JPS6340936A - Voice kana to kanji conversion system - Google Patents

Abstract

PURPOSE:To execute a voice 'Kana' (Japanese syllabary) to 'Kanji' (Chinese character) conversion by using a 'Kana' character-string outputted from an input device, and intonation information outputted from an intonation detecting circuit, as keys, and retrieving a dictionary. CONSTITUTION:For instance, when a word of 'Hashi' is pronounced, an input device 1 detects a 'Kana' (Japanese syllabary) character-string of 'Hashi' corresponding to said word. Simultaneously, an intonation detecting circuit 2 detects the pronunciation of 'Ha' and 'Shi' is high or low. Subsequently, a dictionary retrieving circuit 4 retrieves a dictionary by using the 'Kana' character-string of 'Hashi' as a key, and in case when a homonym exists, an intonation deciding circuit 5 is started, and intonation detected from the input 'Kana' character-string and intonation information of the dictionary are compared and decided. As a result, a notation conforming to a word which is brought to a voice input is determined, and it is transferred to a 'Kanji' (Chinese character) and 'Kana' mixed character-string forming circuit 6, and displayed.

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Industrial Application] This invention relates to a phonetic kana-kanji henhoumanshiki G method for obtaining a character string containing kanji and kana from one pronounced Japanese.

[Prior Art] Figure 3 is a block diagram of a conventional phonetic kana-kanji conversion method disclosed in, for example, Japanese Patent Application Laid-Open No. 60-54041. (3) is a dictionary storage means, (4) is a dictionary retrieval means, and (6) is a means for forming a character string containing kanji and kana.Figure 4 shows a conversion dictionary used in the conventional phonetic kana-kanji conversion method. The reading is associated with the corresponding notation.

Next, the operation will be explained. For example, suppose that the word "Hashi" is pronounced. The voice input means (1) detects the kana character string "Hashi". The dictionary search means (4) retrieves the dictionary shown in Figure @4 from the dictionary storage means (3),
Search for the notation that corresponds to the pronunciation of "Hashi". The kanji-kana mixed character string forming means (6) forms three types of homonyms, ``r-end'', ``author'', and ``彑'', and outputs them for display.

[Problem that the invention seeks to solve]

The conventional phonetic kana-kanji conversion method is structured as described above, so if there are multiple notations for one reading, it is impossible to identify them, and another selection operation method is required. There was a problem.

This invention was made to solve the above-mentioned problems, and automatically determines words that have different intonations in one pronunciation, even if they are homophones.

The purpose of this invention is to obtain a system that can automatically select a desired notation from among these homonyms.

[Means for solving problems]

The phonetic kana-kanji conversion method according to the present invention includes a means for detecting the intonation of a kana character string based on the input voice, which newly has intonation information of the reading in addition to the reading and notation as dictionary information, and The apparatus is provided with means for comparing and determining the intonation and the intonation/non-sound information in the dictionary.

[Effect]

In the phonetic kana kanji henkaku method in this invention.

When a homophone occurs, a desired notation is automatically selected and displayed from among the homophones by further matching the intonation information.

〔Example〕

Hereinafter, an embodiment of the present invention will be described with reference to the drawings. 11
In the g1 diagram
) is a dictionary search means, (5) is an intonation determining means, and (6) is a character string forming means containing kanji and kana characters.

FIG. 2 shows an example of the contents of the dictionary stored by the dictionary storage means (3). In addition to the reading and notation, new intonation information has been added. The pronunciation of ``Hashi'' shown in this example (on the other hand, it expresses cases where there is a difference in pitch between the pronunciations of ``ha'' and ``shi'', or cases where there is no difference in pitch).

The intonation information "00" means "ha" and "
When there is no difference in pitch of pronunciation, 1 notation is ``edge''.
corresponds. Similarly, "lO" is a case where "ha" is higher than "shi", and the notation corresponds to "chopsticks".

Next, I will explain the operation. For example, suppose that the word ``Hashi J'' is pronounced. The voice input means [1] detects the kana character string "Hashi". At the same time, the intonation detection means (2) detects the pitch of the pronunciation of "r wa" and "shi", and if the pitch is "ha" = "shi", it will be "00", "wa J>"
In the case of rL and J, the value is rLOJ, and in the case of "r"<"shi", the value "01" is output.

The dictionary search means (4) searches the dictionary from the dictionary storage means using the kana character string "Hashi" as a key, and obtains three homophones as shown in FIG. If there is no homophone A coined word, the dictionary search means (4) transfers the obtained notation as is to the Kanji-kana mixed character string formation means (6) and displays it, but if there is a ζ homophone synonym as in this example The intonation determining means (5) 8 is activated to compare and determine the intonation detected from the input kana character string and the intonation information in the dictionary. As a result, the voice was written as a single i
A suitable notation is determined, and it is transferred to the Kanji/Kana mixed character string forming means (6) and displayed.

By the way, in this embodiment, if the intonation information stored in the dictionary storage means (3) is different from the intonation of the person who pronounces the pronunciation, the intonation information in the dictionary is corrected.2 For example, several pieces of different intonation information are prepared and selected. It may be used by switching between the two.

[Effects of the Invention] As described above, according to the present invention, a means for automatically classifying homophones based on intonation information is provided, so that a highly accurate phonetic kana-kanji conversion system can be obtained.

[Brief explanation of drawings]

@Figure 1 is a block diagram of a phonetic kana-kanji conversion system according to an embodiment of this invention, Figure 2 is a block diagram of a dictionary according to this invention, and Figure 3 is a diagram of a dictionary configuration according to this invention.
The figure is a completed diagram of a conventional embodiment, and FIG. 4 is a diagram of a conventional dictionary configuration. In the figure, (13 is a voice input means, (2) is an intonation detection means, +3 is a dictionary storage means, and (4) is
! 4! (5); intonation determining means; and (6) means for forming character strings containing kanji and kana characters. Note that the same reference numerals in Figure 1 indicate the same or equivalent parts.

Claims (2)

[Claims] (1) In a phonetic kana-kanji conversion method that obtains a character string containing kanji and kana from pronounced Japanese, there is a voice input means that recognizes the pronounced Japanese sentence and outputs a kana character string, and a voice input means that outputs a kana character string. intonation detection means for detecting intonation of each word of a Japanese sentence; dictionary storage means for storing a dictionary having at least reading and intonation information of a plurality of words and their notations; Dictionary search means for searching the dictionary using the entire or part of the kana character string output from the voice input means and the intonation information output from the intonation detection means as keys to obtain a matching notation. A phonetic kana-kanji conversion method characterized by the following. (2) If the intonation information in the dictionary is different from the intonation of the person who is making the pronouncement, the intonation information of the dictionary can be modified to match the intonation of the person who is making the pronouncement. The phonetic kana-kanji conversion method described in scope 1.