摘要
提出了一种基于双权值神经网络的非特定人连续语音识别的新算法。这种算法可以不经过端点检测和分割,构建连续语音中各不同音节的特征空间覆盖区,可以避免因分割错误而带来的错误识别。通过实验得到了较为满意的识别结果。
Two weight neural network is described. A new dynamic searching algorithm based on two weight neural network is presented. And then it was applied to recognize the continuous speech of speaker-independent. The recognition results can be searched dynamically without endpoint detecting and segmenting. Different feature- space covers are constructed according to different classes of syllables. Some satisfactory recognition results are ob- tained by experiment.
出处
《科学技术与工程》
2010年第19期4797-4800,共4页
Science Technology and Engineering
基金
浙江省自然科学基金项目(Y1090758)资助
关键词
双权值神经网络
连续语音
语音识别
two weight neural networks continuous speech speech recognition
