英语语音优化识别建模仿真分析
|
摘要
随着中国经济高速发展以及全球一体化的进程,英语成为了人们日常交流必不可少的工具,然而对于初学者来说,能够通过语音识别技术将语音信号转化成文本的格式,更有利于快速掌握英语。而且语音识别技术经过多年的发展依然具有巨大的挖掘潜力,面对移动互联网的快速发展,通过对实时通信工具的大数据量的需求爆发,英语语音识别的实时性和系统稳定性越来越受到关注,文中分析了常用的传统语音识别技术,例如动态时间规整、神经网络模型和隐马尔可夫模型等,运用隐马尔可夫模型对语音信号进行处理和识别,提取出特征参数,与经过训练的模型体系进行匹配,找出最优的识别序列。然后在PC平台上,利用MATLAB建模仿真,基本实现了英语语音短句的识别,对于后续的硬件产品实现打下了良好的基础,具有积极的现实意义。
With the rapid development of China's economy and the process of global integration,English has become an indispensable tool for people's daily communication. However,for beginners,voice signals can be converted into text formats through speech recognition technology,which is more conducive to the rapid mastery of English. And speech recognition technology still has tremendous potential after many years of development. Faced with the rapid development of mobile Internet,the real-time and system stability of English speech recognition have been paid more and more attention through the explosion of large data demand for real-time communication tools. This paper analyses the commonly used traditional speech recognition technologies,such as dynamic time regulation,neural network model and hidden Markov model,etc. l,and uses the hidden Markov model to process and recognize the speech signal,extract the characteristic parameters,match with the trained model system,and find out the optimal recognition sequence. Then on PC platform,using MATLAB modeling and simulation,the recognition of English speech short sentences is basically realized,which lays a good foundation for the realization of the hardware products behind,and has positive practical significance.
With the rapid development of China's economy and the process of global integration,English has become an indispensable tool for people's daily communication. However,for beginners,voice signals can be converted into text formats through speech recognition technology,which is more conducive to the rapid mastery of English. And speech recognition technology still has tremendous potential after many years of development. Faced with the rapid development of mobile Internet,the real-time and system stability of English speech recognition have been paid more and more attention through the explosion of large data demand for real-time communication tools. This paper analyses the commonly used traditional speech recognition technologies,such as dynamic time regulation,neural network model and hidden Markov model,etc. l,and uses the hidden Markov model to process and recognize the speech signal,extract the characteristic parameters,match with the trained model system,and find out the optimal recognition sequence. Then on PC platform,using MATLAB modeling and simulation,the recognition of English speech short sentences is basically realized,which lays a good foundation for the realization of the hardware products behind,and has positive practical significance.
引文
[1]周晓武.嵌入式实时英语语音识别系统的设计与研究[J].电子设计工程,2017,25(8):187-190.
[2]赵丹,钟楠.在线连续交互式英语语音智能识别系统设计[J].现代电子技术,2017,40(15):137-140.
[3]张瑞.英语语音合理性优化识别建模仿真研究[J].计算机仿真,2017,34(2):289-292.
[4]秦楚雄,张连海.基于DNN的低资源语音识别特征提取技术[J].自动化学报,2017,43(7):1208-1219.
[5]王建荣,高永春,张句,等.基于Kinect辅助的机器人带噪语音识别[J].清华大学学报:自然科学版,2017,57(9):921-925.
[6]朱荣,李小映.基于Kinect的语音识别技术研究[J].计算机与数字工程,2017,45(6):1211-1215.
[7]周扬,孙玲玲,马德.基于HMM模型的语音识别系统的研究[J].物联网技术,2017,7(10):74-76.
[8]刘蕾.智能播种机嵌入式英语语音识别控制系统设计与研究[J].农机化研究,2018,40(12):240-244.
[9]雷海武,刘任任,刘新.一种基于DTW的动态笔迹识别算法[J].计算技术与自动化,2018,37(1):83-88.
[10]易江燕,陶建华,刘斌,等.基于迁移学习的噪声鲁棒语音识别声学建模[J].清华大学学报:自然科学版,2018,58(1):55-60.
[11]冯志伟,詹宏伟.会话智能代理与语音自动识别[J].外语学刊,2018(1):13-23.
[12]王旭东,王冬霞,周城旭.基于改进BFDNN的远距离语音识别方法[J].电脑知识与技术,2018,14(15):182-185,191.
[13]陈伟,田一明,王喜太,等.基于PSO-BP神经网络的语音识别研究[J].电脑知识与技术,2018,14(1):187-188.
[14]孙彦楠,夏秀渝.基于深度神经网络的关键词识别系统[J].计算机系统应用,2018,27(5):41-48.
[15]王兴刚.英文发音中错误语音自动识别系统设计[J].现代电子技术,2018,41(10):179-182.
[16]陈理,杨雪梅,毕春艳,等.基于子波方差和尺度波长关系的语音识别算法研究[J].现代电子技术,2018,41(10):169-171.
[17]王海坤,潘嘉,刘聪.语音识别技术的研究进展与展望[J].电信科学,2018,34(2):1-11.
[18]程风,翟超,吕志,等.基于语音识别技术的智能家居主控设计[J].工业控制计算机,2018,31(5):29-31.
[19]陈树,于海波.一种改进的特征提取方法在语音识别中的应用[J].传感器与微系统,2018,37(5):154-157.
[2]赵丹,钟楠.在线连续交互式英语语音智能识别系统设计[J].现代电子技术,2017,40(15):137-140.
[3]张瑞.英语语音合理性优化识别建模仿真研究[J].计算机仿真,2017,34(2):289-292.
[4]秦楚雄,张连海.基于DNN的低资源语音识别特征提取技术[J].自动化学报,2017,43(7):1208-1219.
[5]王建荣,高永春,张句,等.基于Kinect辅助的机器人带噪语音识别[J].清华大学学报:自然科学版,2017,57(9):921-925.
[6]朱荣,李小映.基于Kinect的语音识别技术研究[J].计算机与数字工程,2017,45(6):1211-1215.
[7]周扬,孙玲玲,马德.基于HMM模型的语音识别系统的研究[J].物联网技术,2017,7(10):74-76.
[8]刘蕾.智能播种机嵌入式英语语音识别控制系统设计与研究[J].农机化研究,2018,40(12):240-244.
[9]雷海武,刘任任,刘新.一种基于DTW的动态笔迹识别算法[J].计算技术与自动化,2018,37(1):83-88.
[10]易江燕,陶建华,刘斌,等.基于迁移学习的噪声鲁棒语音识别声学建模[J].清华大学学报:自然科学版,2018,58(1):55-60.
[11]冯志伟,詹宏伟.会话智能代理与语音自动识别[J].外语学刊,2018(1):13-23.
[12]王旭东,王冬霞,周城旭.基于改进BFDNN的远距离语音识别方法[J].电脑知识与技术,2018,14(15):182-185,191.
[13]陈伟,田一明,王喜太,等.基于PSO-BP神经网络的语音识别研究[J].电脑知识与技术,2018,14(1):187-188.
[14]孙彦楠,夏秀渝.基于深度神经网络的关键词识别系统[J].计算机系统应用,2018,27(5):41-48.
[15]王兴刚.英文发音中错误语音自动识别系统设计[J].现代电子技术,2018,41(10):179-182.
[16]陈理,杨雪梅,毕春艳,等.基于子波方差和尺度波长关系的语音识别算法研究[J].现代电子技术,2018,41(10):169-171.
[17]王海坤,潘嘉,刘聪.语音识别技术的研究进展与展望[J].电信科学,2018,34(2):1-11.
[18]程风,翟超,吕志,等.基于语音识别技术的智能家居主控设计[J].工业控制计算机,2018,31(5):29-31.
[19]陈树,于海波.一种改进的特征提取方法在语音识别中的应用[J].传感器与微系统,2018,37(5):154-157.