论文信息 - Laterally Interconnected Self-Organizing Maps in Hand-Written Digit Recognition

Laterally Interconnected Self-Organizing Maps in Hand-Written Digit Recognition

An application of laterally interconnected self-organizing maps (LISSOM) to handwritten digit recognition is presented. The lateral connections learn the correlations of activity between units on the map. The resulting excitatory connections focus the activity into local patches and the inhibitory connections decorrelate redundant activity on the map. The map thus forms internal representations that are easy to recognize with e.g. a perceptron network. The recognition rate on a subset of NIST database 3 is 4.0% higher with LISSOM than with a regular Self-Organizing Map (SOM) as the front end, and 15.8% higher than recognition of raw input bitmaps directly. These results form a promising starting point for building pattern recognition systems with a LISSOM map as a front end.

[1] Isabelle Guyon,et al. Neural Network Recognizer for Hand-Written Zip Code Digits , 1988, NIPS.

[2] Martin J. Johnson,et al. Digital Realisation of Self-Organizing Maps , 1988, NIPS.

[3] James A. Pittman,et al. Recognizing Hand-Printed Letters and Digits , 1989, NIPS.

[4] Lawrence D. Jackel,et al. Handwritten Digit Recognition with a Back-Propagation Network , 1989, NIPS.

[5] Kunihiko Fukushima,et al. ALPHANUMERIC CHARACTER RECOGNITION BY THE NEOCOGNITRON , 1990 .

[6] Risto Miikkulainen,et al. Ocular Dominance and Patterned Lateral Connections in a Self-Organizing Model of the Primary Visual Cortex , 1994, NIPS.

[7] Risto Miikkulainen,et al. Topographic Receptive Fields and Patterned Lateral Interaction in a Self-Organizing Model of the Primary Visual Cortex , 1997, Neural Computation.