Hand-written character recognition apparatus

• US 5,526,440 A
• Filed: 11/28/1990
• Issued: 06/11/1996
• Est. Priority Date: 11/30/1989
• Status: Expired due to Fees

• Current Assignee: Sony Corporation (Sony Group Corp.)
• Original Assignee: Sony Corporation (Sony Group Corp.)
• Inventors: Chatani, Masayuki, Yoshida, Kimiyoshi, Tamori, Hirofumi, Sakano, Akio
• Primary Examiner(s): COUSO, YON JUNG
• Application Number:
  US07/619,212
• Time in Patent Office: 2,022 Days
• Field of Search: 382/13, 382/24, 382/10, 382/185, 382/187, 382/202, 382/190, 382/203, 382/204
• US Class Current: 382/202
• CPC Class Codes: G06V 30/196 using sequential comparisons of the image signals with a plurality of references
• CPC Class Codes: G06V 30/36 Matching; Classification
• CPC Class Codes: G06V 30/373 using a special pattern or subpattern alphabet

Abstract

A hand-written character recognition apparatus which checks the relation between the rectangular outline directions of two partial structures of one character.

Claims

1. A hand-written character recognition apparatus comprising:

input means for generating hand-written stroke data;

means for deriving and storing characteristics of the hand-written stroke data;

buffer means for storing rectangular outline and direction data of the hand-written stroke data;

character dictionary means for storing characteristics of strokes of characters and directional relations of rectangular outline data of characters, including rectangular outline and direction data for partial structures of characters; and

character recognition means, including partial structure determining means which determines partial structures of the hand-written stroke data, for determining characters of the hand-written stroke data by comparing characteristics and the directional relations of the rectangular outline data stored in the character dictionary means with respect to the direction data of partial-structures of the hand-written strokes stored in the buffer means.

2. A hand-written character recognition apparatus comprising:

input means for generating hand-written stroke data;

means for deriving and storing characteristics of the hand-written stroke data;

buffer means for storing rectangular outline and direction data of the hand-written stroke data, wherein the rectangular outline data stored in the buffer means includes coordinate data for the vertices and center points of the sides of a hypothetical rectangle surrounding each partial structure represented by the cumulative hand written stroke data;

character dictionary means for storing characteristics of strokes of characters and directional relations of rectangular outline data of characters, including rectangular outline and direction data for partial structures of characters; and

character recognition means, including partial structure determining means which determines partial structures of the hand-written stroke data, for determining characters of the hand-written stroke data by comparing characteristics and the directional relations of the rectangular outline data stored in the character dictionary means with respect to the direction data of partial-structures of the hand-written strokes stored in the buffer means.

3. A hand-written character recognition apparatus as recited in claim 2, wherein the rectangular outline data stored in the buffer means includes coordinate data for the vertices and center points of the sides of a hypothetical rectangle surrounding a partial structure represented by the cumulative hand written stroke data minus the last inputted hand written stroke data.

4. A hand-written character recognition method comprising the steps of:

generating hand-written stroke data;

deriving and storing characteristics of the hand-written stroke data;

storing rectangular outline and direction data of the hand-written stroke data;

storing characteristics of strokes of characters and directional relations of rectangular outline data of characters, including rectangular outline and direction data for partial structures of characters; and

determining characters of the hand-written stroke data, including determining partial structures of the hand-written stroke data, by comparing characteristics and the directional relations of the stored rectangular outline data with respect to the stored direction data of partial-structures of the hand-written strokes.