资源简介
该算法可用于信号处理的多方面,盲源分离,特征提取,模式识别等等
代码片段和文件信息
% This demo script shows how to perform SFA on user-defined function
% spaces. The user has to overwrite the functions EXPANSION and
% XP_DIM. In this example the two functions are located in the
% subdirectory sfa_tk/demo/expansion_demo/ . Further information can be
% found on the online documentation.
% init graphics
figure; clf; set(gcf ‘Position‘ [618 173 575 727]);
% !!! The following lines installs the user-defined EXPANSION and XP_DIM
% functions defined in the expansion_demo/ subdirectory.
% You might need to change this line if you don‘t start the script from
% the demo/ directory.
addpath expansion_demo/
% create the input signal
T = 5000;
t = linspace(0 2*pi T);
x1 = -sin(t)+2*cos(11*t).^4;
x2 = cos(11*t);
x = [x1; x2]‘;
% show the input signal
clf; subplot(311); plot(x);
title(‘input signals‘);
% linear SFA won‘t be able to recover the slowest source signal sin(t)
% since it is nonlinearly mixed
fprintf(‘\nLINEAR SFA\n\n‘);
y = sfa1(x);
subplot(312); plot(y(:1));
title(‘signal exctracted by linear SFA‘);
% the user-defined expansion function contains a to-the-fourth nonlinearity
% which allows SFA to reconstruct the slowest source signal sin(t).
% if the user-defined functions are really used appropriate messages
% should have been printed on the screen.
fprintf(‘\nEXPANDED SFA\n\n‘);
y = sfa2(x);
subplot(313); plot(y(:1));
title([‘signal exctracted by nonlinear SFA in a‘ ...
‘ user-defined function space‘]);
% !!! Remove the user-defined EXPANSION and XP_DIM functions defined in
% the expansion_demo/ subdirectory. You might need to change this line
% if you don‘t start the script from the demo/ directory
rmpath expansion_demo/
相关资源
- 自适应波束形成的经典算法大全
- 语音信号处理 预加重分帧加窗 端点检
- 时间序列的分析——模型的识别与预
- 电力系统课程设计
- 物体的运动轨迹预测扩展卡尔曼滤波
- 粒子群与遗传混合算法优化
- 各种聚类程序,包括生成聚类树、k
- ORL人脸库的PCA人脸识别,GUI界面
- RBF神经网络对iris分类
- UKF与EKF算法应用比较
- 动态矩阵控制算法dmc
- 演化博弈仿真
- 多方向的Gabor特征提取
- BP神经网络的非线性系统建模
- 时域特征的求取,例如峭度、裕度等
- PCA降维,有详细的注释,你值得拥有
- 数值分析实验报告-数值分析实验报告
- ROSL低秩分解算法,用于图像分析.ra
- 影像组学工具包
- 数学建模作业论文——有两种外形相
- 随机信号分析实验
- 混沌粒子群算法
- 图像数字图像课程设计--数字水印设计
- 这里实现了四种SVM工具箱的分类与回
- 最小二乘法模型辨识算法
- 经典功率谱估计与现代功率谱估计的
- 小波变换的特征提取
- 输入电压前馈Buck变换器的研究-输入电
- 投影追踪法
- 数字信号处理实验3用FFT对信号作频谱
评论
共有 条评论