GENETIC ALGORITHM FOR CNN TEMPLATE LEARNING

Authors
KOZEK T ROSKA T CHUA LO
Citation
T. Kozek et al., GENETIC ALGORITHM FOR CNN TEMPLATE LEARNING, IEEE transactions on circuits and systems. 1, Fundamental theory andapplications, 40(6), 1993, pp. 392-402
Citations number
15
Categorie Soggetti
Engineering, Eletrical & Electronic
Journal title
IEEE transactions on circuits and systems. 1, Fundamental theory andapplications → ACNP
ISSN journal
10577122
Volume
40
Issue
6
Year of publication
1993
Pages
392 - 402
Database
ISI
SICI code
1057-7122(1993)40:6<392:GAFCTL>2.0.ZU;2-J
Abstract
A new learning algorithm for space invariant cellular neural networks (CNN's) is described. Learning is formulated as an optimization proble m. Exploration of any specified domain of stable CNN's is possible by the current approach. Templates are derived using a genetic optimizati on algorithm. Details of the algorithm are discussed and several appli cation results are shown. Using this algorithm propagation-type and gr ay-scale-output CNN's can also be designed.