摘要
在科技快速发展的今天,人工智能技术日益成熟,包括机器人、语音识别、图像识别以及大师系统的等技术也在被不断地尝试使用在人们生活的各个方面。语音识别技术在现如今不管是PC端还是移动端都有很多应用,从苹果公司采用的Siri语音助手到今天微软的小娜语音助手,越来越证实语音识别技术在将来有很大的发展空间。基于Mel倒谱特征和RBF神经网络的语音识别改进算法结果表明,与现有的语音识别技术对比时,语音识别率有较大的提高,能够达到语音识别的改进的预期效果。
With the rapid development of science technology in nowadays, not only the artificial intelligence technology including robot, speech recognition, image recognition, and the master system technology gradually mature, but also be used in different aspects of people's lives.Speech recognition technology is already applied in PC and mobile client. Since Apple used speech recognition technology in Siri voice assistant, up until today Microsoft published the voice assistant which use the neural network technology, Cortana, it confirmed that speech recognition technology has enormous development space in the future. Speech recognition based on Mel frequency cepstral features and RBF neural network shows that compared to existing speech recognition technology, the accuracies of speech recognition has improved a lot, and it can reach the expected effect.