JPS63106698A - Voice input kana-kanji converter - Google Patents
Voice input kana-kanji converterInfo
- Publication number
- JPS63106698A JPS63106698A JP61252603A JP25260386A JPS63106698A JP S63106698 A JPS63106698 A JP S63106698A JP 61252603 A JP61252603 A JP 61252603A JP 25260386 A JP25260386 A JP 25260386A JP S63106698 A JPS63106698 A JP S63106698A
- Authority
- JP
- Japan
- Prior art keywords
- input
- kana
- speech
- kanji
- voice
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000006243 chemical reaction Methods 0.000 claims description 21
- 239000002245 particle Substances 0.000 description 6
- 238000000034 method Methods 0.000 description 4
- 238000010586 diagram Methods 0.000 description 2
- 235000016496 Panda oleosa Nutrition 0.000 description 1
- 240000000220 Panda oleosa Species 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 230000003442 weekly effect Effects 0.000 description 1
Abstract
(57)【要約】本公報は電子出願前の出願データであるた
め要約のデータは記録されません。(57) [Abstract] This bulletin contains application data before electronic filing, so abstract data is not recorded.
Description
【発明の詳細な説明】
五先分互
本発明は、音声認識によってかな漢字文書を作成する装
、置に係るものであり１例えば、ＯＣＲ入力のかな漢字
変換装置等にも応用可能なものである。DETAILED DESCRIPTION OF THE INVENTION The present invention relates to a device for creating a Kana-Kanji document by speech recognition, and can be applied to, for example, a Kana-Kanji conversion device for OCR input.
従ｍ桁
従来の音声入力かな漢字変換装置においては、入力可能
な音声の種類が１種類であるために、入力された音声の
確からしさを示す尤度や、限られた文法情報しか用いる
ことができないままかな漢字変換を行っていた。しかし
、この様な、かな漢字変換方法では充分に効率のよい結
果が得られな一鳥。Conventional voice input kana-kanji conversion devices can only input one type of voice, so only the likelihood that indicates the certainty of the input voice and limited grammatical information can be used. Performed makana kanji conversion. However, this type of kana-kanji conversion method does not provide sufficiently efficient results.
−−ｈ
本発明は、上述のごとき実情に鑑みてなされたちの′で
、特に、入力された文字（列）を効率よくかな漢字変換
を行うことを目的としてなされたものである。--h The present invention was developed in view of the above-mentioned circumstances, and is particularly aimed at efficiently converting input characters (sequences) into kana-kanji.
週ニーー成。Weekly nee-sei.
本発明は、上記目的を達成するために、音声分析部把、
複数の異なる種類の音声を認識する認識部と、認識結果
のかな漢字変換をおこなうかな漢生変換部とを具備し、
入力された音声の認識結果の種類が、直前に入力されて
認識されていた音声の種類と異なり、かつ、直後に入力
された音声の種類とも異なる時、該入力音声は特別な文
法的情報をもっている可能性があると判断してかな漢字
変換をおこなうことを特徴としたものである。以下、本
発明の実施例に基づいて説明する。In order to achieve the above object, the present invention includes a speech analysis unit,
It is equipped with a recognition unit that recognizes a plurality of different types of speech, and a kana-kanji conversion unit that converts the recognition results into kana-kanji.
When the type of recognition result of input speech is different from the type of speech that was input and recognized immediately before, and also different from the type of speech input immediately after, the input speech has special grammatical information. The feature is that it performs kana-kanji conversion after determining that there is a possibility that Hereinafter, the present invention will be explained based on examples.
第１図は１本発明の一実施例を説明するための全体ブロ
ック構成図で１図中、１は入力部、２は音声分析部、３
は音声認識部、３ａは単語認識部。FIG. 1 is an overall block configuration diagram for explaining one embodiment of the present invention. In the figure, 1 is an input section, 2 is a voice analysis section, and 3
3a is a speech recognition unit, and 3a is a word recognition unit.
３ｂは単音節認識部、４１多音声種類判定部、５はかな
漢字変換部、６は出力部で、本発明は、複数の異なる音
声を認識する認識部をもち、認識された音声の種類の情
、報をかな漢字変換部で効率よく活用してかな漢字変換
を行うもので、前記複数の異なる音声を、たとえば単音
節と、それ以外（単語又は文節も含む、以下“単語″と
代表して記す）とするものである。3b is a monosyllable recognition unit, 41 is a multi-phonetic type determination unit, 5 is an ephemeral Kanji conversion unit, and 6 is an output unit. , information is efficiently utilized in the kana-kanji conversion unit to perform kana-kanji conversion, and the plurality of different sounds are, for example, monosyllables and other sounds (including words and phrases, hereinafter referred to as "words"). That is.
第１図において、入力部１より入力された入力音声は、
音声分析部２で特徴量に変換されて、単音節認識部３ｂ
’と単語認識部３ａへ入力されると共に判定部４へ入力
される。判定部４では通常知られている方法で単音節で
あるか単語であるかを判定する。その判定結果により、
単音節認識部３ｂ又は単語認識部３ａの認識結果をかな
漢字変換部５へ送る。かな漢字変換部５においては、従
来知られている方法でかな漢字変換を行う（例えば。In FIG. 1, the input voice input from the input unit 1 is
It is converted into a feature amount by the speech analysis unit 2, and then the monosyllable recognition unit 3b
' is input to the word recognition unit 3a and also input to the determination unit 4. The determining unit 4 determines whether the word is a single syllable or a word using a commonly known method. Based on the judgment results,
The recognition results of the monosyllable recognition section 3b or the word recognition section 3a are sent to the kana-kanji conversion section 5. The kana-kanji conversion unit 5 performs kana-kanji conversion using a conventionally known method (for example.
文節かな漢字変換や最長一致法など）が、Ｔ！１．在入
力された認識結果が単音節であり、かつ直前に入力され
ていた音声の認識結果が単語である時、次の入力音声の
種類の判定待ちになる。そして、次の音声の種類が単語
と判定された時、単語と単語にはさまれた単音節は、助
詞や接続詞などの可能性が高いと判断され、かな漢字変
換がおこなわれる。Bunsetsu kana-kanji conversion, longest match method, etc.) are T! 1. When the currently input recognition result is a monosyllable and the recognition result of the previous input voice is a word, the system waits for determination of the type of the next input voice. When the next sound type is determined to be a word, it is determined that the monosyllables between the words are likely to be particles or conjunctions, and kana-kanji conversion is performed.
以下に簡単な例を示す。A simple example is shown below.
「古拙にかえるが落ちる」という文章の中で、「古拙」
「かえる」 「おちる」が単語登録されていると、す
ると１発声は。In the sentence ``The frog falls in the ancient times,''
If ``frog'' and ``fall'' are registered as words, then one utterance will be...
「フルイケ・二・カエル・ガ・オチル」となる。It becomes "Furuike Ni Kaeru Ga Ochiru".
そこで、「フルイケ」 「二」と入力がおこなわれてい
た段階で、「二」という単音節の前は、ｒフルイケ」と
いう単語であるため１次の音声の入力待ちとなる。続く
、「カエル」が単語と判定された時点で、「二」は単語
と単語にはさまれた単音節であるので、助詞、接続詞の
可能性が高いと判断され、かな漢字変換部において、助
詞の［に」と決定される。又、この時、「に」が「す」
に誤認されたとすると、この場合も前記と同様に「カエ
ル」が単語と判定された時点で「す」は単語と単語には
さまれた単音節であると判断され、助詞。Therefore, at the stage where "Furuike" and "Two" have been input, since the word "r Furuike" comes before the single syllable "Two", the input of the first voice is waited. Next, when "Frog" is determined to be a word, it is determined that since "Two" is a single syllable sandwiched between two words, it is likely to be a particle or conjunction, and the Kana-Kanji converter converts it into a particle. It is determined that the Also, at this time, “ni” becomes “su”
In this case, as in the case above, when ``kaeru'' is determined to be a word, ``su'' is judged to be a monosyllable sandwiched between two words, and becomes a particle.
接続詞の可能性が高いと判断されるが、「すＪは、助詞
、接続詞にはならない、そこで、この場合には誤認識の
可能性が高いと判断する。「す」は母音が「ｉ」であり
、「ミ」 「二」 「イ」などの中でよく誤認識を引き
起こす、その中で、助詞。It is judged that there is a high possibility that it is a conjunction, but ``suJ cannot be a particle or a conjunction, so in this case it is judged that there is a high possibility of misrecognition.``su'' has a vowel of ``i''. Among them, particles that often cause misrecognition such as "mi", "two", and "i".
接続詞となり得るのは「二」であり、接続を調べてみれ
ば、うまくかな漢字変換を行うことが出来る。``Two'' can be a conjunction, and if you look into the conjunction, you can successfully convert kana to kanji.
この様にすれば、かな漢字変換中、たとえば、解析不能
文字が出てきた時、他の候補の中から又は候補外からも
適当な文字を選び出し、効率のよいかな漢字変換を行う
ことができる。In this way, during kana-kanji conversion, for example, when an unanalyzable character appears, an appropriate character can be selected from other candidates or even from outside the candidates, and efficient kana-kanji conversion can be performed.
なお１以上に、入力が音声の場合を例にして説明したが
、たとえば、単語を漢字、単音節をひらがななどに置き
換えることによりＯＣＲ入力装置にも適用できることは
容易に理解できよう。Although the above description has been made using the case where the input is voice, it is easy to understand that it can also be applied to an OCR input device by replacing words with kanji, monosyllables with hiragana, etc.
効果
以上の説明から明らかなように１本発明によると、効率
の高いかな漢字変換を行うことができ、オペレータの負
担を軽くすることが出来る。Effects As is clear from the above description, according to the present invention, it is possible to perform kana-kanji conversion with high efficiency, and the burden on the operator can be lightened.
第１図は、本発明の一実施例を説明するための全体ブロ
ック構成図である。
１・・・入力部、２・・・音声分析部、３・・・音声認
識部。
３ａ・・・単語認識部、３ｂ・・・単音節認識部、４・
・・音声種類判定部、５・・・かな漢字変換部、６・・
・出力部。FIG. 1 is an overall block configuration diagram for explaining one embodiment of the present invention. 1... Input section, 2... Speech analysis section, 3... Speech recognition section. 3a... word recognition section, 3b... monosyllable recognition section, 4.
...Speech type determination unit, 5...Kana-Kanji conversion unit, 6...
・Output section.
Claims (2)
る認識部と、認識結果のかな漢字変換をおこなうかな漢
字変換部とを具備し、入力された音声の認識結果の種類
が、直前に入力されて認識されていた音声の種類と異な
り、かつ、直後に入力された音声の種類とも異なる時、
該入力音声は特別な文法的情報をもっている可能性があ
ると判断してかな漢字変換をおこなうことを特徴とする
音声入力かな漢字変換装置(1) Equipped with a speech analysis section, a recognition section that recognizes multiple different types of speech, and a kana-kanji conversion section that converts the recognition results into kana-kanji, so that the type of recognition result for the input speech is the same as the one immediately before the input. When the type of voice is different from the type of voice that was input and recognized, and also different from the type of voice input immediately after,
A voice input kana-kanji conversion device characterized by determining that the input voice may have special grammatical information and converting it into kana-kanji.
認識誤りを上記特別な文法的情報を用いて訂正すること
を特徴とする特許請求の範囲第（１）項に記載の音声入
力かな漢字変換装置(2) Speech input according to claim (1), characterized in that recognition errors in input speech sandwiched between different types of speech are corrected using the special grammatical information. Kana-Kanji conversion device
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| JP61252603A JPS63106698A (en) | 1986-10-23 | 1986-10-23 | Voice input kana-kanji converter |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| JP61252603A JPS63106698A (en) | 1986-10-23 | 1986-10-23 | Voice input kana-kanji converter |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| JPS63106698A true JPS63106698A (en) | 1988-05-11 |
Family
ID=17239661
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| JP61252603A Pending JPS63106698A (en) | 1986-10-23 | 1986-10-23 | Voice input kana-kanji converter |
Country Status (1)
| Country | Link |
|---|---|
| JP (1) | JPS63106698A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JPH0446397A (en) * | 1990-06-14 | 1992-02-17 | Nec Corp | Continuous voice recognition system |
-
1986
- 1986-10-23 JP JP61252603A patent/JPS63106698A/en active Pending
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JPH0446397A (en) * | 1990-06-14 | 1992-02-17 | Nec Corp | Continuous voice recognition system |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| EP0834862A3 (en) | Method of key-phrase detection and verification for flexible speech understanding | |
| US5995934A (en) | Method for recognizing alpha-numeric strings in a Chinese speech recognition system | |
| JPS6147440B2 (en) | ||
| KR19980070329A (en) | Method and system for speaker independent recognition of user defined phrases | |
| CN110942767A (en) | Recognition labeling and optimization method and device for ASR language model | |
| CN109859746B (en) | TTS-based voice recognition corpus generation method and system | |
| JPS63106698A (en) | Voice input kana-kanji converter | |
| Heeman et al. | Incorporating POS tagging into language modeling | |
| JPS6316766B2 (en) | ||
| JPH06131500A (en) | Character recognizing device | |
| JPH10116093A (en) | Voice recognition device | |
| JPS6219899A (en) | Sentence voice recognition equipment | |
| JPS60188995A (en) | Sentence enunciation | |
| JPS62180462A (en) | Voice input kana-kanji converter | |
| JP3009709B2 (en) | Japanese speech recognition method | |
| Gorin et al. | On improving speech recognition and keyword spotting with automatically generated morphological units | |
| JPS6344700A (en) | Word detection system | |
| JPS62164097A (en) | Voice discrimination system | |
| JPH01241671A (en) | Alphabet/kana converting system | |
| JP2000214881A (en) | Apparatus and method for sound recognition linguistic model generation | |
| JPS62166400A (en) | Voice wordprocessor | |
| JPS61189599A (en) | Symbol comparator | |
| JPS62197861A (en) | Voice input kana-kanji converter | |
| JPS62139070A (en) | Voice input kana-to-kanji converter | |
| JPH0614375B2 (en) | Character input device |