PCA: Performs principle component analysis for image files
In Osburg/eigenfaces: PCA on Eigenfaces

Description Usage Arguments Details Value References Examples

Calculates the data covariance matrix of the original data. Returns eigenvectors (and eigenvalues) of the covariance matrix as an 'imageset_ef' object.

1	PCA(td, showEigenvals = TRUE, quick = FALSE)

`td`	an object of class 'imageset_ef', training data.
`showEigenvals`	logical vector (TRUE or FALSE).
`quick`	logical vector (TRUE or FALSE).

td is the 'imageset_ef'-object where the images are saved as 'image_ef' objects. showEigenvals determins whether the eigenvalues are returned in addition to the eigenvectors (FALSE means only the eigenvectors are returned). When the number of pixels of each images is much bigger than the number of images in td it is faster to diagonalize t(A) %*% A instead of the covariance matrix. However, this causes that only a subset of the eigenvectors of the covariance matrix is returned. If quick is set TRUE, this option is activated.

list of length 1 (when showEigenvals <- FALSE; contains an 'imageset_ef' object consisting of the eigenvectors as 'image_ef' objects) or 2 (when showEigenvals <- TRUE; additionally contains a list of the eigenvalues).

Marinovsky F., Wagner P., Gesichtserkennung mit Eigenfaces, FH Zittau/Görlitz

# Import Olivetti-faces
olivetti <- system.file("extdata","olivetti_X.csv",package="eigenfaces")
td <- load_imageset_ef(olivetti, c(64,64))
PCA(td, showEigenvals = FALSE, quick=TRUE)