基于概率神经网络的数字音识别

本文主要利用概率神经网络和动态时间规整技术来实现数字音的识别研究。结论是在利用概率神经网络进行语音识别时可以达到比较高的识别率，此外动态时间规整函数的加入，解决了神经网络的模板规整问题。作为语音识别技术的基础，其中包含了小波的基础理论，语音的预处理，DTW技术，端点检测等基础技术。对于神经网络的加入，更加有利于深入了解神经网络这一新兴技术。

The recognition of digital audio based on Probabilistic Neural Network

In this paper, the use of probabilistic neural networks and dynamic time warping technology to achieve the recognition of digital audio. Concluded that the use of probabilistic neural network for speech recognition can be achieved at relatively high recognition rate, in addition to dynamic time warping function of the addition of neural networks to solve the problem of the template structured. As the basis of voice recognition technology, which contains the basis of wavelet theory, the voice of the pretreatment, DTW technology, basic technology, such as endpoint detection. The addition of the neural network, more conducive to a deeper understanding of the neural network of new and emerging technologies.