllable recognition means 1 and homophones obtained from the kana-kanji conversion means 3 on the display section 6 in correspondence with the key arrangement of the keyboard 7.

Next, the operation of the apparatus configured as described above will be explained using a specific example shown in FIGS. 2 and 3.

Now, if you input the sounds "se" and "tsu",
For example, when trying to obtain the kanji ``se'', when the first syllable ``se'' is input, the syllable recognition means 1 recognizes as follows.
“te” in “se”, “ze”, and “he” in order of certainty
・Suppose that we have obtained five candidates for "be". The display means 6 displays these six candidates on the display section 6 in correspondence with the key arrangement of a predetermined portion of the keyboard, such as function keys, for example, as shown at 10 in FIG. 2a. In this case, there is no recognition error and "se" is obtained as the first candidate, and the next candidate is not selected using the key. Next, when "tsu" was uttered, there was an error in syllable recognition, and as shown in Figure 2, Part 11, "su" was obtained as the first candidate, and "tsu" was the third candidate. Suppose it became. At this time, key 1 from the keyboard corresponding to the display of "tsu"
By pressing 2, the candidate "tsu" is selected.

The kana-kanji conversion means 3 converts the input kana string "setsu" into kana-kanji. Assume that ``section'', ``setsu'', ``juni'', and ``ori'' are obtained as a result. The display means 5 shows the display section 6 as follows.
Similar to Figures 2a and 2b, homophones are displayed as shown in 13 of Figure 3-.

When the operator presses the key 14 that corresponds to the desired kanji ``taku'' from among the homophones displayed on the keyboard 7, the second selection means 4 selects ``taku'' and displays it through the display means 5. Displayed in section 6.

As described above, according to this embodiment, the syllable recognition means 1 outputs a plurality of syllable recognition candidates, the display means 6 displays them in correspondence with the corresponding selection keys of the keyboard 7, and the keys are manually operated. The target syllable candidate is selected by the first selection means 2, and then the homophone obtained by the Kana-Kanji conversion means 3 is displayed in response to the input in the same manner as the syllable recognition candidate, and the second one is selected by the instruction from the keyboard 7. By selecting the target kanji using the selection means, it is possible to provide an audio word processor that is highly operable and capable of high-speed text input.

Note that each component in this embodiment can be implemented programmatically by replacing part or all of it with a computer. It is also possible to provide a separate key for the display corresponding to the function key in Fig. 2, so that it can be erased as needed by operating this key, and the input text can be displayed at the same position as the display. be.

Effects of the Invention As described above, the audio word processor of the present invention displays the key arrangement of a predetermined part of the keyboard, such as a group of function keys, as well as the next candidate for syllable recognition and the homophone for kana-kanji conversion in correspondence with this key arrangement. By making the selection by dividing the same function key, it is possible to speed up the input of text by voice, and its industrial value is enormous.

[Brief explanation of the drawing]

FIG. 1 is a block diagram of a voice word processor according to an embodiment of the present invention, and FIGS. 2 and 3 are diagrams showing display examples in the same embodiment. 1...Syllable recognition means, 2...First selection means, 3...Kana-kanji conversion means, 4...
...Second selection means, 6...Display means, 6...
...display section, 7...keyboard. Name of agent: Patent attorney Toshio Nakao and one other person
Figure 2 (S(b)

Claims (1)

[Claims] a syllable recognition unit that recognizes input speech syllable by syllable and sequentially outputs a plurality of recognition candidates; a display unit that displays the plurality of recognition candidates; and one of the recognition candidates is selected by an instruction from a keyboard. a first selection means for selecting the first selection means;
a kana-kanji conversion means that converts the syllable string sequentially obtained according to the selection result of the selection means to kana-kanji and outputs the corresponding kanji and homonyms, and selects one of the homonyms according to instructions from the keyboard. a second selection means, the display means displays a plurality of recognition candidates obtained by the syllable recognition means and a plurality of homophones obtained from the kana-kanji conversion means together with a key arrangement of a predetermined portion of the keyboard. A voice word processor characterized in that the voice word processor is configured to be displayed on the display unit in correspondence with a key arrangement.