JPS59176835A - Method and device for producing sound input sentence - Google Patents

Abstract

PURPOSE:To facilitate easy operation for the production of a sentence and at the same time to improve the processing efficiency by producing a sound input sentence without confirming successively the KANA (Japanses syllabary) character string corresponding to the sound input. CONSTITUTION:The KANA character string confirmed by a sound confirming part 3 is collated with the confirmation correspondence career list contained in a file device 1. The presence or absence is checked for the next candidate KANA character string different from said confirmed KANA character string. The result of this check is added to an analysis route. A KANA/KANJI (Chinese character) converting part 6 analyzes the form elements for all KANA character strings corresponding to each analysis route and produces a key KANJI-KANA character string for each route. The KANJI-KANA character train having the minimum number of form elements is selected and displayed on a display 4. When the presence of an error is informed from a console 5, the next candidate KANJI-KANA sentence is displayed. When this sentence is correct, it is sent to a sentence editing part 8. Then a series of sentences are edited.

Description

DETAILED DESCRIPTION OF THE INVENTION (a) Technical Field of the Invention The present invention relates to a text creation device, and more particularly to a device for converting a voice-input sentence into a kanji (kanji) or kana-mixed sentence using an electronic circuit.

(b) Technical Background Recently, text creation devices called Japanese word processors have become popular, and their input methods can be broadly divided into two types: the kanji tablet method and the "kana" kanji conversion method.

The former allows you to search for characters in a dictionary by selecting the required character from the kanji/kana characters displayed on the tablet, while the latter allows you to enter sentences word by word using kana characters. , the necessary word is converted into Kanji for each input, and in the case of homophones, the reconversion is repeated to obtain the necessary Kanji.

(C) Prior art and problems In conventional document creation devices, character tablets such as those mentioned above are used, or "kana" and "kana" characters are input from the keyboard.
It was mainly used to input characters such as numbers.

On the other hand, the development of the most natural voice input/output method as a man-machine dialogue method is progressing, and attempts have been made to apply the voice input method to the above-mentioned text creation devices. is a system that sequentially displays ``kana'' input by voice on a display screen to create a ``kana'' character string, and converts the ``kana'' character string into ``kana'' into kanji.
Since it is necessary to constantly check the "kana" characters appearing on the display screen, it is not as easy to use as the keyboard input method, and improvements have been desired.

(d) Object of the Invention An object of the present invention is to enable text creation by voice input without sequentially checking the "kana" character string corresponding to the voice input.

(e) Structure of the Invention The present invention provides a recognition correspondence history table (for example, when a certain "kana" character "chi" is input by voice, when creating a voice input text using "kana" kanji conversion, If it is recognized as “wa”, other “
A table consisting of statistical data of the recognition frequency in each case, such as when the character is recognized as a kana character, is stored in advance in the same file as the voice pattern, and the recognition correspondence history table is stored each time the voice input sentence creation process starts. is transferred to the text creation department, and the recognized "kana" character string is compared with the recognition correspondence history table to verify whether there is a possible "kana" character string other than the recognized "kana" character string, and to determine whether there is a possible "kana" character string. If a “kana” character string exists, an analysis route corresponding to this is established,
It is characterized in that each of the separately possible "kana" character strings is sequentially analyzed for each analysis route and a sentence containing kanji "kana" is output.

Examples of Tf1 invention Embodiments of the present invention will be described below with reference to the drawings.

FIG. 1 and FIG. 2 are functional block diagrams of a conventional voice input sentence creation device and a voice input sentence creation device according to the present invention, respectively.

In FIG. 1, a standard voice pattern for each voice input operator is registered in advance in a file device 1. The operator's voice input from the microphone 2 is processed by the voice recognition unit 3 into 1 syllable (generally "kana" character)! After the pattern is decomposed into digits, it is compared with the standard speech pattern of the file device 1 and recognized as the most similar syllable in the standard speech pattern, that is, the "kana" character.

The character string ``1 Kana'' that corresponds to the input voice recognized in this way is displayed on the display 4, and if it is determined that the voice input ``Kana'' character has been correctly recognized, the character string ``Kana'' is displayed on the console 5. The execution is started by instructing execution of conversion or by detecting a punctuation mark, etc., and the recognized “kana” character string is converted into “kana” by the “kana” kanji conversion unit 6.
(The method is to divide the "kana" character string into various units called morphemes while referring to a word dictionary, and check the connection information between each morpheme included in the morpheme, the longest morpheme, etc. decomposes the "kana" character string into parts of speech based on the longest match method and examines the connections between the parts of speech), and selects the most probable word arrangement, that is, a combination of kanji and kana characters, and this is the kanji "kana". It is displayed on the display 4 as a mixed sentence.

If there is an error in the "kana" character string displayed on the display 4, the operation console 5 notifies you of this, and the user inputs the "kana" character again aloud from the microphone 2 to check whether it has been correctly recognized. Check on Display 4. In addition, if there is an error in the kanji character string "kana" displayed on the display 4, input that fact from the console 5, and the next candidate "kana" character string will be output to the speech recognition unit 3. (The voice recognition unit 3 sets multiple candidate recognition "kana" character strings based on the degree of closeness between the input voice and the voice pattern of the file device 1, and compares the distance between each candidate with a threshold value.) (1) Obtaining similarity information (1) Forming nodes between candidates whose distance is less than the threshold value and selecting each candidate in turn), "Kana" Kanji conversion is performed for the "Kana" character string of the next candidate, The corresponding kanji "kana" mixed sentence is output on the display 4.

In the present invention, a recognition correspondence history table as described above is stored in advance in the file device 1 that stores standard voice patterns for each voice input operator. Further, as shown in FIG. 2, a "kana" preprocessing section 9 and a "kana" postprocessing section 10 are provided before and after the "kana" kanji conversion section 6.

In FIG. 2, with the start of the voice input sentence creation process,
The recognition correspondence history table stored in the file device 1 is transferred to the file device 11 of the “kana” preprocessing section 9.

The voice input “kana” characters from microphone 2 are sent to file device 1.
The voice pattern and the feature points are compared, sequentially recognized by the voice recognition section 3, and stored in the file device 11 as a recognized "kana" character string. In this case, in order to verify whether there is a possibility that the input "kana" character was misrecognized, the recognized "kana" character string is compared with the recognition correspondence history table, and the recognized "kana" character string is compared with the recognition correspondence history table.
If there is a possibility that there is a different “kana” character string from the “kana” character string, it is added to the analysis route of the next candidate “kana” character string in the speech recognition unit 3, and the other “kana” character string is analyzed. A route will be established to analyze each.

Thereafter, in the same manner as before, in the "kana" kanji conversion unit 6, morphological analysis is sequentially performed on all "kana" character strings corresponding to the above root while referring to the word dictionary stored in the file device 7. , the most probable representative kanji "kana" mixed string is created for each route, and among these representative kanji "kana" mixed strings, the number of morphemes that make up the kanji "kana" mixed string is the smallest. is selected.

This is displayed on the display 4 as a sentence containing the kanji "kana", and its validity is determined by the voice input operator.

If it is determined that there is an error in the kanji ``kana'' character displayed on the display 4, input that information from the console 5, and the ``gana'' kanji converter 6 will select the next candidate. The transmission of the kanji "kana" mixed string is requested, and the corresponding kanji "gana" mixed string is displayed on the display 4.

When a message indicating that a sentence containing the correct kanji "kana" has been output is input from the console 5, this is temporarily stored in the sentence editing section 8, and
Sentences containing individual kanji "kana" are edited as a series of sentences.

In the present invention, the “kana” character string corresponding to the finally selected kanji “kana” mixed string and the “kana” character string recognized by the speech recognition unit 3 are processed in the “kana” post-processing unit 10. File device 1 is compared based on the correspondence relationship.
The data in the recognition correspondence history table stored on 1, that is, the recognition correspondence degree value is updated.

The updated recognition correspondence history table is transferred to the file device 1 when the voice input operation is completed.

The sentence containing the kanji "kana" created as described above is printed out by the printer 12.

1 and 2, reference numeral 13 represents an output control device for the display 4 and printer 12. In FIG.

(8) Effects of the Invention According to the present invention, there is no need to sequentially check the "kana" characters that are input by voice and convert them into "kana", and it is not necessary to convert them into "kana" or "kana" characters of any given length. It is sufficient to judge the correctness of recognition/conversion based on the state, which has a great effect of facilitating the voice input sentence creation operation and improving processing efficiency.

[Brief explanation of the drawing]

FIG. 1 and FIG. 2 are functional block diagrams of a conventional voice input sentence creation device and a voice input sentence creation device according to the present invention, respectively. In the figure, 1 is a file device, 2 is a microphone, 3 is a voice recognition unit, 4 is a display, 5 is a console, and 6 is a “kana”
7 and 11 are file devices, 8 is a text editing section, 9 is a "kana" pre-processing section, 10 is a "kana" post-processing section, 12 is a printer, and 13 is an output control device. See 1 figure

Claims (4)

[Claims] (1) Voice input sentence creation blade using “kana” kanji conversion
In the formula, the recognized "kana" character string is compared with the recognition correspondence history table to verify the presence or absence of a possible "kana" character string other than the recognized "kana" character string, and to determine the presence or absence of a possible "kana" character string separately from the recognized "kana" character string. If it exists, a corresponding analysis route is established, and for each analysis route, each of the separately possible "kana" character strings is sequentially analyzed and a sentence containing kanji "kana" is output. Voice input text creation method. (2) The recognition correspondence history table is stored in the same file as the voice pattern for each voice input operator, and is transferred to the text creation section at the time of voice input operation. Voice input text creation method. (3) Compare the correspondence of each “kana” character in the recognized “kana” character string and the “gana” character string corresponding to each “kana” character and the output kanji “kana” mixed sentence, The voice input sentence creation method according to claim 1, wherein the content of the recognition correspondence history table is updated based on the correspondence relationship. (4) In a voice input text creation device that uses “kana” Kanji conversion, create and control multiple analysis routes based on a voice pattern file storing a recognition correspondence history table and the recognition correspondence history table and voice input information. for"
The Kana/Kanji preprocessing unit compares each “Kana” character in the recognized “Kana” character string with the corresponding “Kana” character in the “Kana” character string corresponding to the output Kanji “Kana” mixed sentence. and a "kana" post-processing unit for updating the contents of the recognition correspondence history table based on the correspondence relationship.