Method and apparatus for decoding handwritten characters

• US 7,359,551 B2
• Filed: 10/15/2002
• Issued: 04/15/2008
• Est. Priority Date: 10/15/2001
• Status: Expired due to Fees

• Current Assignee: Silverbrook Research Pty Limited
• Original Assignee: Silverbrook Research Pty Limited
• Inventors: Napper, Jonathon Leigh
• Primary Examiner(s): Ahmed; Samir
• Assistant Examiner(s): RASHIDIAN, MOHAMMAD M
• Application Number:
  US10/492,168
• Publication Number:
  US 20050175241A1
• Time in Patent Office: 2,009 Days
• Field of Search: 382/187, 382/259, 382/181, 382/185, 382/186, 382/164, 382/171, 382/173, 382/179, 382/177
• 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
• CPC Class Codes: G06V 30/244 using graphical properties, e.g. alphabet type or font

Abstract

Disclosed is a method of performing written character recognition, including sequentially comparing a given input character with one of a series of stored character models. The step of comparing further includes the steps of: decomposing the given input character into one or more segments in accordance with a segmentation scheme associated with the present character model; evaluating the segment or segments against a segment model associated with the present character model, such that the evaluation produces a score which measures the degree of conformity with the model; repeating said step of comparing for a plurality of character models in the series of character models; and selecting the character model which is assigned the highest score and classifying the input character as the character associated with that character model.

Claims

1. A method of performing written character recognition on an input character, the method comprising the steps of:

(a) for each of a plurality of character models;

(i) decomposing the input character into one or more segments in accordance with a segmentation scheme associated with the character model; and

(ii) evaluating said segment or segments against a segment model associated with the present character model, such that the evaluation produces a score indicative of conformity with the model; and

(b) selecting the character model that produced the highest score and classifying the input character as the character associated with that character model;

wherein the step of decomposing the input character into one or more segments is performed such that each segment has a minimum length according to the following equation;

$l_{<br><br>\min }=\mathrm{\lambda <br><br>}<br><br>\underset{i=2}{\overset{n}{\sum <br><br>}}<br><br><br><br><br><br>p_i-p_{i-1}<br><br>$

where

l_min=minimum length of segment

p_i=coordinates of point i

n =number of points in the stroke

λ

=minimum length threshold.

2. A method as claimed in claim 1 wherein the character for recognition is one of a letter, a number or a punctuation mark.

3. A method as claimed in claim 1 wherein if the given input character can not be segmented according to the segmentation scheme associated with the present character model, then the present character model is rejected as a possible match for the given input character.

4. A method as claimed in claim 1 wherein the evaluation procedure assigns a score which rewards closer conformance to an ideal.

5. A method as claimed in claim 4 wherein the evaluation procedure uses fuzzy logic rules.

6. A method as claimed in claim 1 wherein each character model may have associated with it a plurality of segmentation schemes to facilitate recognition of different forms of the same character.

7. A method as claimed in claim 1 wherein different rules for decomposing the input character into one or more segments may be defined for different character models.

8. Apparatus for performing handwritten character recognition, the apparatus including:

an input surface responsive to the movement of a stylus thereupon for producing an input stimulus;

a memory;

a processor for receiving said input stimulus and processing said input stimulus, wherein said processor is adapted to perform the method in accordance with any claim 1.