Estimating Spoken Words by Vowel and Consonant Identification Using Machine Lip Reading

DOI HANDLE

Bibliographic Information

Other Title
  • 機械読唇を用いた母音認識および子音認識による発話語推定

Abstract

In recent years, with the development of speech recognition technology, communication between people has become smoother. However, it is not as smooth in noisy environments and for people with speech disabilities. In this study, we propose a machine lipreading method that estimates the content of speech based on video information alone. Machine lipreading of Japanese generally involves vowel identification. Therefore, we attempt to perform consonant recognition in addition to vowel recognition, and compare these results with a word corpus to estimate the speech content.

Journal

Keywords

Details 詳細情報について

Report a problem

Back to top