Use of Distinctive Features for Automatic Speech Recognition

摘要

One of the most critical and yet unsolved problems in phonetic recognition is thetransformation of the continuous speech signal to a discrete representation for accessing words in the lexicon. In order to find an efficient description of speech for recognition tasks, our research investigates the use of distinctive features. Distinctive features are a small set of linguistic units which have the potential advantage of enabling us to describe contextual and coarticulatory variations in speech more parsimoniously and thus make more effective use of available training data. To access the usefulness of distinctive features, we focus our inquiry on three questions. First is there a particular spectral representation that will yield superior performance over others. Second, how would the extraction and use of acoustic attributes affect classification performance when compared to the direct use of the spectral representation. Finally, are there performance advantages in introducing an intermediate linguistic representation between the signal and the lexicon.