Statistical modeling for speech recognition
The demand of intelligent machines that may recognize the spoken speech and respond in a natural voice has been driving speech research. The challenging in speech recognition systems due to the language nature where there are no clear boundaries between words, the phonetic beginning and ending are...
| Main Authors: | , , , |
|---|---|
| Format: | Conference or Workshop Item |
| Language: | English |
| Published: |
2012
|
| Subjects: | |
| Online Access: | http://irep.iium.edu.my/24784/ http://irep.iium.edu.my/24784/1/2003C.pdf |
| Summary: | The demand of intelligent machines that may recognize the spoken speech and respond in a natural voice has been driving speech
research. The challenging in speech recognition systems due to the language nature where there are no clear boundaries between
words, the phonetic beginning and ending are influenced by neighbouring words, in addition to the variability in different
speakers speech: male or female, young or senior, loud or low speech, read or spontaneous, emotional or formal, fast or slow
speaking rate and the speech signal can be affected with environment noise. To avoid these difficulties the data driven statistical
approach based on large quantities of spoken data is used. The performance of speech recognition systems is still far worse than
that of humans. This is partly caused by the use of poor statistical models. In this paper, a comprehensive study of statistical
methods for speech and language processing are presented. The role of signal processing in creating a reliable feature set for the
recognizer and the role of statistical methods in enabling the recognizer to recognize the words of the spoken input sentence as
well as the meaning associated with the recognized word sequence were presented. |
|---|