[R] PCA on image data

[Monica Pisica]

Hi,

You have the values of the principal component in "scores" (for each "location" where you have a RGB reading) and the eigenvectors in "loadings", see 
?princomp

So your first pca component would be:

comp1 <- pca$scores[,1]

Now you can transform this in the matrix you need to display your image.

Hope this helps,

Monica


PS. I think prcomp() has the retx argument, not princomp() .... my answer pertains to princomp....

------------------------------------------------------------
Message: 14
Date: Thu, 3 Jul 2008 02:15:43 -0700 (PDT)
From: Bio7 
Subject: [R] PCA on image data
To: r-help at r-project.org
Message-ID: 
Content-Type: text/plain; charset=us-ascii


Dear R users,

i would like to apply a PCA on image data for data reduction. 
The image data is available as three matrices for the 
RGB values. At the moment i use

x <- data.frame(R,G,B)#convert image data to data frame
pca<-princomp(x,retx = TRUE)

This is working so far. 
>From this results then i want to create a new matrix
from the first (second..) principal component. Here i stuck.
So my question is how can i create this matrix from the results of the PCA.
(Maybe there is also a faster method available for PCA?)

Any answers or advices are appreciated

With kind regards

M.Austenfeld



_________________________________________________________________
Windows Live Hotmail is giving away Zunes. Enter for your chance to win.

M_Mobile_Zune_V3