说明：  针对语音识别的实时性要求提出了一种新的基于支持向量机的OOV快速拒识算法并将其应用于基于DSP 实现的孤立词语音识别系统中。
(Speech Recognition real-time requirements of a new support vector machines based on the vocabulary rapid rejection algorithm which can be used for the DSP-based Isolated Speech Recognition System.)