Thread Subject:

princomp for dimension reduction

hii
I am doing a project on eye tracking.for training the classifier I have to look at 4 corners of the computer screen and take images..I have taken 30 images for each corner.so I have 120 images.I have to reduce the dimension of 120 images and project them into new coordinate system of 3 dimension.I kept each image in the colomn of a matrix so total colomns are 120 and rows are rows X colomns of each image.
I am using princomp function to do principal component analysis.
I dont know which output parameter out of coeff,score,latent gives me new corrdinates.

please help me..

thank you