论文部分内容阅读
本文介绍了一个可以脱离计算机的小词汇表语音识别系统，提出了一种新的神经网络结构并采用了模糊逻辑来实现该系统，使之能在噪声较大的实际环境中进行不定人的实时语音识别，识别率约９０％。
This paper introduces a small vocabularies speech recognition system that can be separated from the computer, puts forward a new neural network structure and uses fuzzy logic to realize the system so that it can perform random human real-time in the noisy real environment Speech recognition, recognition rate of about 90%.