节点文献
一种基于多特征的带噪语音信号端点检测与音节分割算法
An Endpoint Detection and Syllable Separation Algorithm Based on Multi-characteristic for Noisy Speech Signal
【摘要】 语音信号的端点检测和音节分割直接决定语音识别率。在传统方法的基础上提取语音信号的多个特征参数,并综合利用各个参数的特性进行检测和分割,提高了端点检测和音节分割的准确度。
【Abstract】 The endpoint detection of speech signal is an important step in speech recognition. It determines directly the recognition rate. Several characteristic parameters of speech signal are extracted based on the traditional methods for endpoint detection and syllable separation, and the accuracy of endpoint detection and syllable separation is improved.
【关键词】 端点检测;
音节分割;
自相关函数;
【Key words】 endpoint detection; syllable separation; autocorrelation function;
【Key words】 endpoint detection; syllable separation; autocorrelation function;
【基金】 陕西省自然科学基础研究计划项目(2003F23)
- 【文献出处】 电声技术 ,Audio Engineering , 编辑部邮箱 ,2005年07期
- 【分类号】TN912.3
- 【被引频次】14
- 【下载频次】340