Modeling the Manifolds of Images of Handwritten
Digits.
Geoff Hinton   Peter Dayan   Mike Revow
IEEE Transactions on Neural Networks, 8, 65-74.
Abstract
This paper describes two new methods for modeling the manifolds of
digitized images of handwritten digits. The models allow a priori
information about the structure of the manifolds to be combined with
empirical data. Accurate modeling of the manifolds allows digits to be
discriminated using the relative probability densities under the
alternative models. One of the methods is grounded in principal components
analysis, the other in factor analysis. Both methods are based on locally
linear low-dimensional approximations to the underlying data manifold.
Links with other methods that model the manifold are discussed.
compressed postscript   pdf