Pre-processing of envelope images for optical character recognition

Author

Downton, A.C. ; Leedham, C.G.

Author_Institution

Dept. of Electron. Syst. Eng., Essex Univ., UK

fYear

1988

Firstpage

27

Abstract

A feasibility study was undertaken to investigate methods of preprocessing envelope images to extract the address region from the image in the presence of the other data and presort the addresses into subclasses suitable for recognition by an optical character recognition system with separate recognition channels for machine and handwritten address classes. Preliminary results are presented that are based on trials using a sample of nearly 1000 envelope images. The data show a successful address extraction rate of around 98% and a correct classification rate of about 90%

Keywords

computerised picture processing; optical character recognition; postal services; address extraction; address region; character style classification; feature extraction; handwritten addresses; image preprocessing; machine-written addresses; optical character recognition; picture processing; presorting; sorting; Character recognition; Data engineering; Data mining; Handwriting recognition; Image recognition; Image segmentation; Optical character recognition software; Pixel; Statistics; Systems engineering and theory;

fLanguage

English

Publisher

ieee

Conference_Titel

Pattern Recognition, 1988., 9th International Conference on

Conference_Location

Rome

Print_ISBN

0-8186-0878-1

Type

conf

DOI

10.1109/ICPR.1988.28164

Filename

28164