RECOGNIZING HANDWRITING INPUT USING ROTATABLE SUPPORT LINES

• US 20150234588A1
• Filed: 05/01/2015
• Published: 08/20/2015
• Est. Priority Date: 03/14/2013
• Status: Active Grant

• Current Assignee: Cerence Operating Company (Cerence Inc.)
• Original Assignee: Nuance Communications, Inc. (Microsoft Corporation)
• Inventors: Andersson, Jonas, Morwing, Jonas
• Granted Patent: US 9,304,681 B2
• Time in Patent Office: Days
• US Class Current: 1/1
• CPC Class Codes: G06F 18/20 Analysing
• CPC Class Codes: G06F 3/04883 for inputting data by handwriting, e.g. gesture or text
• CPC Class Codes: G06V 30/10 Character recognition
• CPC Class Codes: G06V 30/1475 Inclination or skew detection or correction of characters or of image to be recognised
• CPC Class Codes: G06V 30/32 Digital ink

Abstract

Software, firmware, and systems are described for identifying characters in a handwritten input received from a user on an input device, irrespective of an angle that the input is received at. In one implementation, the system establishes an anchor point and distances from the anchor point to reference support lines. A set of candidate characters is identified based on received handwritten input. The system estimates support lines for each of the candidate characters. The system ranks the candidate characters based on a total deviation measurement from the expectation for each candidate, where the expectation in part is based on the established distance from the established anchor point to reference support lines, and identifies a best-ranked candidate based at least in part on a smallest total deviation measurement.

Claims

1. A method for identifying characters in a handwritten input on a touch-sensitive device, the method comprising:

establishing an anchor point on the touch-sensitive device;

establishing reference support lines,

wherein the reference support lines include first and second reference support lines,

wherein the reference support lines are associated with a reference angle around the anchor point,

wherein a first reference distance separates the anchor point and a point of the first reference support line proximate to the anchor point, and a second reference distance separates the anchor point and a point of the second reference support line proximate to the anchor point;

receiving a handwritten user input via the touch-sensitive device;

estimating candidate support lines for each candidate character in a set of candidate characters,

wherein each candidate support line is associated with a candidate angle around the anchor point,

wherein the candidate support lines for each candidate character include a first candidate support line and a second candidate support line, and

wherein a first candidate distance separates the anchor point and a point of the first candidate support line proximate to the anchor point, and a second candidate distance separates the anchor point and a point of the second candidate support line proximate to the anchor point;

for each candidate character, determining a deviation of the handwritten user input from the candidate character, wherein the deviation is based at least in part on;

an angle deviation based at least in part on a difference between the candidate angle associated with the candidate character and the reference angle,

a scale deviation determined based at least in part on a comparison between a difference of the first reference distance and the second reference distance and a difference of the first candidate distance and the second candidate distance,

a position deviation determined based at least in part on a difference between the first reference distance and the first candidate difference, or

a combination thereof;

ranking the candidate characters based at least in part on respective deviations for the candidate character; and

identifying a best-ranked candidate character based at least in part on a smallest deviation.

2. The method of claim 1, wherein the method further comprises:

establishing new reference support lines,

wherein the new reference support lines are positioned around the anchor point at an angle selected based at least in part on the candidate angle of the best-ranked candidate character,

wherein the new reference support lines include a first new reference support line and a second new reference support line, and

wherein the first reference distance separates the anchor point and a point of the first new reference support line closest to the anchor point, and the second reference distance separates the anchor point and a point of the second new reference support line closest to the anchor point;

receiving a second handwritten user input via the touch-sensitive device;

identifying a set of candidate characters based on the second handwritten user input;

for each candidate character for the second handwritten user input, determining a deviation of the second handwritten user input from the candidate character, wherein the deviation is determined using the new reference support lines;

ranking the candidate characters for the second handwritten user input based at least in part on respective deviations for the candidate characters of the second handwritten user input; and

identifying a best-ranked candidate character for the second handwritten user input based at least in part on a smallest deviation.

3. The method of claim 1, further comprising:

receiving a selection of one of the candidate characters from the user; and

setting new reference support lines having the candidate angle around the anchor point corresponding to the selected one of the candidate characters.

4. The method of claim 1, wherein the touch-sensitive device has a circular shape and the anchor point is at a center of the touch-sensitive device.

5. The method of claim 1, wherein a distance between the reference support lines is equal to an expected height of a character of the handwritten input.

6. A system for identifying characters in a handwritten input on a touch-sensitive device, the system comprising:

at least one processor;

a memory;

at least one touch-sensitive input device, coupled to the at least one processor, configured to receive a handwritten user input, wherein the touch sensitive device has an established anchor point;

a reference support line module configured to establish reference support lines,

wherein the reference support lines include a first reference support line and a second reference support line,

wherein the reference support lines are associated with a reference angle around the anchor point,

wherein a first reference distance separates the anchor point and a point of the first reference support line closest to the anchor point, and a second reference distance separates the anchor point and a point of the second reference support line closest to the anchor point;

a candidate character module configured to identify, based on the handwritten user input, a set of candidate characters;

a candidate support line module configured to estimate a set of candidate support lines for each of the candidate characters of the set of candidate characters,

wherein each of the sets of candidate support lines includes a first candidate support line and a second candidate support line,

wherein each of the sets of candidate support lines are associated with a candidate angle around the anchor point, and

wherein, for each of the sets of candidate support lines, a first candidate distance separates the anchor point and a point of the first candidate support line closest to the anchor point, and a second candidate distance separates the anchor point and a point of the second candidate support line closest to the anchor point;

a deviation module configured to compute a total deviation of the handwritten user input from each the candidate character, wherein each total deviation is based at least in part on;

an angle deviation based at least in part on a difference between the candidate angle associated with that candidate character and the reference angle,

a scale deviation determined based at least in part on a comparison between a difference of the first reference distance and the second reference distance and a difference of the first candidate distance and the second candidate distance,

a position deviation determined based at least in part on a difference between the first reference distance and the first candidate difference, or

any combination thereof; and

a ranking module configured to;

rank the candidate characters based at least in part on the respective total deviation for each candidate character; and

identify a best-ranked candidate character based at least in part on a smallest total deviation.

7. The system of claim 6,

wherein the reference support line module is further configured to establish new reference support lines,

wherein the new reference support lines are positioned around the anchor point at an angle selected based at least in part on the candidate angle of the best-ranked candidate character,

wherein the new reference support lines include a first new reference support line and a second new reference support line, and

wherein the first reference distance separates the anchor point and a point of the first new reference support line closest to the anchor point, and the second reference distance separates the anchor point and a point of the second new reference support line closest to the anchor point;

wherein the touch-sensitive device is further configured to receive a second handwritten user input;

wherein the candidate character module is further configured to identify a set of candidate characters based on the second handwritten user input;

wherein the deviation module is further configured to, for each candidate character for the second handwritten user input, determine a total deviation of the second handwritten user input from the candidate character, wherein the total deviation is determined using the new reference support lines; and

wherein the ranking module is further configured to;

rank the candidate characters for the second handwritten user input based at least in part on respective total deviations for the candidate characters of the second handwritten user input; and

identify a best-ranked candidate character for the second handwritten user input based at least in part on a smallest total deviation.

8. The system of claim 6, wherein the reference support line module is further configured to set new reference support lines having a candidate angle around the anchor point corresponding to a user-selected one of the candidate characters.

9. The system of claim 6, further comprising displaying the reference support lines on the touch-sensitive device.

10. A computer-readable storage medium storing instructions that, when executed by a computing system, cause the computing system to perform operations for identifying characters in a handwritten input on a touch-sensitive device, the operations comprising:

receiving a handwritten user input via the touch-sensitive device;

identifying a set of candidate characters based on the handwritten user input;

estimating candidate support lines for each of the candidate characters;

associating reference support lines for each candidate character;

for each candidate character,

measuring a deviation between the estimated support lines and reference support lines to determine one or more deviations from an expectation;

ranking each candidate character based on a total deviation measurement for each candidate character; and

identifying a best-ranked candidate character based at least in part on a smallest total deviation measurement.

11. The computer-readable storage medium of claim 10, wherein each reference support line is associated with a corresponding reference angle around an established anchor point on the touch-sensitive device.

12. The computer-readable storage medium of claim 11,

wherein each of the estimated candidate support lines is associated with a candidate angle around an established anchor point on the touch-sensitive device, and

wherein each of the candidate support lines is at a corresponding candidate distance from the established anchor point.

13. The computer-readable storage medium of claim 12, wherein measuring the deviation for at least one candidate character comprises determining an angle deviation between the candidate angle and the reference angle.

14. The computer-readable storage medium of claim 10, wherein measuring the deviation for at least one candidate character comprises determining a scale deviation based at least in part on a comparison between a difference in distances of the reference support lines and a difference in distances of the candidate support lines.

15. The computer-readable storage medium of claim 10, wherein measuring the deviation for at least one candidate character comprises determining a position deviation based at least in part on a position difference between a support line of the candidate support lines and a support line of the reference support lines.

16. The computer-readable storage medium of claim 10, wherein the operations further comprise:

establishing new reference support lines positioned around an anchor point of the touch-sensitive device at an angle selected based at least in part on a candidate angle of the identified best-ranked candidate character.

17. The computer-readable storage medium of claim 10, wherein the operations further comprise:

receiving a selection of one of the candidate characters from the user; and

setting new reference support lines having the candidate angle around the point.

18. The computer-readable storage medium of claim 10, wherein a distance between the reference support lines is equal to an expected height of a character of the handwritten input.

19. The computer-readable storage medium of claim 10,

wherein the handwritten user input includes a curve,

wherein the candidate characters are ranked further based at least in part on a deviation between an estimated candidate angle for each candidate character and an angle of an orthogonal projection point of a mass center of the curve to a baseline of reference support lines around the point for the candidate character;

and

wherein the reference support lines for each candidate character are associated with an angle around the point selected based at least in part on the candidate angle for the respective candidate characters.

20. The computer-readable storage medium of claim 10, wherein the ranking is further based at least in part on a pure shape distance between one of the candidate characters and the received handwritten user input.