An introduction to genetic algorithms
An introduction to genetic algorithms
Comparison of normalization methods for character recognition
ICDAR '95 Proceedings of the Third International Conference on Document Analysis and Recognition (Volume 2) - Volume 2
Pattern Classification (2nd Edition)
Pattern Classification (2nd Edition)
Signature Verification Using a Bayesian Approach
IWCF '08 Proceedings of the 2nd international workshop on Computational Forensics
Zoning methods for handwritten character recognition: A survey
Pattern Recognition
| Hi-index | 0.00 |
Classifying an unknown input is a fundamental problemin pattern recognition. A common method is to define a distancemetric between patterns and find the most similar patternin the reference set. When patterns are in binary featurevector form, there have been two approaches to improvethe performance over the equal-weighted Hamming distancemetric. One is to give different weights to different featuresusing an optimization technique, and the other is to use asimilarity measure that gives full credit to features present inboth patterns and the less credit to those absent from bothpatterns. Both approaches have been reported to performbetter than the naïve Hamming distance approach. In thispaper, we propose to combine these two approaches using agenetic algorithm to optimize weights. Experimental resultsshow that this method is superior to conventional measuresin an OCR application.