T
Torn Ueda
Researcher at National Archives and Records Administration
Publications - 2
Citations - 9
Torn Ueda is an academic researcher from National Archives and Records Administration. The author has contributed to research in topics: Syllable & Speaker recognition. The author has an hindex of 1, co-authored 2 publications receiving 9 citations.
Papers
More filters
Proceedings ArticleDOI
Voice-activated word processor with automatic learning for dynamic optimization of syllable-templates
TL;DR: A learning method in which the syllable templates are automatically optimized, based on speaker-dependent recognition system, showed an average syllable recognition accuracy of 71.0% without and 82.5% with automatic learning.
Journal ArticleDOI
Voice-activated word processor with automatic learning for dynamic optimization of syllable-templates
TL;DR: A voice-activated word processor has been realized which provides text input by articulated, phrase-by-phrase Japanese speech, using an automatic learning algorithm to improve syllable recognition accuracy.