Title:
【発明の名称】類似文字識別方法
Document Type and Number:
Japanese Patent JP2963474
Kind Code:
B2
Abstract:
PURPOSE:To identify similar characters with high accuracy by comparing an upper blank ratio or a left blank ratio with a standard value prepared in advance concerning specified characters. CONSTITUTION:The picture of a document is read by a scanner and the segmentation processing of lines and characters is executed. In the case of the segmentation processing, the upper and lower sides of a circumscribed rectangle in a line character string are detected as upper and lower reference lines Lu and Ld and an interval between the lines is defined as a standard width W of the line. Concerning the respective characters, the interval between the upper reference line Lu and the upper end of each character is calculated as the upper blank. The blank ratio is calculated under the condition of blank ratio = upper blank/W and by comparing this value with the standard blank ratio concerting the characters prepared in advance, identification is executed between 'yo', KANA (Japanese syllabary) and 'yo' (small-sized, making a compound word) and between 'tsu (KANA)' and 'tsu' (small-sized, making a compound word). In the case of longitudinal writing, the processing is executing while paying attention to the left blank ratio.
More Like This:
Inventors:
MINEWAKI TAKAKUNI
Application Number:
JP25009189A
Publication Date:
October 18, 1999
Filing Date:
September 26, 1989
Export Citation:
Assignee:
RIKOO KK
International Classes:
G06K9/62; G06K9/68; G06K9/46; (IPC1-7): G06K9/62; G06K9/46; G06K9/68
Domestic Patent References:
JP62117090A | ||||
JP62187988A | ||||
JP62169285A | ||||
JP1108691A | ||||
JP63216189A | ||||
JP1171080A | ||||
JP285985A | ||||
JP60134394A | ||||
JP59167783A | ||||
JP59109979A | ||||
JP59231681A |
Attorney, Agent or Firm:
Makoto Suzuki