THIS PAGE INCORPORATES CASCADING STYLE SHEET! TO VIEW PAGE CORRECTLY PLEASE UPDATE TO A CSS COMPLIANT BROWSER
 |
Natural Basis Functions and Topographic Memory for Face Recognition |
pdf 615 KB
ps.gz 1.36 MB
ps 3.64 MB
@Article{Rao:IJCAI95,
author = {Rajesh P.N. Rao and Dana H. Ballard},
title = {Natural Basis Functions and Topographic Memory for Face Recognition},
journal = {Proc International Joint Conf. on Artificial Intelligence},
year = {1995},
pages = {10 - 17},
abstract = { Recent work regarding the statistics of nat
ural images has revealed that the dominant
eigenvectors of arbitrary natural images closely
approximate various oriented derivativeof
Gaussian functions; these functions have also
been shown to provide the best fit to the recep
tive field profiles of cells in the primate striate
cortex. We propose a scheme for expression
invariant face recognition that employs a fixed
set of these ``natural'' basis functions to gener
ate multiscale iconic representations of human
faces. Using a fixed set of basis functions obvi
ates the need for recomputing eigenvectors (a
step that was necessary in some previous ap
proaches employing principal component anal
ysis (PCA) for recognition) while at the same
time retaining the redundancyreducing prop
erties of PCA. A face is represented by a set of
iconic representations automatically extracted
from an input image. The description thus ob
tained is stored in a topographicallyorganized
sparse distributed memory that is based on a
model of human longterm memory first pro
posed by Kanerva. We describe experimental
results for an implementation of the method
on a pipeline image processor that is capable
of achieving near realtime recognition by ex
ploiting the processor's framerate convolution
capability for indexing purposes. }
}