Beth A. Carlson

Learn More
Speech recognizers trained with quiet wide-band speech degrade dramatically with high-pass, low-pass, and notch filtering, with noise, and with interruptions of the speech input. A new and simple approach to compensate for these degradations is presented which uses mel-filter-bank (MFB) magnitudes as input features and missing feature theory to dynamically(More)
Classical speaker and language recognition techniques can be applied to the classification of unknown utterances by computing the likelihoods of the utterances given a set of well trained target models. This paper addresses the problem of grouping unknown utterances when no information is available regarding the speaker or language classes or even the total(More)
  • Dewitt C Seward, Iv, Victor V Zue, Marc A Zissman, Lincoln Laboratory, Thesis Supervisor +10 others
  • 2007
Hidden Markov models are powerful tools for acoustic modeling in speech recognition systems. However, detailed analysis of their performance in specific experiments can be difficult. Two tools were developed and implemented for the purpose of analyzing hidden Markov model experiments: an interactive Viterbi backtrace viewer, and a multi-dimensional scaling(More)
  • 1