基于高斯混合模型（GMM）的说话人识别系统

说话人识别系统,界面友好，可以对实时录音的说话人进行识别(Speaker recognition system, user-friendly, real-time recording can be carried out to identify the speaker)

说明：  基于统计模型的谱减法，程序复杂度不高，优于传统谱减法．(statistical model based on the spectral subtraction, the complexity of the procedure is not high and superior to the traditional spectral subtraction.)

最权威和最全面的语音识别的特征提取的 IEEE的期刊论文1980(The most authoritative and comprehensive speech recognition feature extraction of the IEEE Journal Papers 1980)

通常语音信号在增强时会出现混响现象，演讲者为了消除背景混响，不得不频繁地偏转头部的方向，这样会造成脉冲响应的不断改变。我们结合盲解卷法和频谱消去法来提高逆滤波器的滤波效果。我们利用输入语音信号间的相关系数矩阵计算出稳定、精确的室内脉冲响应的逆滤波器，而这些输入信号无需测量室内的脉冲响应就能被观测到。逆滤波能够消除早期的反射，这些反射包含混响中的绝大部分能量。之后，用频谱消去法来抑制逆滤波后的信号的尾部混响。本方法在实际适应性方面的表现通过具体的实验进行了验证，结果表明盲解卷法和频谱消去法的结合相较于单独使用一种方法，能够提供一个更优越的演说环境。(Usually voice signal will be enhanced when the reverberation phenomenon, the speaker in order to eliminate the background reverberation, and had to frequently deflect the direction of the head, this will cause the impulse response of changing. We combine the blind deconvolution and spectral elimination method to improve the filtering effect of the inverse filter. We use the correlation coefficient matrix between the input speech signal to calculate the stable, accurate inverse filter of the indoor impulse response, and these input signals do not need to measure the room impulse response can be observed. Inverse filtering to eliminate early reflections, these reflections contains most of the energy in the reverberation. After the spectrum elimination method to suppress the tail reverberation of the signal after the inverse filtering. Performance in the actual adaptation of the method by specific experimental validation results show that the blind deconvolution and spectrum to eliminate )

C语言fft和ifdct的函数，可以直接使用(C language functions fft and ifdct can be directly used)

matlab 实现的基于GMM的说话人识别系统(matlab implementation of the GMM-based speaker recognition system)

端点检测算法采用现阶段比较精确的时频方差检测法，模式识别和模板匹配采用的是改进后的DTW算法（主要是限定了DTW的搜索路径，进一步精确了DTW的平行四边形的形状，进一步的减少匹配中的参数存贮量以及多余的搜索路径）。(Endpoint detection algorithm uses more precise at this stage variance time-frequency detection, pattern recognition and template matching is based on the improved DTW algorithm (DTW is mainly limited to the search path, and further precision of the DTW parallelogram shape, and further reduction of matching the parameters in the storage volume as well as the extra search path).)

基于动态时间规整与隐马尔科夫模型的自动语音识别系统(Automatic speech recognition system based on dynamic time warping and hidden Markov model)

图像识别，可以识别出图片中的需要的水果。mtalab实现的。(Image recognition, the image can be identified in fruit required. mtalab achieved.)

1、文件夹中包含了经典功率谱估计和自适应均衡算法两个实验的所有程序。 2、R.m、LMS.m、LMSmain.m为自适应均衡算法的程序： R.m用来计算输入信号的自相关矩阵及其特征值； LMS.m为时域LMS算法，用统计的方法仿真得出不同信道参数和不同步长下的学习曲线； LMSmain.m为实验主程序，按照实验要求中的具体数据得到实验结果和曲线。 3、functionx.m、fzhouqitu.m、spectrum.m、bt.m、bart_lett.m、welch.m、SPECTRUMmain.m为经典谱估计的程序： functionx.m产生需要进行谱估计的函数； fzhouqitu.m用来计算信号周期图的函数； spectrum.m是用周期图法进行谱估计的函数 bt.m是用BT图法进行谱估计的函数 bart_lett.m是用BARTLETT法进行谱估计的函数 welch.m是用WELCH法进行谱估计的函数 SPECTRUMmain.m是主程序，按照实验要求中的具体数据得到实验结果和曲线。 (a document folder contains the classic power spectrum estimation and adaptive equalization algorithm for the two experiments all the procedures. 2, R.m, LMS.m, LMSmain.m adaptive equalization algorithm for the procedure : R.m used to calculate the input signal and the correlation matrix eigenvalue; LMS.m too LMS algorithm, use statistical simulation method come to a different channel parameters and the synchronous learning curve; LMSmain.m main program for the experiment, according to the experimental requirements of the specific data and the curve of the experimental results. 3, functionx.m, fzhouqitu.m, spectrum.m. bt.m, bart_lett.m, welch.m. SPECTRUMmain.m classical spectrum estimation procedures : functionx.m have a need for spectrum estimation function; fzhouqitu.m used to calculate t)