基于DSR系统的嵌入式语音拨号系统设计  

The Design of Embedded Voice Dialing System Based on DSR Mode

作　　者：肖薇[1] 李奇[1] 冯南[2] 

机构地区：[1]东南大学自动控制系,南京210096 [2]Motorola南京软件中心,南京211102

出　　处：《科学技术与工程》2004年第12期1016-1018,1023,共4页Science Technology and Engineering

摘　　要：采用分布式语音识别(DSR)模式来设计和实现手机的语音拨号系统:在手机端进行特征参数的提取,然后通过无线信道传输到服务器端进行识别与训练,最后把结果送回到手机端。通过试验发现,该语音拨号系统识别率在90％以上,并且大大提高了语音质量和识别的速度。The Distributed Speech Recognition (DSR) to design and implement the voice dialing system of the mobile phone are introduced. The feature parameters are extracted on the site of client (phone), then transmitted over a channel to the server to train and recognize, and the results are sent back to the client (phone). Through the experiment, the recognition rate of this voice dialing system is above 90%. The quality of the voice and the speed of the recognition process are improved greatly.

关 键 词：分布式语音识别(DSR) 嵌入式 语音拨号系统 特征提取 

分 类 号：TN912.34[电子电信—通信与信息系统]