摘要
设计了以DSP为核心的语音识别系统。该系统采用小波变换算法对噪声进行处理,提高了系统的语音识别性能。最后通过实验数据验证了该算法在提高语音识别系统性能应用中的有效性。该系统的研制为语音识别真正走向实用化提供了一种可行的方法。
The speech recognition system which made DSP as core was designed. In this system,the wavelet transform algorithm was adopted to deal with the noise, and the performance of speech recognition was improved. Finally,the test data proved that the algorithm was valid in improving the performance of speech recognition. The system＇ s development provided a feasible method for the practical utility of the speech recognition.
出处
《低压电器》
北大核心
2008年第14期12-15,58,共5页
Low Voltage Apparatus
基金
辽宁省教育厅项目(20060700)
关键词
小波变换
DSP
降噪
隐马尔可夫模型
信号处理
语音识别系统
wavelet transform
digital signal processor （DSP）
denoise
hidden markov model （HMM）
signal process
speech recognition system
