基于小波分析的语音识别的研究
|
摘要
语音识别技术的应用,本质上在于它能将输入的语音转化为语言代码,能够大幅度降低代码率,便于存储和传输,而且也容易被计算机或专用信息处理单元理解其含义,从而开发出更广泛的应用。例如,机器能听懂人类的自然语言。能够有效去除语音信号中的噪声是当今业界研究的热点问题,有很重要的理论价值和实用意义。
本论文研究被噪声干扰的语音信号的去噪和识别课题。首先对语音信号和噪声的特性进行了分析;接着对语音识别系统的预处理、语音信号分析方法、特征提取、模板训练和模板匹配方法进行了论述;语音识别率的提高需要提取准确的语音特征参数,最好的办法就是对待识别语音进行降噪处理。
本论文选取小波变换阈值去噪原理去除噪声。在对众多小波函数的分析中选择了sym8小波基和Heursure阈值选择规则,在‘sln'重调方法的前提下,分别采用硬阈值法、软阈值法和双变量阈值法,以及不同的小波分解层数进行了实验,得出采用双变量阈值法和5层尺度分解得到比较好的去噪效果和较小的信号损失的成果,对解决小波基选择和小波阈值选择的两个难点问题提供了一个可行的方法。
The application of speech recognition technology allows the input speech signal to be changed into speech code. With the technology, not only the data of the speech, which is transferred and storied in code mode, is less than that in original way, but the speech code is easier processed by computer or other information process unit. Therefore, the speech recognition technology can be applied in many fields, for example, a machine can understand out language. Efficient speeches de-noise which is a research focus in IT is meaningful for real world and has high theoretical value.
The theme is about de-noising of speech with noise and speech recognition. Firstly, the feature of speech signal and noise is introduced, and then the components of the speech recognition system, such as preprocessing, means of speech signal analysis, feature extraction, the training and the matching of speech template, are discussed. To increase the rate of speech recognition, the parameter of speech feature should be extracted accurately, signal de-noise is the best way to achieve the goal.
The 'sym8' wavelet and 'Heusure' threshold rule are chosen. Under the 'sln' readjustment method, hard, soft and double threshold are separately adopted in the experiments of different layer wavelet. The results of the experiment support 5 layers criterion decomposition with the double threshold, with which we can get good de-noise effect and reduce the lost of signal. And the study provide a effective method of wavelet and threshold selection
本论文研究被噪声干扰的语音信号的去噪和识别课题。首先对语音信号和噪声的特性进行了分析;接着对语音识别系统的预处理、语音信号分析方法、特征提取、模板训练和模板匹配方法进行了论述;语音识别率的提高需要提取准确的语音特征参数,最好的办法就是对待识别语音进行降噪处理。
本论文选取小波变换阈值去噪原理去除噪声。在对众多小波函数的分析中选择了sym8小波基和Heursure阈值选择规则,在‘sln'重调方法的前提下,分别采用硬阈值法、软阈值法和双变量阈值法,以及不同的小波分解层数进行了实验,得出采用双变量阈值法和5层尺度分解得到比较好的去噪效果和较小的信号损失的成果,对解决小波基选择和小波阈值选择的两个难点问题提供了一个可行的方法。
The application of speech recognition technology allows the input speech signal to be changed into speech code. With the technology, not only the data of the speech, which is transferred and storied in code mode, is less than that in original way, but the speech code is easier processed by computer or other information process unit. Therefore, the speech recognition technology can be applied in many fields, for example, a machine can understand out language. Efficient speeches de-noise which is a research focus in IT is meaningful for real world and has high theoretical value.
The theme is about de-noising of speech with noise and speech recognition. Firstly, the feature of speech signal and noise is introduced, and then the components of the speech recognition system, such as preprocessing, means of speech signal analysis, feature extraction, the training and the matching of speech template, are discussed. To increase the rate of speech recognition, the parameter of speech feature should be extracted accurately, signal de-noise is the best way to achieve the goal.
The 'sym8' wavelet and 'Heusure' threshold rule are chosen. Under the 'sln' readjustment method, hard, soft and double threshold are separately adopted in the experiments of different layer wavelet. The results of the experiment support 5 layers criterion decomposition with the double threshold, with which we can get good de-noise effect and reduce the lost of signal. And the study provide a effective method of wavelet and threshold selection
引文
1 飞思科技产品研发中心.MATLAB6.5辅助小波分析与应用.北京:电子科技出版社,2003:18-26
2 于鹏,徐义芳,曹志刚.基于加权特征值补偿的说话人识别.2002.8(6):513-S17
3 王一世.数字信号处理.北京:北京理工大学出版社,1997:134-146
4 李蕴华.将倒谱参数与基音信息有效结合进行说话人辩认.信号处理,2000.6(1)}5-89
5 刘刚,屈梁生.自适应阈值选择和小波消噪方法研究.信号处理,2002,18(6):509-512
6 江铭炎,郝宇.基于小波变换的语音增强去噪方法.山东大学学报(自然科学版),2001(2):201--209.
7 汤宝平等.基于平移不变的小波去噪方法及应用.重庆大学学报(自然科学版),2002,25(3):1-5
8 陈峰,成新民.基于小波变换的信号去噪技术及实现.现代电子技术,2005,(3):11-13
9 陈尚勤,罗承烈,杨雪.近代语音识别.四川:电子科技大学出版社,1991:135-137
10 李建平等.小波分析方法的应用.重庆:重庆大学出版社.1999.
11 沈亚强,金洪霞,刘旭.基于子波变换的语音去噪方法.信号处理,2000,16(3):221-226
12 沈亚强,低信噪比语音信号端点检测与自适应滤波,电子测量与仪器学报,第15卷,第(?)期,2001年3月
13 苏剑波,徐波.应用模式识别技术导论:人脸识别与语音识别.上海:上海交通大学出版社,2001:103-104 111-119
14 吴淑珍,冯成林.噪声环境下语音识别方法研究.北京大学学报(自然科学版),2001,37(3):365-370
15 吴是淑珍,吴阿华.说话人识别的参量研究和语音库建设.北京大学学报(自然科学版),1995.1(3):317-320
16 杨福生.小波变换的工程分析与应用.北京:科学出版社,1999:1-3,20-50
17 张军英.说话人识别的现代方法与技术.西安:西北大学出版社,1996:1-14,48-50}80-101
18 张贤达,现代信号处理.清华大学出版社,1995:125-136
19 张维强,徐晨.一种基于平移不变的小波阈值去噪算法.现代电子技术,2003,(6):29-31
20 屈丹,王炳锡.语言辨识的矢量量化方法(VQ).信息工程大学学报,2002,(3):54-57
21 胡昌华等.基于MATLAB的系统分析与设计小波分析.西安:西安电子科技大学出版 社,2000.
22 易克初,田斌,付强.语音信号处理.北京:国防工业出版社,2000:8-10 1-2
23 郑海波.李志远.基于小波包变换的一种降噪算法.合肥工业大学学报(自然版),2001,24(6):459-462
24 郑治真,沈萍.小波变换及其MATLAB工具的应用.北京:地震出版社,2001
25 赵红恰.基于小波变换阈值的信号去噪.现代雷达,2001,(2):37-39
26 赵力.语音信号处理.北京:电子工业出版社,2003:1-4
27 赵瑞珍等.基于小波变换系数区域相关性的阈值滤波算法.西安电子科技大学学报(自然科学报),2001,28(3):324-327
28 秦前清.实用小波分析.西安:西安电子科技大学出版社,1995:1-4.29-32
29 C.Charies,J.P.Rasson.Wavelet de-noising of Poisson-distn'buted data and applications.Computational Statistics.Data Analysis,2003,(43):139-148
30 Dai-fei Guo,Wei-Hong zhu,Zhen-Ming Guo and Jian-qiang Zhang.A study of wavelet Thresholding Denoisng.IEEE.Proceeding of ICSP2000,2000:329-332
31 Haitao Fang,Deshuang Huang.Noise reduction in lidar signal based on discrete wavelet transform.Optics Communications,2004,(233):67-76
32 Kanedera Noboru,Arai Takayuki,Hermansky Hynek,et al.On the Relative Importance of Various Components of the Modulation Spectrum for Automatic Speech Recognition.Speech Communication,1999,(28):43-55
33 L.R.Rabiner,Biing-Hwang Juang.Fundamentals of Speech Recognition.New Jersey,PTR Prentice Hall,1993
34 Lei Zhang,Paul Bao.Denoising by spatial correlation thresholding.IEEE Trans oncircuits and systems for video technology,2003,13(6):535-538
35 Pasti,L.Walczak,B.Massart,D.L.Reschiglian,P.Optimization of signal de-noising in discrete wavelet transform,Chemometrics and Intelligent Laboratory Systems,1999,48(1):21-34
36 Richard,P.Lippmann.Speech Recognition by Machines and Humans.Speech Communication,1997,(22):1-15
37 Silvia Baeehelli,Srena Papi.Filtered wavdet thresholding methods.Journal of Computation and Applied Mathematics.2004,(64-165):39-52
38 Taichiu Hsung,Daniel Pakkong Lun and K.C.Ho.Optimizing the multiwavelet shrinkage de-noising.IEEE Trans.on signal processing.2005,53(1):240-251
39 T.R.Downie and B.W.Silverman.The Discrete Multiple Wavelet Transform and Thresholding methods.IEEE Trans on signal processing,1998,46(9):2558-2561
40 Xichao Yin,Pu Han,Jun Zhang,Fengqi Zhang,Ningling Wang.Application of wavelet transform in signal denosing.Proceeding of the Second International Conference on Machine Learing and Cybernetics,2003,(11):436-441
41 Zhen Xianlin,Guo xiangsong,Wen xue.Two improved methods on wavelet image de-noising.Proceedings of the Second International Conference on Machine Learding and Cybernetics,2003,(11):2979-2983
2 于鹏,徐义芳,曹志刚.基于加权特征值补偿的说话人识别.2002.8(6):513-S17
3 王一世.数字信号处理.北京:北京理工大学出版社,1997:134-146
4 李蕴华.将倒谱参数与基音信息有效结合进行说话人辩认.信号处理,2000.6(1)}5-89
5 刘刚,屈梁生.自适应阈值选择和小波消噪方法研究.信号处理,2002,18(6):509-512
6 江铭炎,郝宇.基于小波变换的语音增强去噪方法.山东大学学报(自然科学版),2001(2):201--209.
7 汤宝平等.基于平移不变的小波去噪方法及应用.重庆大学学报(自然科学版),2002,25(3):1-5
8 陈峰,成新民.基于小波变换的信号去噪技术及实现.现代电子技术,2005,(3):11-13
9 陈尚勤,罗承烈,杨雪.近代语音识别.四川:电子科技大学出版社,1991:135-137
10 李建平等.小波分析方法的应用.重庆:重庆大学出版社.1999.
11 沈亚强,金洪霞,刘旭.基于子波变换的语音去噪方法.信号处理,2000,16(3):221-226
12 沈亚强,低信噪比语音信号端点检测与自适应滤波,电子测量与仪器学报,第15卷,第(?)期,2001年3月
13 苏剑波,徐波.应用模式识别技术导论:人脸识别与语音识别.上海:上海交通大学出版社,2001:103-104 111-119
14 吴淑珍,冯成林.噪声环境下语音识别方法研究.北京大学学报(自然科学版),2001,37(3):365-370
15 吴是淑珍,吴阿华.说话人识别的参量研究和语音库建设.北京大学学报(自然科学版),1995.1(3):317-320
16 杨福生.小波变换的工程分析与应用.北京:科学出版社,1999:1-3,20-50
17 张军英.说话人识别的现代方法与技术.西安:西北大学出版社,1996:1-14,48-50}80-101
18 张贤达,现代信号处理.清华大学出版社,1995:125-136
19 张维强,徐晨.一种基于平移不变的小波阈值去噪算法.现代电子技术,2003,(6):29-31
20 屈丹,王炳锡.语言辨识的矢量量化方法(VQ).信息工程大学学报,2002,(3):54-57
21 胡昌华等.基于MATLAB的系统分析与设计小波分析.西安:西安电子科技大学出版 社,2000.
22 易克初,田斌,付强.语音信号处理.北京:国防工业出版社,2000:8-10 1-2
23 郑海波.李志远.基于小波包变换的一种降噪算法.合肥工业大学学报(自然版),2001,24(6):459-462
24 郑治真,沈萍.小波变换及其MATLAB工具的应用.北京:地震出版社,2001
25 赵红恰.基于小波变换阈值的信号去噪.现代雷达,2001,(2):37-39
26 赵力.语音信号处理.北京:电子工业出版社,2003:1-4
27 赵瑞珍等.基于小波变换系数区域相关性的阈值滤波算法.西安电子科技大学学报(自然科学报),2001,28(3):324-327
28 秦前清.实用小波分析.西安:西安电子科技大学出版社,1995:1-4.29-32
29 C.Charies,J.P.Rasson.Wavelet de-noising of Poisson-distn'buted data and applications.Computational Statistics.Data Analysis,2003,(43):139-148
30 Dai-fei Guo,Wei-Hong zhu,Zhen-Ming Guo and Jian-qiang Zhang.A study of wavelet Thresholding Denoisng.IEEE.Proceeding of ICSP2000,2000:329-332
31 Haitao Fang,Deshuang Huang.Noise reduction in lidar signal based on discrete wavelet transform.Optics Communications,2004,(233):67-76
32 Kanedera Noboru,Arai Takayuki,Hermansky Hynek,et al.On the Relative Importance of Various Components of the Modulation Spectrum for Automatic Speech Recognition.Speech Communication,1999,(28):43-55
33 L.R.Rabiner,Biing-Hwang Juang.Fundamentals of Speech Recognition.New Jersey,PTR Prentice Hall,1993
34 Lei Zhang,Paul Bao.Denoising by spatial correlation thresholding.IEEE Trans oncircuits and systems for video technology,2003,13(6):535-538
35 Pasti,L.Walczak,B.Massart,D.L.Reschiglian,P.Optimization of signal de-noising in discrete wavelet transform,Chemometrics and Intelligent Laboratory Systems,1999,48(1):21-34
36 Richard,P.Lippmann.Speech Recognition by Machines and Humans.Speech Communication,1997,(22):1-15
37 Silvia Baeehelli,Srena Papi.Filtered wavdet thresholding methods.Journal of Computation and Applied Mathematics.2004,(64-165):39-52
38 Taichiu Hsung,Daniel Pakkong Lun and K.C.Ho.Optimizing the multiwavelet shrinkage de-noising.IEEE Trans.on signal processing.2005,53(1):240-251
39 T.R.Downie and B.W.Silverman.The Discrete Multiple Wavelet Transform and Thresholding methods.IEEE Trans on signal processing,1998,46(9):2558-2561
40 Xichao Yin,Pu Han,Jun Zhang,Fengqi Zhang,Ningling Wang.Application of wavelet transform in signal denosing.Proceeding of the Second International Conference on Machine Learing and Cybernetics,2003,(11):436-441
41 Zhen Xianlin,Guo xiangsong,Wen xue.Two improved methods on wavelet image de-noising.Proceedings of the Second International Conference on Machine Learding and Cybernetics,2003,(11):2979-2983