JPS60122429A - Monosyllable voice input system - Google Patents

Abstract

PURPOSE:To obtain the character display of an intended foreign word by designating the foreign word input so as to reference a foreign word correcting table thereby adding a proper monosyllable candidate for character display. CONSTITUTION:An input voice is collated with a monosyllable standard pattern 2 registered in advance at a monosyllable recognizing section 1, and the candidate group of each syllable as the result of recognition is fed to a character string forming section 3 together with information such as similarity and characteristic amount of similarity. The forming section 3 references a foreign word correction table 4 based on the designation of foreign word input. The monosyllable candidate suitable for the character display is added to form a character string and the result is fed to a Kanji (Chinese character) Kana (Japanese syllabary) mixed sentence converting section 5, where a foreign word dictionary 6 is collated, the candidate is displayed and a Japanese voice is transmitted while selecting the intended candidate.

Description

DETAILED DESCRIPTION OF THE INVENTION (1) Technical field of the invention The present invention relates to foreign words used in audio word processors,
The present invention relates to a monosyllabic speech input method that can easily correct when the recognition result obtained when monosyllabic recognition is performed on input speech differs from normal character display (reading).

(2) Prior Art and Problems Conventionally, in Japanese audio word processors, the audio of a foreign word cannot be directly displayed in Japanese Katakana, but is sometimes displayed as a combination of two characters. For example, "file"
If you enter a foreign word ``'' and recognize it as a monosyllable, it will say ``
With a recognition device that does not recognize ``F'', ``7A'' becomes ``HA''.
If it is recognized as ``Baile'', it is inappropriate as a character display, and it is necessary to correct the recognized ``Ha'' to ``Fa''. Of course, the voice input is "F", "A", "E1."
If you input "Le", the text will be displayed properly, but the voice input will be unnatural (2), and you will be using the voice word processor without its features. Similarly, in the case of ``fo'', even if it becomes ``ho'' when recognized as a single syllable, there are cases where it is desired to display the character as ``fo''. In this way, even if a single katakana character is recognized in monosyllabic recognition, it may not be appropriate as Japanese if it is not expressed as two katakana characters in character display. However, since the input specification for foreign words is in Katakana, there is no one that has the function of converting two characters for one character.

Also, when inputting foreign words by voice, fluctuations in pronunciation (
There are vowel extensions), and simply recognizing them as they are will result in a monosyllabic character string that is not in the foreign word dictionary, and it will not be possible to obtain the intended foreign word. For example, when "ri" in "computer", "motor", etc. is input by voice, in monosyllabic recognition, long sound processing is applied after the monosyllabic candidate "ta 1".
"tar" is recognized. Therefore, in this case, it is necessary to correct "tar" to "yu".

That is, in monosyllabic recognition, a long sound is added to one kana character and it is recognized, and in character display, it may not be possible to match it with a dictionary unless it is expressed using that one kana character.

(6) The same applies to "sir".

As mentioned above, when the recognition result of monosyllabic recognition differs from the written expression, it is necessary to correct it, but in the conventional method, it is necessary to perform this correction to obtain the intended foreign word. No measures were taken.

(3) Purpose of the Invention The purpose of the present invention is to easily correct the recognition result when monosyllabic recognition is performed on the input speech of a foreign word used in a spoken word processor, which differs from the normal character display. It is an object of the present invention to provide a monosyllabic voice input method that can be used.

(4) Structure of the Invention In order to achieve the above object, the monosyllabic speech input method of the present invention inputs unknown input speech into a monosyllabic recognition unit, compares it with a monosyllabic standard pattern registered in a dictionary in advance, and obtains the recognition result. Put the monosyllable candidate group into the character string creation section, specify the foreign word input, refer to the foreign word correction table, add appropriate monosyllable candidates to the character display, and check against the foreign word dictionary. Yoshii (4) This is characterized by obtaining the character representation of the foreign word shown in the figure.

(5) Embodiments of the invention Generally, when a foreign word is to be recognized, the pronunciation of the foreign word has many unstable elements, and the content of the pronunciation and the content written in Katakana may not match. There are many.

Therefore, even if the recognition result is correct, it often happens that the foreign word cannot be input because it does not match the contents of the foreign word dictionary.

First, based on statistical data, we have a matrix that indicates which monosyllable out of 101 monosyllables a foreign word will be, and then create a system by having the individual pronounce the foreign word after registering the monosyllables. , prepare a loanword correction table with a matrix of speech input persons, and refer to this table during recognition, and add appropriate monosyllable candidates to the character display (reading) of monosyllables that are not subject to recognition. I can do it. In this case, the recognition result of "Ha" can be corrected to "F".

(5) In addition, by creating rules based on statistically generated pronunciation fluctuations of foreign words, storing them in the same table as above, and referring to the table during recognition, similarly, monosyllabic candidates can be generated. can be added.

As a result, the recognition result of "shi" can be corrected to "yu".

The figure is a configuration explanatory diagram of an embodiment of the present invention.

In the figure, the input speech is input to the monosyllable recognition unit 1, and it is compared with the monosyllable standard pattern 2 registered in the dictionary in advance (pattern learning), and each monosyllable candidate group of the N recognition results is determined from the recognition result. It is sent to the character string creation unit 3 along with information such as similarity and likelihood features. This character string creation unit 3 inputs the foreign word input specification (katakana specification) along with the sent information.
Refer to foreign word correction table 4 (single syllable error learning).

In this foreign word correction table, monosyllable candidates corresponding to the above-mentioned character display outside the U recognition target and fluctuations in utterance are stored together with the speech input person's matrix.

In either case, a character string is created by adding appropriate monosyllable candidates to the character display, and sent to the Kanji/Kana-mixed sentence conversion section 5.

(6) The kanji/kana mixed p-sentence conversion unit 5 compares the sent character string with the foreign word dictionary 6 (frequency learning), performs kanji/kana mixed p-sentence conversion, displays candidates, and While selecting the selected one, input the Japanese voice.

(6) As described in detail, according to the present invention, when the recognition result when monosyllable recognition is performed on the input speech of a foreign word used in an audio word processor is different from the character notation in Katakana. In order to perform the correction, the system refers to the foreign word correction table according to the foreign word input specification, adds appropriate monosyllable candidates to the character display, and obtains the intended foreign word. As a result, as mentioned above, even if the monosyllable candidate in the recognition result of the input speech is recognized as "bar 1" or "ho 1", the monosyllable candidate added from the loanword correction table is "7a" or "ho1". Similarly, even if a single syllable candidate as a U recognition result of the input speech is recognized as ``sir'' or ``sir'' by long sound processing, Loan word correction technique (7) The monosyllable candidates added every time are ``ta 1'' or ``su'', so that the intended loan word can be obtained.

[Brief explanation of drawings]

The figure is an explanatory diagram of the configuration of an embodiment of the present invention. In the figure, 1 is a monosyllable recognition unit, 2 is a monosyllable standard pattern, 3 is a character string creation unit, 4 is a foreign word correction table, and 5 is a large mixture of kanji and kana. The conversion part 6 indicates a word-hakujiho (loan word). Patent applicant Fujitsu Ltd. Sub-agent Patent attorney 1) Yoshishige Saka (8)

Claims (3)

[Claims] (1) Enter the unknown input speech into the monosyllabic recognition section, match it with the monosyllabic standard pattern registered in the dictionary in advance, enter the monosyllabic candidate group resulting from the recognition into the character string creation section, and specify the foreign word input. A monosyllabic method characterized in that the character display of the intended foreign word is obtained by further referring to a loan word correction table, adding appropriate monosyllabic candidates to the character display, and comparing with a word dictionary of the loan word. Voice input method. (2) Even if the monosyllabic candidate in the recognition result of the input speech is recognized as "p" or "尭", the monosyllabic candidate added according to the loanword correction table is "fa" or "huo". A monosyllabic voice input method according to claim (1). (3) Even if the monosyllable candidate in the recognition result of the input speech is recognized as "shi" or "sah" by the long sound processing, the monosyllable candidate added by the loan word correction table is (1) "ri". ” or “su”, the monosyllabic voice input method according to claim (1).