AHPS

Text localize and Detection

高斯过程小例子，提供三种不同协方差函数，选择不同协方差函数，设定样本之后，对高斯过程采样，画出3d图(Gaussian process small example, offers three different covariance function, choose a different covariance function, after setting the sample, the sampling of the Gaussian process, draw 3d Figure)

文章展示了基于高斯混合模型的语音频谱预测方法。频谱预测可能在传包过程中预防丢包这方面起到大作用。期望最大化算法用两倍或三倍的连续语音因素来测试模型。模型被用来设计第一，儿等指令预测量。预测表用频谱分配状态来估计并和一个简单的参考模型对比。最好的预测表得到一个平均频率扭曲值是0.46dB小于参考模型(This paper presents methods for speech spectrum prediction based on Gaussian mixture models. Spectrum prediction may be useful in a packet transmission system where the sensitivity to packet losses is a major problem. Models of speech are trained by the Expectation Maximization algorithm using pairs, triples etc. of consecutive cepstral vectors. The models are used to design first, second etc. order predictors. The prediction schemes are evaluated using the spectral distortion criterion and compared to a simple reference method. The best prediction scheme obtains an average spectral distortion that is 0.46 dB less than for the reference method. )

DBScan算法，在matlab中实现(DBScan algorithm, to achieve in matlab)

求取Multi-Dimensional Maximally Stable Extremal Regions，下载于美国加利福尼亚大学洛杉矶分校(strike Multi-Dimensional Ex Maximally Stable tremal Regions, downloads in the United States the University of California at Los Angeles)

DOA估计与快拍数的关系的MUSIC算法，MATLAB仿真源代码，可完成绘图。(DOA estimation snapshot relationship between the number of MUSIC algorithm, MATLAB simulation source code, to be completed by the drawing.)

例1 采用动量梯度下降算法训练 BP 网络。 例2 采用贝叶斯正则化算法提高 BP 网络的推广能力。在本例中，我们采用两种训练方法，即 L-M 优化算法（trainlm）和贝叶斯正则化算法（trainbr），用以训练 BP 网络，使其能够拟合某一附加有白噪声的正弦样本数据。(Example 1 uses the momentum gradient descent algorithm to train the BP network. Example 2 uses the Bayesian regularization algorithm to improve the generalization ability of BP network. In this example, we use two training methods, the LM optimization algorithm (trainlm) and the Bayesian regularization algorithm (trainbr), to train the BP network to fit a sine sample with white noise data. )

一共有15个m文件，80几个函数，5000多行代码。包含了一般软件的几大块：界面、参数、算法、仿真、演示等。熟悉GUI的朋友感兴趣的问题都有所涉及：比如参数传递，多个界面的操作，m文件构造复杂界面，存取文件，多个axes，读取图片，显示图片(A total of 15 m paper, 80 in several functions, more than 5000 lines of code. Contains several large pieces of software in general: the interface, parameters, algorithms, simulation, presentation and so on. Familiar with the GUI of the Friend of interest issues are involved: for example, transmission parameters, the operation of a number of interface, m file structure is complicated interface, access to documents, multiple axes, to read pictures, show the picture)

说明：  MATLAB在信号处理中的应用1 信号及其表示 信号的基本运算 信号的能量和功率 (MATLAB in signal processing of the application of a Signal and its representation Signal basic operations The signal energy and power )

SPEAKER RECOGNITION MELBANK