3_c_search_code_plus_2_FDTD_matlab_code

说明：  利用AMDF法求语音信号的基音周期（结果为时频图）(Pitch detection by AMDF algorithm)

介绍了一种新的滤波方法，并给出了具体应用对象和理论分析，可以为非线性滤波研究提高新的方向(Introduces a new filtering method, and gives the specific application objects and theoretical analysis, nonlinear filter can improve the new direction of)

通过sift算子，实现两幅图像的快速、精准匹配，效果较好(Sift operator, two images fast, accurate matching, better)

用于提取语音信号的能量谱的函数,ampp(s,a,b,c)，x为输入信号，a,b,c用于指定subplot函数的作图位置(A function for extracting the energy spectrum of the speech signal, ampp (s, a, b, c), x is the input signal, a, b, c is used to specify the location of subplot mapping function)

contraharmonic filter and anti_blurring

LFM信号的分数傅里叶变换分析，包含LFM信号分数阶谱图产生及FrFt子程序，及峰值搜索。非常好用。(LFM signal' s fractional Fourier transform, including the LFM signal generation and FrFt fractional spectrum of subroutines, and peak search. Very easy to use.)

cstering in mtlb.to cstetring in

相关向量机(RVM)的matlab源程序，包含快速算法，内含代码使用说明。 RVM采取是与支持向量机相同的函数形式稀疏概率模型，对未知函数进行预测或分类。 优点： (1) 不仅仅输出预测目标量的点估计值,还可以输出预测值的分布. (2) 使用更少数量的支持向量,从而显著减少输出目标量预测值的计算时间. (3) RVM不需要估计过多的参数. (4) RVM对是否满足Mercer 定理的核函数没有限制,适应性更好(Relevance Vector Machine (RVM) of the matlab source code, including fast algorithm that contains code for use. RVM support vector machine is taken the same functional form sparse probabilistic model to predict the unknown function or classification. Advantages: (1) is not only the amount of output predicted target point estimates, but also the distribution of predicted values ​ ​ can be output. (2) using a smaller number of support vectors, thereby significantly reducing the output target amount predicted value calculation time. (3) RVM does not require excessive parameter estimation. (4) RVM meets Mercer theorem on the kernel function is not limited, and better adaptability)

说明：  matlab下实现拉格朗日插值与三次样条插值，附实验报告(Lagrange interpolation matlab to achieve with cubic spline interpolation, with a lab report)

Particle Swarm Optimization