Segmentation of handwritten patterns using stroke and jump vectors

• US 5,533,147 A
• Filed: 11/02/1994
• Issued: 07/02/1996
• Est. Priority Date: 05/29/1990
• Status: Expired due to Term

• Current Assignee: Canon Kabushiki Kaisha (Canon Inc.)
• Original Assignee: Canon Kabushiki Kaisha (Canon Inc.)
• Inventors: Matsubayashi, Kazuhiro, Arai, Tsunekazu, Takeuchi, Shinsuke
• Primary Examiner(s): Boudreau, Leo
• Assistant Examiner(s): Johns, Andrew W.
• Application Number:
  US08/334,003
• Time in Patent Office: 608 Days
• Field of Search: 382/3, 382/9, 382/13, 382/24, 382/25, 382/26, 382/177, 382/179, 382/187, 382/202, 382/203, 382/204, 382/209
• US Class Current: 382/179
• CPC Class Codes: G06V 30/347 Sampling; Contour coding; Stroke extraction
• CPC Class Codes: G06V 30/373 using a special pattern or subpattern alphabet

Abstract

Disclosed are a pattern recognition method and apparatus which eliminate the complications of handwritten input by storing two or more patterns in memory beforehand, matching input patterns with the stored patterns, and delimiting and recognizing individual patterns from the input patterns based upon the pattern matching.

Claims

1. A pattern delimiting method, comprising the steps of:

preparing a dictionary storing a string of stroke data each of which defined from a point of start of contact to a point of end of contact and jump vector data, which is represented as a movement from an end of contact point of one stroke to a start of contact point of a next stroke and feature data of the string corresponding to each other, said feature data indicating whether or not the string of stroke data and jump vector data forms an individual pattern;

matching an inputted string of stroke data and jump vector data with the string of stroke data and jump vector data stored in the dictionary beforehand and obtaining feature data of the matched string of stroke data and jump vector data from the dictionary; and

delimiting a string of stroke data and jump vector data composing an individual pattern from the inputted string of stroke data based on the feature data obtained from the matching.

2. The delimiting method according to claim 1, wherein said feature data includes data showing whether or not a last stroke data in the set of stroke data is at an end of the individual pattern.

3. The delimiting method according to claim 1, wherein said stroke data corresponds to patterns represented by coordinates.

4. The pattern delimiting method according to claim 1, wherein said storing of stroke data and jump vector data are inputted from a digitizer.

5. The pattern delimiting method according to claim 1, wherein said coordinate data is inputted as data representing a position where an input pen is pressed on an input-tablet having electrode lines wired in vertical and horizontal directions.

6. A pattern delimiting apparatus comprising:

means for storing a dictionary including a string of stroke data each of which is defined from a point of start of contact to a point of end of contact and jump vector data, each of which is represented as a movement from an end of contact point of one stroke to a start of contact point of a next stroke, and feature data of the string corresponding to each other, said feature data indicating whether or not the string of stroke data and jump vector data forms an individual pattern;

means for matching an inputted string of stroke data and jump vector data with the string of stroke data and jump vector data stored in the dictionary beforehand and obtaining feature data of the matched string of stroke data and jump vector data from the dictionary; and

means for delimiting a string of stroke data and jump vector data composing an individual pattern from the inputted string of stroke data based on the feature data obtained from said matching means.

7. The delimiting apparatus according to claim 6, wherein said feature data includes data showing whether or not a last stroke data in the set of stroke data is at an end of the individual pattern.

8. The delimiting apparatus according to claim 6, wherein said stroke data corresponds to inputted patterns represented by coordinates.

9. The pattern delimiting method according to claim 6, wherein said storing of stroke data and jump vector data are inputted from a digitizer.

10. The pattern delimiting method according to claim 6, wherein said coordinate data is inputted as data representing a position where an input pen is pressed on an input-tablet having electrode lines wired in vertical and horizontal directions.

11. A pattern delimiting method comprising the steps of:

inputting coordinate data of patterns and contact data of a pointing device;

extracting one or more strokes and one or more jump vectors from said coordinate data and said contact data, wherein each of said strokes is defined from a point of start of contact to a point of end of contact of the pointing device and each of said jump vectors is represented as a movement from an end of contact point of one stroke to a start of contact point of a next stroke; and

judging whether each of said jump vectors is in the middle of an individual pattern or between two individual patterns based upon a set of the jump vector and strokes before and after the jump vector, and an order thereof.

12. The pattern recognition method according to claim 11, wherein said coordinate data is inputted from a digitizer.

13. The pattern recognition method according to claim 11, wherein said coordinate data is inputted as data representing a position where an input pen is pressed on an input-tablet having electrode lines wired in vertical and horizontal directions.

14. A pattern delimiting apparatus comprising:

means for inputting coordinate data of patterns and contact data of a pointing device;

means for extracting one or more strokes and one or more jump vectors from said coordinate data and said contact data, wherein each of said strokes is defined from a point of start of contact to a point of end of contact of the pointing device and each of said jump vectors is represented as a movement from an end of contact point of one stroke to a start of contact point of a next stroke; and

means for judging whether each of said jump vectors is in the middle of an individual pattern or between two individual patterns based upon a set of the jump vector and strokes before and after the jump vector, and an order thereof.

15. The pattern recognition method according to claim 14, wherein said coordinate data is inputted from a digitizer.

16. The pattern recognition method according to claim 14, wherein said coordinate data is inputted as data representing a position where an input pen is pressed on an input-tablet having electrode lines wired in vertical and horizontal directions.