JPS6335998B2 - - Google Patents
Info
- Publication number
- JPS6335998B2 JPS6335998B2 JP57178481A JP17848182A JPS6335998B2 JP S6335998 B2 JPS6335998 B2 JP S6335998B2 JP 57178481 A JP57178481 A JP 57178481A JP 17848182 A JP17848182 A JP 17848182A JP S6335998 B2 JPS6335998 B2 JP S6335998B2
- Authority
- JP
- Japan
- Prior art keywords
- word
- phoneme
- vowel
- devoiced
- vowels
- 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.)
- Expired
Links
- 239000011159 matrix material Substances 0.000 description 1
Description
（産業上の利用分野）
本発明は、入力音声に対し、まず音素認識を行
ない、この認識音素系列を音素表記された単語辞
書中の各単語と照合して単語を認識する単語音声
認識方法に関するものである。
（従来例の構成とその問題点）
従来のこの種の単語音声認識方法について第１
図とともに説明する。
第１図に示すように、入力単語音声を分析し、
この入力単語音声の特徴を抽出して入力単語音声
を構成する音素を認識し、この認識された音素系
列を音素表記された単語辞書中の各単語とコンフ
ユージヨンマトリツクス（以下、C.M.と略記す
る）を用いて照合し、尤度を計算し、尤度の大き
い単語を認識単語とするものである。ここで、第
１表は音素表記された単語辞書の一例を示してお
り、また、第２表は単語辞書の音素表記法の一例
を示している。
第１表に示したように、単語辞書における無声
化母音に関する音素表記は、日本語の性質によ
り、無声子音に挾まれた場合の母音は、無声化す
るとして、無声子音に挾まれた母音のみを無声化
母音として表記し、その他の場合の母音は有声母
音として表記していた。従来例では、音素認識段
階に
(Industrial Application Field) The present invention relates to a word speech recognition method that first performs phoneme recognition on input speech and then recognizes the word by comparing the recognized phoneme sequence with each word in a word dictionary in which phonemes are expressed. It is something. (Structure of conventional example and its problems) First, regarding this type of conventional word speech recognition method
This will be explained with figures. As shown in Figure 1, the input word speech is analyzed,
The features of this input word sound are extracted to recognize the phonemes that make up the input word sound, and this recognized phoneme sequence is combined with each word in the word dictionary in which the phoneme is expressed and a confusion matrix (hereinafter abbreviated as CM). ), the likelihood is calculated, and the word with the highest likelihood is selected as the recognized word. Here, Table 1 shows an example of a word dictionary in phoneme notation, and Table 2 shows an example of the phoneme notation method of the word dictionary. As shown in Table 1, due to the nature of the Japanese language, the phoneme representation of unvoiced vowels in word dictionaries is such that vowels that are sandwiched between voiceless consonants are devoiced, and only vowels that are sandwiched between voiceless consonants are devoiced. were written as unvoiced vowels, and vowels in other cases were written as voiced vowels. In the conventional example, at the phoneme recognition stage,
【表】【table】
【表】【table】
Claims (1)
系列を得、この認識音素系列と音素表記された単
語辞書中の各単語とを照合し尤度を計算して単語
を認識するに際し、前記認識音素系列の母音の無
声化修正のとき、無声子音が連続する場合および
無声子音と有声子音が続く場合とで挿入する無声
化母音の表記を区別するとともに、前記単語辞書
中の各単語の母音を表記する音素信号を、正常に
発声される有声母音、無声化する母音および無声
化し易い母音の３つに区分して表わすことを特徴
とする単語音声認識方法。1 Perform phoneme recognition on the input speech to obtain a recognized phoneme sequence, compare this recognized phoneme sequence with each word in the word dictionary in which the phoneme is expressed, calculate the likelihood, and recognize the word. When correcting the devoicing of vowels in a phoneme series, the notation of the devoiced vowel to be inserted is distinguished between cases where there are consecutive unvoiced consonants and cases where a voiceless consonant and a voiced consonant follow, and the vowels of each word in the word dictionary are A word speech recognition method characterized in that a phoneme signal to be written is divided into three categories: a normally uttered voiced vowel, a devoiced vowel, and a vowel that is easily devoiced.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| JP57178481A JPS5968795A (en) | 1982-10-13 | 1982-10-13 | Recognition of word voice |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| JP57178481A JPS5968795A (en) | 1982-10-13 | 1982-10-13 | Recognition of word voice |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| JPS5968795A JPS5968795A (en) | 1984-04-18 |
| JPS6335998B2 true JPS6335998B2 (en) | 1988-07-18 |
Family
ID=16049228
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| JP57178481A Granted JPS5968795A (en) | 1982-10-13 | 1982-10-13 | Recognition of word voice |
Country Status (1)
| Country | Link |
|---|---|
| JP (1) | JPS5968795A (en) |
-
1982
- 1982-10-13 JP JP57178481A patent/JPS5968795A/en active Granted
Also Published As
| Publication number | Publication date |
|---|---|
| JPS5968795A (en) | 1984-04-18 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US5333275A (en) | System and method for time aligning speech | |
| JP2004258658A (en) | Continuous speech recognition method using inter-word phoneme information and device thereforfor | |
| JPS62235998A (en) | Syllable identification system | |
| Greibus et al. | The phoneme set influence for Lithuanian speech commands recognition accuracy | |
| Hunt | Speaker adaptation for word‐based speech recognition systems | |
| JPS6335998B2 (en) | ||
| JP3621624B2 (en) | Foreign language learning apparatus, foreign language learning method and medium | |
| JP3378547B2 (en) | Voice recognition method and apparatus | |
| Downey et al. | A decision tree approach to task-independent speech recognition | |
| JPS59121400A (en) | Voice processor | |
| JPS61149997A (en) | Voice recognition equipment | |
| JPS6027433B2 (en) | Japanese information input device | |
| KR920009961B1 (en) | Unlimited korean language synthesis method and its circuit | |
| Pisarn et al. | Improving Thai spelling recognition with tone features | |
| Kalith | Ibralebbe Mohamed Kalith, David Asirvatham and Ismail Raisal | |
| JPS5837698A (en) | Conversion method for voice input japanese language typewriter | |
| JPH04127199A (en) | Japanese pronunciation determining method for foreign language word | |
| Kaur et al. | Automatic marking of Punjabi syllables boundaries in a sound file | |
| Arslan | A new universal language for speech recognition applications | |
| KR970050115A (en) | Speech Recognition Method of Variation Unit using Korean Variation Grouping Tree | |
| JPH08171396A (en) | Speech recognition device | |
| JPH0667685A (en) | Speech synthesizing device | |
| JPS60182499A (en) | Voice recognition equipment | |
| Yubune | Comprehension of patterns of phonetic simplification in English by non-native speakers: A cognitive account | |
| JPS6180298A (en) | Voice recognition equipment |