Login| Sign Up| Help| Contact|

Patent Searching and Data


Title:
DICTIONARY DATA ORIGINATING SYSTEM IN CHARACTER RECOGNITION SYSTEM
Document Type and Number:
Japanese Patent JPS63282897
Kind Code:
A
Abstract:

PURPOSE: To simplify the origination of dictionary data capable of corresponding to various characters by dividing a segmented character into meshes and analyzing character background structure in accordance with the ratio of the characters in the divided meshes.

CONSTITUTION: A character is segmented and then a character frame is divided into several meshes less than 32 vertical and horizontal meshes and the meshes to be a background are noticed to analyze rough background structure. Since each mesh is expressed by one bit, a dictionary data can be expressed by the small number of bits corresponding to the number of divided meshes. At the time of forming dictionary data, the background structure of plural data is analyzed in each character (0W9, AWZ) based on real teaching, and every appearance of different background structure, new dictionary data are registered. Consequently, the system can be allowed to correspond to noises or changes of a character form (various character forms) and the dictionary origination based upon real teaching can be simplified.


Inventors:
NISHIDA YOSHIE
HATA SEIJI
YANO SOICHI
Application Number:
JP11687887A
Publication Date:
November 18, 1988
Filing Date:
May 15, 1987
Export Citation:
Click for automatic bibliography generation   Help
Assignee:
HITACHI LTD
HITACHI KEIYO ENG
International Classes:
G06K9/46; G06K9/62; G06K9/68; G06K9/52; (IPC1-7): G06K9/46; G06K9/68
Attorney, Agent or Firm:
Katsuo Ogawa