论文部分内容阅读
该文分析了人耳耳蜗及中枢听觉神经对声信号处理的机理，主要包括耳蜗对声信号的频率分析、柯蒂氏器官毛细胞的从振动机械能到电化学能的转换、听神经的侧抑制神经网络（ＬＩＮ）和增益调整四个层次，建立各个层次上听觉的数学模型。模型的最终输出即为用于识别的语音特征。辨识实验结果也表明此听觉模型特征能较好地表征语音本质。
This paper analyzes the mechanism of acoustic signal processing in cochlear and central auditory nerve of human being, including the frequency analysis of acoustic signals in cochlea, the conversion from mechanical vibration energy to electrochemical energy in Cortina’s organ cells, Network (LIN) and gain adjustment of the four levels, the establishment of mathematical models of auditory at all levels. The final output of the model is the speech feature used for recognition. The experimental results also show that the characteristics of the auditory model can better characterize the nature of the speech.