基于MIM模型的语音识别引擎SDSE及应用

Speech Recognition Engine SDSE Based on Mutual Information Matching and App lication

语音识别引擎SDSE提供一种通用的实时语音输入、分析和识别接口,通过与其它软件模块的连接和DDE数据通信实现人机语音通信交互操作,可应用于WINDOWS平台的PC和各类多媒体信息终端。该文介绍语音识别引擎SDSE的模块结构以及设计与实现方法,对SDSE的语音处理、互信息匹配模型MIM的运用进行了描述,同时也简要介绍了SDSE在语音识别浏览器VoiceIE和WORD文本输入中的具体应用。

Speech is u nd erstood as the most natural way for human to operate machine such as computer,P DA and other multimedia information terminals and the one of representation of i ntelligent computer is hearing perception.This paper describes a speech engine SDSE that is based on mutual information matching model MIM and can be called f or use by other software modules with embedded or support connection type and DD E data transfer.The application of mutual information matching algorithm MIM an d other speech processing such as feature detection and real-time speech input analysis are described in detail after introduction of system structure.At t he end,two applications of SDSE are presented,one is VoiceIE for Internet WWW browsing and another is VoiceWord for WORD text input.