Fusion of handwritten word classifiers

被引：115

作者：

Gader, PD

Mohamed, MA

Keller, JM

机构：

[1] Dept. of Comp. Eng. and Comp. Sci., University of Missouri-Columbia, Columbia

来源：

PATTERN RECOGNITION LETTERS | 1996年 / 17卷 / 06期

关键词：

handwriting recognition; fuzzy integral; Borda count; Sugeno measures; Choquet integral; Sugeno integral; classifier fusion; neural networks;

D O I：

10.1016/0167-8655(96)00021-9

中图分类号：

TP18 [人工智能理论];

学科分类号：

081104 ; 0812 ; 0835 ; 1405 ;

摘要：

Methods for fusing multiple handwritten word classifiers are compared on standard data. A novel method based on data-dependent densities in a Choquet fuzzy integral is shown to outperform neural networks, Borda and weighted Borda counts, and Sugeno fuzzy integral.

引用

页码：577 / 584

页数：8

共 18 条

[1] COMBINING MULTIPLE NEURAL NETWORKS BY FUZZY INTEGRAL FOR ROBUST CLASSIFICATION [J].