ICA vs. PCA as Preprocessing for Appearance-based Object Recognition

M.A. Vicente and C. Fernández (Spain)


Computer vision, face recognition, PCA, ICA


The main goal of this paper is to show experimentally the equivalence between whitened PCA and ICA in the context of appearance-based object recognition: if no feature selection is carried out, ICA and whitened PCA perform equally in any data set provided a rotational invariant classifier is chosen.

Important Links:

Go Back