基于排列组合熵的语音端点检测技术研究

语音端点检测对于后续语音识别分析有着重要的意义，论文研究基于排列组合熵的语音端点检测技术。分析排列数和窗长对检测性能的影响，对含不同噪声类型及不同信噪比的中英文语音样本进行了端点检测实验。尤其在低信噪比情形下，检测方法要优于谱熵特征检测方法。该方法具有较优的稳健性和实时性等特点，为强背景噪声下的语音端点检测提供了新的研究途径。

Application of permutation entropy measure in detecting speech

As to sequential speech recognition,detecting the ending point of speech signals has shown its important role.This paper proposes a new speech detection method based on permutation entropy.Permutation and window size on detecting performance was analyzed.The comparison experiments of speech signals corresponding to different SNR and noise type was designed using the measure of complexity behaviors.Simulation results indicate that the method has better performance than frequency-entropy, especially in detecting low-SNR signal.The method has good robustness and strong real-time capabilities.It provides a new approach for detecting endpoints of weak signals under noisy environment.