Independent components of face images: A representation for face recognition

Marian Stewart Bartlett and Terrence J. Sejnowski

Proceedings of the 4th Annual Jount Symposium on Neural Computation, Pasadena, CA, May 17, 1997. Proceedings can be obtained from the Institute for Neural Computation, UCSD 0523, La Jolla, CA 92093.

Abstract

Methods for obtaining representations of face images based on independent component analysis (ICA) are presented. A global ICA representation is compared to a global representation based on principal component analysis (PCA) for recognizing faces across changes in lighting and changes in pose. For each set of face images, a set of statistically independent source images was found through an unsupervised learning algorithm that maximized the mutual information between the input and the output of a nonlinear transformation (Bell \& Sejnowski, 1995). These source images comprised the kernels for the representation. The independent component kernels gave superior class discriminability to the principal component kernels. Recognition across changes in pose with the ICA representation was 93\%, compared to 87\% with a PCA representation, and across changes in lighting ICA gave 100\% correct recognition, compared to 90\% with PCA.