Chinese Speech Recognition System With Very Large Vocabulary

摘要

This paper described a speaker-dependent Chinese speech recognition system on the platform of SUN-SPARC 10 with very large vocabulary. This system was based on the SCHMM. Considering the special characteristics of Chinese, this paper creatively put forward a new revised beam search algorithm, named vowel_part_lookahead beam search algorithm, Experiment proved looking a vowel part ahead in an utterance could give valuable information about the optimality of the current partial path, with this new information a more robust and harder paining can be done.