HMM

自己编写的用于传感器信号滤波的算法程序，并有真实信号处理的数据演示-I have written for the sensor signal filtering algorithms procedures and have the authentic signal processing of data presentation

利用杜宾算法求lpcc系数，非常有用-Durbin algorithm for the use of lpcc coefficient, very useful

用VC实现的语音库的建立的程序. 首先建立单音节语音库,然后将所有的单音节语音文件合成一个大的.wav文件,并且制定索引机制,便于快捷查找-VC realize voice with the establishment of library procedures. First of all, the establishment of monosyllabic speech database, and then all of the monosyllabic voice synthesis of a large document. Wav files, and to develop indexing mechanisms to facilitate the quick search

简洁适用的基音周期提取程序,采用自相关和平均幅度差函数相结合的方法-Simple application of the pitch period extraction procedures, the use of auto-correlation and the average magnitude difference function method of combining

FGHFGFHDFHFGHDFGHDFGHDFGHDFGHDFGHGHDFGHGH

解除tcp筛选的的算法和软件实现，用vc编写，请大家下载-lifting tcp screening algorithms and software used to prepare vc, please download

采用MELP编制的压缩与解码程序，在MATLAB环境下可直接运行，语音2.4kb/s-MELP (Mixed-Excitation Linear Predictive) Vocoder is a new 2400 bps Federal standard speech coder. This Matlab code of MELP can run directly in MATLAB condition.

这是一个字符识别程序，主要利用关系码和方向码来识别汉字或者英文字符的手写识别-This is a character recognition procedures, using mainly between code and code to identify the direction of Chinese or English characters handwriting recognition

首先对语音信号的预处理和特征提取问题进行了讨论,提取了三个有效的语音识别特征参数—LPC系数、LPC倒谱系数和Mel频率倒谱系数(MFCC);其次着熏讨论了语音识别中常用的两种识别方法—基于模板匹配的DTW和基于统计膜型的HMM在语音识别中的应用,分别讨论了各自的模型训练和识别的方 法,对HMM在实际应用中的一些具体问题也进行了讨论;最后对常规HMM提出了一些改进方法,包括显式状态驻留和随机分段模型(SM),重点讨论了SM在 语音识别中的训练和识别方法,并通过一个孤立数字语音识别系统验证了其比常规HMM更能描述语音信号的特征。

采样小波包分解语音信号，分解为3层，并求出分解系数-Speech signal using wavelet packet decomposition，decomposing 6 level，and attain decomposition coefficient