Initial evaluation of a preselection module for a flexible large vocabulary speech recognition system in telephone environment

We are improving a flexible, large-vocabulary, speaker-independent, isolated-word recognition system in a telephone environment, originally designed as an integrated system doing all the recognition process in one step. We have transformed it by adopting the hypothesis-verification paradigm. In this paper, we describe the architecture and results of the hypothesis subsystem. We show the system evolution and the modifications adopted to face such a difficult task, achieving significant improvements using automatically clustered phoneme-like units, semi-continuous HMMs and multiple models per unit. The system behavior for vocabulary-dependent and vocabulary-independent tasks and for vocabularies up to 10,000 words are tested.