Fusion of handwritten numeral classifiers based on fuzzy and genetic algorithms

Fuzzy and genetic algorithms are used to develop an approach for fusioning multiple handwritten numeral classifiers. A computational scheme of the Choquet (fuzzy) integral serves as a data fusion tool, whereas genetic algorithms are implemented to optimize the derivation of fuzzy densities which play a very important role for the calculation of fuzzy measures and fuzzy integrals. Several experimental results are provided to illustrate the effectiveness of this methodology.