Assessment for automatic speech recognition: I Comparison of assessment methods
article
The use of three types of vocabularies (cockpit-control words, digits, and initial consonants) was compared for the assessment of five speech recognizers. The goal of this study is to compare various assessment methods from application oriented to carefully controlled laboratory situations. It was found that the discrimination between various recognizer (input) conditions is improved for more difficult vocabularies. Confusions between stimuli and responses of testwords can be used as a diagnostic tool for prediction of performance and developments.
Topics
TNO Identifier
7766
Source
Speech Communication, 12(3), pp. 241-246.
Pages
241-246
Files
To receive the publication files, please send an e-mail request to TNO Repository.