We present a graph-based method for rapid, accurate search through prototypes for transformation-invariant pattern classification. Our method has in theory the same recognition accuracy as other recent methods based on "tangent distance" [Simard et al., 1994], since it uses the same categorization rule. Nevertheless ours is significantly faster during classification because far fewer tangent distances need be computed. Crucial to the success of our system are 1) a novel graph architecture in which transformation constraints and geometric relationships among prototypes are encoded during learning, and 2) an improved graph search criterion, used during classification. These architectural insights are applicable to a wide range of problem domains. Here we demonstrate that on a handwriting recognition task, a basic implementation of our system requires less than half the computation of the Euclidean sorting method.
[1]
Patrice Y. Simard,et al.
Learning Prototype Models for Tangent Distance
,
1994,
NIPS.
[2]
Yann LeCun,et al.
Efficient Pattern Recognition Using a New Transformation Distance
,
1992,
NIPS.
[3]
Patrice Y. Simard.
Efficient Computation of Complex Distance Metrics Using Hierarchical Filtering
,
1993,
NIPS.
[4]
Y. Le Cun,et al.
Comparing different neural network architectures for classifying handwritten digits
,
1989,
International 1989 Joint Conference on Neural Networks.
[5]
Yann LeCun,et al.
Tangent Prop - A Formalism for Specifying Selected Invariances in an Adaptive Network
,
1991,
NIPS.