Handling of diacritic points

• US 20060193519A1
• Filed: 02/28/2006
• Published: 08/31/2006
• Est. Priority Date: 02/28/2005
• Status: Active Grant

• Current Assignee: Cerence Operating Company (Cerence Inc.)
• Original Assignee: Zi Decuma AB
• Inventors: Sternby, Jakob
• Granted Patent: US 7,596,272 B2
• Time in Patent Office: Days
• US Class Current: 382/186
• CPC Class Codes: G06V 30/2272 with lexical matching
• CPC Class Codes: G06V 30/347 Sampling; Contour coding; Stroke extraction

Abstract

A method for recognition of a handwritten pattern comprises selecting core points among a sequence of detected points of the handwritten pattern. The core points are selected for use in segmenting and recognizing the handwritten pattern. The method may further comprise identifying potential diacritics in the sequence of points of the handwritten pattern, determining features of each core point, assigning at least one feature associated with identified potential diacritics to each core point of a subset of core points, and comparing the handwritten pattern to templates.

Claims

1. A method for recognition of a handwritten pattern comprising one or more curves and representing a plurality of symbols, said method comprising:

detecting the handwritten pattern as a sequence of points along the one or more curves of the handwritten pattern,

identifying potential diacritics in the sequence of points of the handwritten pattern,

selecting core points among the sequence of points, wherein said core points are selected for use in segmenting the handwritten pattern and recognizing these segments of the handwritten pattern as symbols,

determining features of the one or more curves at or in the vicinity of each core point,

assigning at least one feature associated with a identified potential diacritics to each core point of a subset of core points, and

comparing the handwritten pattern to templates, each template representing at least one symbol or part of a symbol, said comparing comprising;

stepwise analyzing the core points in sequence, wherein core points represent possible segmentation points and sequences of core points from a first possible segmentation point to a second possible segmentation point represent possible symbols, said analyzing being performed by

matching the features of sequences of core points that either start with the first core point or the last core point of a previous sequence of core points to said templates and calculating a distance value, and

assigning a cumulative distance value to the last core point in the matched sequence of core points, said cumulative distance value being a sum of a distance value assigned to the first core point in the sequence and the calculated distance value,

whereby a smallest cumulative distance value for all sequential core points is assigned to the last core point and corresponds to a sequence of matched templates which represent a plurality of symbols as a possible recognition result of the handwritten pattern.

2. The method according to claim 1, wherein said comparing is performed by

forming a graph having nodes and arches connecting the nodes, wherein each node represents a core point corresponding to a possible segmentation point between two symbols in the handwritten pattern and each arch represents a path along a sequence of core points from one node to another node, said path corresponding to a possible symbol in the handwritten pattern,

assigning at least one distance value to each path by matching the features of the sequence of core points to said templates, and

determining at least the path through the graph from the first node to the last node with the smallest cumulative distance value, said path corresponding to a sequence of matched templates which represent a plurality of symbols as a possible recognition result of the handwritten pattern.

3. The method according to claim 2, wherein several possible recognition results are determined corresponding to paths through the graph having cumulative distance values below a threshold value.

4. The method according to claim 2, wherein said graph is formed by sequentially adding core points as nodes and wherein said assigning of at least one distance value to each path is performed during forming of the graph.

5. The method according to claim 2, wherein said assigning of at least one distance value comprises assigning each path the lowest distance value corresponding to the template that best matches the features of the sequence of core points.

6. The method according to claim 1, wherein the act of assigning at least one feature associated with a detected diacritic is performed before the act of comparing the handwritten pattern to templates.

7. The method according to claim 1, wherein a number of core points being selected is limited by the shape of said one or more curves.

8. The method according to claim 7, wherein the number of core points being selected is limited to points holding information of directional changes of said one or more curves.

9. The method according to claim 1, wherein said selecting of core points comprises selecting points which are local extreme points of said one or more curves in a direction orthogonal to a main writing direction of the handwritten pattern.

10. The method according to claim 9, wherein the core points are classified as N core points, S core points and M core points, wherein an N core point being a local maxima core point in a direction orthogonal to a main writing direction, an S core point being a local minima core point in a direction orthogonal to a main writing direction and an M core point being an intermediate core point to N and S core points, further comprising the act of selecting the core points of said subset of core points as N and M core points.

11. The method according to claim 1, wherein said selecting of core points comprises selecting points in which an absolute value of the derivative of the one or more curves exceeds a threshold value.

12. The method according to claim 1, wherein said selecting of core points comprises selecting a core point where a deviation of the one or more curves from a straight line between two adjacent core points exceeds a threshold value.

13. The method according to claim 1, wherein the determining of features comprises determining a y-coordinate of the core point, said y-coordinate defining a position in a direction orthogonal to a main writing direction.

14. The method according to claim 1, wherein the determining of features comprises determining a direction of the curve towards the core point and a direction of the curve from the core point.

15. The method according to claim 1, wherein the determining of features comprises determining a x-coordinate displacement from a previous core point, said x-coordinate defining a position in a direction parallel with a main writing direction.

16. The method according to claim 1, wherein the determining of features comprises determining whether a pen has been lifted up or pressed down in the core point.

17. The method according to claim 1, wherein the distance value between a core point in the handwritten pattern and a point of a template is calculated as a weighted sum of the distance between each feature of the core point and the point of the template.

18. The method according to claim 1, wherein the handwritten pattern is compared to templates, wherein a plurality of templates represent different appearances of the same symbol.

19. The method according to claim 1, wherein the possible recognition result of the handwritten pattern is compared to a dictionary.

20. The method according to claim 1, wherein specific sequences of matched templates are disallowed.

21. The method according to claim 1, wherein a weight is given to the cumulative distance value for specific sequences of matched templates.

22. The method according to claim 1, wherein the handwritten pattern is written in cursive writing.

23. A device for recognition of a handwritten pattern comprising one or more curves and representing a plurality of symbols, said device comprising:

means performing the acts of any one of the preceding claims.

24. A computer readable memory device having stored thereon instructions that are executable by a computer, the instructions being capable of causing a computer to execute a method according to claim 1.