Utterance verification for spontaneous Mandarin speech keyword spotting

Utterance verification represents an important technology in the design of user-friendly speech recognition systems. In this paper we introduce four confidence measures of utterance verification for Mandarin speech keyword spotting. The spotting system adopts a two-stage strategy, with recognition followed by verification and the basic unit of the system is syllable. In the second stage, a keyword verification function with four different confidence measures is investigated. Experimental results show that utterance verification with the four confidence measures all can reduce false alarm rate and system performance of the third confidence measure is best