Research on Uyghur Broadcast News Continues Speech Sensitive-Word Spotting System

In this paper we introduce research and implementation of HMM based Uyghur Broadcast News Sensitive-word Spotting System uses MATLAB. Our system has three advantages described below: First, the speech corpus is very small because it has finite number sensitive-word. Second, the broadcast news word pronunciation is clear and the property of word speed is regular, so this is benefit to improve the recognition rate. Third, in our system, we used whole-word models as the basic speech unit, because whole words have the property that their acoustic representation is well defined and it is easer to segmentation of the beginning and the end of the word.