Handwritten Digit Recognition by a Mixture of Local PrincipalComponent Analysis

Authors:
Bailing Zhang;Minyue Fu;Hong Yan
Affiliations:
Department of Electrical and Computer Engineering, The University of Newcastle, NSW 2308, Australia;Department of Electrical and Computer Engineering, The University of Newcastle, NSW 2308, Australia;Department of Electrical Engineering, University of Sydney, NSW 2006, Australia Email: bailing@ee.usyd.edu.au
Venue:
Neural Processing Letters
Year:
1998

Citing 9
Cited 0

Self-organization and associative memory: 3rd edition

Self-organization and associative memory: 3rd edition
Original Contribution: Least mean square error reconstruction principle for self-organizing neural-nets

Neural Networks
Representation and separation of signals using nonlinear PCA type learning

Neural Networks
Self-organized formation of various invariant-feature filters in the adaptive-subspace SOM

Neural Computation
Dimension reduction by local principal component analysis

Neural Computation
Parameter extraction from population codes: A critical assessment

Neural Computation
Modeling the manifolds of images of handwritten digits

IEEE Transactions on Neural Networks
Decision-based neural networks with signal/image classification applications

IEEE Transactions on Neural Networks
Efficient classification for multiclass problems using modular neural networks

IEEE Transactions on Neural Networks

Quantified Score

Hi-index	0.00

Visualization

Abstract

Mixture of local principal component analysis (PCA) has attractedattention due to a number of benefits over global PCA. The performance ofa mixture model usually depends on the data partition and local linearfitting. In this paper, we propose a mixture model which has theproperties of optimal data partition and robust local fitting. Datapartition is realized by a soft competition algorithm called neural ’gas‘and robust local linear fitting is approached by a nonlinear extension ofPCA learning algorithm. Based on this mixture model, we describe a modularclassification scheme for handwritten digit recognition, in which eachmodule or network models the manifold of one of ten digit classes.Experiments demonstrate a very high recognition rate.