11-OtherLinearTechniques

problems of optical communications

说明：  Matlab交互式多模型UKF和EKF滤波程序（附说明文档）(Matlab interactive multi model UKF and EKF filter program (with explanatory document))

说明：  EOF为经验正交分解，实现数据的时空分离。是处理气象资料的经典程序(EOF for the empirical orthogonal decomposition, spatial and temporal separation of data. To deal with the classic procedures for meteorological data)

threshold for 2 adjacent images

Deblurring Images Using a Regularized Filter This example shows how to use regularized deconvolution to deblur images. Regularized deconvolution can be used effectively when constraints are applied on the recovered image (e.g., smoothness) and limited information is known about the additive noise. The blurred and noisy image is restored by a constrained least square restoration algorithm that uses a regularized filter.

说明：  基于HMM的说话人识别系统，包括提取MFCC特征，特征对比等。(A speaker recognition system based on HMM,included the extraction of MFCC features and features comparison)

基于MATLAB实现的说话人识别程序，分别用bp、pnn、som、rbf、lvq等算法，对语音文件进行训练和测试，效果不错。~..~ 下面说明一下bprengong程序： 数据分别用来训练和测试两部分。 具体程序分为两部分，第一部分为：计算识别模型 变量v是mfcc处理以后的矢量。因为数据可能长短不一，所以放在同一进行截取。p的每一行代表一个语音数据（共15个）。变量Pr为每一行的最大最小值。变量T为目标值。输出神经元个数为15。 在训练阶段，如果用于训练的输入训练样本的类别标号为i（即语音数据的标号），则训练时设第i个节点的期望输出设为1。其余节点期望输出均为0。 在识别时，当一个未知类别的样本作用到输入端时，考查各输出节点的输出，并将这个样本的类别判定为输出值最大的那个节点对应的类别。 (err)

通过高斯函数导数检测图像边缘，实质是方向可调小波变换检测图像边缘。(Derivative of Gaussian function through the image edge detection, in essence, Steerable wavelet transform image edge detection.)

提供了完整的增量式支持向量机的matlab原代码程序，便于初学者学习(Provides a complete incremental SVM matlab program source code, easy for beginners to learn)

说明：  有各种小波包分解的程序和子程序，能够有效快速地对振动信号进行分解。(There are various wavelet packet decomposition programs and subroutines, which can decompose vibration signals effectively and quickly.)