搜索资源列表
kiuqie
- 有PMUSIC 校正前和校正后的比较,AHP层次分析法计算判断矩阵的最大特征值,完整的基于HMM的语音识别系统。- A relatively before correction and after correction PMUSIC, Calculate the maximum eigenvalue judgment matrix of AHP, Complete HMM-based speech recognition system.
mangfun
- 完整的基于HMM的语音识别系统,有PMUSIC 校正前和校正后的比较,未来线路预测,分析误差。- Complete HMM-based speech recognition system, A relatively before correction and after correction PMUSIC, Future line prediction, error analysis.
fengning
- 完整的基于HMM的语音识别系统,比较了软阈值,硬阈值及当今各种阈值计算方法,包含特征值与特征向量的提取、训练样本以及最后的识别。- Complete HMM-based speech recognition system, Comparison of soft threshold and hard threshold and today various threshold calculation method, Contains the eigenvalue and eigenvector ex
CNN
- 用matlab实现的BP神经网络,比较基础,对已有的语音信号进行分类,留了500个作为检验-Using matlab to achieve the BP neural network basis of comparison, the existing classification of the speech signal, leaving a 500 as a test
1
- 计算语音信号的归一化自相关函数,学习归一化自相关函数基因周期检测算法,并用实验给出结果并与原方法比较。-Calculating a speech signal normalized autocorrelation function, learning normalized autocorrelation function of genes cycle detection algorithm, and experimental results are given and compared with
MATLAB
- 语音信号经过低通,高通,带通,带阻四个滤波器,比较滤波前后的功率谱,以及去噪能力-Voice signal through the low-pass, high-pass, band-pass, four band-elimination filter, compared before and after the power spectrum of the filter, and the denoising ability
bss-fullrank
- 最近提出来的采用高斯混合模型进行盲源分离的工具,比较有效,产生的结果可以应用于语音识别上-Blind source separation using GMM
ningqen_v33
- 比较了软阈值,硬阈值及当今各种阈值计算方法,采用的是脉冲对消法,完整的基于HMM的语音识别系统。- Comparison of soft threshold and hard threshold and today various threshold calculation method, It uses a pulse of consumer law, Complete HMM-based speech recognition system.
pantou_v50
- 有PMUSIC 校正前和校正后的比较,脉冲响应的相关分析算法并检验,完整的基于HMM的语音识别系统。- A relatively before correction and after correction PMUSIC, Related impulse response analysis algorithm and inspection, Complete HMM-based speech recognition system.
qaijiu
- 有PMUSIC 校正前和校正后的比较,语音信号的采集与处理,数字信号处理课设,是学习PCA特征提取的很好的学习资料。- A relatively before correction and after correction PMUSIC, Acquisition and Processing of the speech signal, digital signal processing class-based, Is a good learning materials to learn PCA
dereverberation
- 去混响的一种方法,使用两段法去混响。内附有语音文件方便比较。-a kind of dereverberation method
谱减法
- 谱减法是一种比较简单的语音降噪的方法,在语音信号具有较高的信噪比的情况下具有良好的降噪功能。(Spectral Subtraction)
DTW
- 动态时间规整用于语音识别对齐和对比,例子是librosa 提取MFCC 特征,用DTW比较识别(Dynamic time warping is used for speech recognition, alignment, and contrast. The example is librosa, which extracts MFCC features and compares them with DTW)
DTW
- Dynamic Time Warping(DTW)诞生有一定的历史了(日本学者Itakura提出),它出现的目的也比较单纯,是一种衡量两个长度不同的时间序列的相似度的方法。应用也比较广,主要是在模板匹配中,比如说用在孤立词语音识别(识别两段语音是否表示同一个单词),手势识别,数据挖掘和信息检索等中。(The birth of Dynamic Time Warping (DTW) has a certain history (Itakura, a Japanese scholar), and it
pujian
- 运用谱减法,对语音信号进行去噪处理并比较处理前后的信噪比(Using spectral subtraction to denoise the speech signal and compare the signal to noise ratio before and after processing)
EMD
- 语音增强 提供io接口及处理的工具箱等,语音增强效果比较好(Speech enhancement provide the IO interface and the processing toolbox, and so on, the speech enhancement effect is better)