Method and device for recognition of a handwritten pattern

• US 7,437,001 B2
• Filed: 06/05/2007
• Issued: 10/14/2008
• Est. Priority Date: 03/07/2001
• Status: Active Grant

• Current Assignee: Cerence Operating Company (Cerence Inc.)
• Original Assignee: Zi Decuma AB
• Inventors: Morwing, Jonas, Sparr, Gunnar
• Primary Examiner(s): Chawan; Sheela C
• Application Number:
  US11/806,928
• Publication Number:
  US 20080002884A1
• Time in Patent Office: 497 Days
• Field of Search: 382186-189, 382/177, 382/159, 382/174, 382/197, 382/229, 382/178, 283/67, 283/122, 283/187, 283/192, 705/12, 235/51
• US Class Current: 382/186
• CPC Class Codes: G06V 30/1423 the instrument generating sequences of position coordinates corresponding to handwriting preprocessing or recognising digital ink G06V30/32

Abstract

A method for recognition of a handwritten pattern comprises the steps of forming (4) a representation of the handwritten pattern, forming (6) at least two subconfigurations by dividing the representation of the handwritten pattern, and processing the subconfigurations. The step of processing comprises the steps of comparing (8) each subconfiguration with reference configurations, selecting (10) at least one subconfiguration candidate for each subconfiguration among the reference configurations based on said step of comparing, and determining (12) at least one candidate pattern consisting of one selected subconfiguration candidate for each subconfiguration. The method further comprises the steps of comparing (14) the representation of the handwritten pattern to the candidate pattern, and computing (16) a cost function in order to find a closest matching candidate pattern. The cost function is based on the comparing of the representation of the handwritten pattern to the candidate pattern and on the comparing of the subconfigurations to the selected subconfiguration candidates in the candidate pattern.

Claims

1. A method for recognition of a handwritten pattern, comprising:

forming a representation of the handwritten pattern,

forming at least two subconfigurations by dividing the representation of the handwritten pattern,

processing the subconfigurations, said processing comprising;

comparing each subconfiguration with reference configurations,

selecting at least one subconfiguration candidate for each subconfiguration among the reference configurations based on said comparing of each subconfiguration, and

determining at least one candidate pattern consisting of one selected subconfiguration candidate for each subconfiguration,

comparing the representation of the handwritten pattern to the candidate pattern, and

computing a cost function based on the comparing of the representation of the handwritten pattern to the candidate pattern and on the comparing of the subconfigurations to the selected subconfiguration candidates in the candidate pattern in order to find a closest matching candidate pattern.

2. The method according to claim 1, wherein determining a candidate pattern comprises merging one selected subconfiguration candidate for each subconfiguration to a candidate pattern, said merging being made according to a predefined rule.

3. The method according to claim 2, wherein comparing the representation of the handwritten pattern to the candidate pattern comprises measuring a difference of a positioning of the subconfigurations relative to each other compared to a positioning of the selected subconfiguration candidates relative to each other in the candidate pattern.

4. The method according to claim 3, wherein measuring the difference of the positioning comprises:

transforming the representation of the handwritten pattern to make it resemble the candidate pattern as much as possible, and

computing a geometric proximity between the representation of the handwritten pattern and the transformed representation of the handwritten pattern.

5. The method according to claim 3, wherein measuring the difference of the positioning comprises:

transforming the candidate pattern to make it resemble the representation of the handwritten pattern as much as possible, and

computing a geometric proximity between the candidate pattern and the transformed candidate pattern.

6. The method according to claim 3, wherein measuring the difference of the positioning comprises computing a geometric proximity between the representation of the handwritten pattern and the candidate pattern.

7. The method according to claim 1, wherein comparing the representation of the handwritten pattern to the candidate pattern comprises measuring a difference of a positioning of the subconfigurations relative to each other compared to a positioning of the selected subconfiguration candidates relative to each other in the candidate pattern.

8. The method according to claim 7, wherein measuring the difference of the positioning comprises:

transforming the representation of the handwritten pattern to make it resemble the candidate pattern as much as possible, and

computing a geometric proximity between the representation of the handwritten pattern and the transformed representation of the handwritten pattern.

9. The method according to claim 7, wherein measuring the difference of the positioning comprises:

transforming the candidate pattern to make it resemble the representation of the handwritten pattern as much as possible, and

computing a geometric proximity between the candidate pattern and the transformed candidate pattern.

10. The method according to claim 7, wherein measuring the difference of the positioning comprises computing a geometric proximity between the representation of the handwritten pattern and the candidate pattern.

11. The method according to any one of the preceding claims, wherein forming at least two subconfigurations by dividing the representation of the handwritten pattern comprises forming different sets of subconfigurations by dividing the representation of the handwritten pattern in alternative ways.

12. The method according to claim 11, wherein processing the subconfigurations is performed for each set of subconfigurations.

13. The method according to claim 1, wherein the recognition of the handwritten pattern is independent of a group of transformations.

14. A device for recognition of a handwritten character, comprising:

means for detecting the handwritten pattern and forming a representation of the handwritten pattern,

means for forming at least two subconfigurations by dividing the representation of the handwritten pattern,

a database, in which reference configurations are stored,

means for comparing each subconfiguration with the reference configurations,

means for selecting, for each subconfiguration, at least one subconfiguration candidate from the reference configurations in the database based on said comparing of each subconfiguration,

means for determining at least one candidate pattern consisting of one selected subconfiguration candidate for each subconfiguration,

means for comparing the representation of the handwritten pattern with the candidate pattern, and

means for computing a cost function based on the comparing of the representation of the handwritten pattern with the candidate pattern and on the comparing of the subconfigurations to the selected subconfiguration candidates in the candidate pattern in order to find a closest matching candidate pattern.