Diane C Simmons

US Patent:

20020085758, Jul 4, 2002

Filed:

Nov 21, 2001

Appl. No.:

09/991553

Inventors:

Mohammed Ayshi - Endicott NY, US
M. Kimmel - Endicott NY, US
Diane Simmons - Cortland NY, US

International Classification:

G06K009/48

US Classification:

382/199000

Abstract:

A character recognition system and method implement a classification system according to which each character of a set of characters is identifiable at least partially on the basis of a set of character feature types indicative of the identity of the character, the set of character feature types including (i) bars; (ii) lakes and (iii) bays. A character image is algorithmically scanned in each direction of a predetermined set of scan directions in order to extract data relating to bars, lakes and/or bays. A feature vector is assembled based on the algorithmically extracted data and rendered accessible to character recognition apparatus. The character recognition apparatus is adapted to recognize a character corresponding to an assembled feature vector at least partially on the basis of the quantity of each of (i) bars; (ii) lakes and (iii) bays indicated in the feature vector and to provide an output identifying the recognized character.