this function performs zca-whitening to a 2- or 3- dimensional image
ZCAwhiten(image, k, epsilon)a matrix, data frame or 3-dimensional array where the third dimension is equal to 3
an integer specifying the number of components to keep when svd is performed (reduced dimension representation of the data)
a float specifying the regularization parameter
a matrix or 3-dimensional array where the third dimension is equal to 3
Whitening (or sphering) is the preprocessing needed for some algorithms. If we are training on images, the raw input is redundant, since adjacent pixel values are highly correlated. When using whitening the features become less correlated and all features have the same variance.
http://ufldl.stanford.edu/wiki/index.php/Whitening
path = system.file("tmp_images", "1.png", package = "OpenImageR")
image = readImage(path)
res = ZCAwhiten(image, k = 20, epsilon = 0.1)