Simple feature extraction for handwritten character recognition

Author

Pedrazzi, P. ; Colla, A.M.

Author_Institution

Elsag Bailey-Finmeccanica S.p.A., Genova, Italy

Volume

3

fYear

1995

fDate

23-26 Oct 1995

Firstpage

320

Abstract

This paper deals with a simple and effective set of features for (handprinted) character representation in automatic reading systems. These features, computed within regularly placed windows spanning the character bitmap, consist of a combination of average pixel density and measures of local alignment along some directions. Patterns from different databases call be accommodated by choosing a variable window size. These features used in conjunction with a neural classifier (MLP) yielded a very high accuracy on several handprinted character databases, including NIST´s ones. Moreover they are easily implementable in VLSI, with throughputs as high as 250,000 characters/sec

Keywords

VLSI; document image processing; feature extraction; handwriting recognition; image classification; multilayer perceptrons; MLP; NIST; VLSI; automatic reading systems; average pixel density; character bitmap; feature extraction; handprinted character databases; handwritten character recognition; local alignment measures; neural classifier; variable window size; Character recognition; Counting circuits; Feature extraction; NIST; Neural networks; Noise robustness; Optical character recognition software; Throughput; Uncertainty; Very large scale integration;

fLanguage

English

Publisher

ieee

Conference_Titel

Image Processing, 1995. Proceedings., International Conference on

Conference_Location

Washington, DC

Print_ISBN

0-8186-7310-9

Type

conf

DOI

10.1109/ICIP.1995.537640

Filename

537640