资源简介
slow feature analysis 慢特征分析matlab源代码,参考文献:Wiskott, L. and Sejnowski, T.J. (2002), "Slow Feature Analysis:Unsupervised Learning of Invariances",
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/
- 上一篇:matlab音频端点检测
- 下一篇:混合高斯模型的高阶统计量研究与仿真
评论
共有 条评论