Code covered by the BSD License  

Highlights from
Active Shape Model (ASM) and Active Appearance Model (AAM)

image thumbnail

Active Shape Model (ASM) and Active Appearance Model (AAM)

by

 

16 Feb 2010 (Updated )

Cootes 2D/3D Active Shape & Appearance Model for automatic image object segmentation and recognition

[Evalues, Evectors, x_mean]=PCA(x)
function [Evalues, Evectors, x_mean]=PCA(x)
% PCA using Single Value Decomposition
% Obtaining mean vector, eigenvectors and eigenvalues
%
% [Evalues, Evectors, x_mean]=PCA(x);
%
% inputs,
%   X : M x N matrix with M the trainingvector length and N the number
%              of training data sets
%
% outputs,
%   Evalues : The eigen values of the data
%   Evector : The eigen vectors of the data
%   x_mean : The mean training vector
%
%
s=size(x,2);
% Calculate the mean 
x_mean=sum(x,2)/s;

% Substract the mean
x2=(x-repmat(x_mean,1,s))/ sqrt(s-1);

% Do the SVD 
%[U2,S2] = svds(x2,s); 
[U2,S2] = svd(x2,0);

Evalues=diag(S2).^2;
Evectors=bsxfun(@times,U2,sign(U2(1,:)));

Contact us