Natural input recognition tool

• US 20030216913A1
• Filed: 05/14/2002
• Published: 11/20/2003
• Est. Priority Date: 05/14/2002
• Status: Active Grant

• Current Assignee: Microsoft Technology Licensing LLC (Microsoft Corporation)
• Original Assignee: Microsoft Corporation
• Inventors: Garside, Adrian J., Winkler, David V., Keely, Leroy B., Soin, Ravipal, Geidl, Erik M.
• Granted Patent: US 7,380,203 B2
• Time in Patent Office: Days
• US Class Current: 704/235
• 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 recognition tool according to the invention intelligently recognizes natural input before it is passed to a destination or target application. More particularly, the recognition tool according to the invention provides better formatting for text recognized from natural input, based upon the context in which the text is being inserted into a target application. The recognition tool also provides various tools for correcting inaccurately recognized text. For example, the recognition tool allows a user to select only a part of an inaccurate text, and then identifies alternate text candidates based upon the selected portion of the inaccurate text. Further, when the user selects text containing multiple words for correction, the recognition tool provides cross combinations of alternate text candidates for the user'"'"'s selection. Still further, if the user replaces inaccurate text by submitting a new natural input object, the recognition tool ensures that the text recognized from the new natural input object is different from the inaccurate text been replaced. The recognition tool additionally improves the recognition experience after recognized text has been provided to the target application. The recognition tool provides the target application with the original natural input object for the recognized text, along with the alternate text candidates for that original natural input object. Thus, the target application can use the alternate text cabinets to correct inaccurately recognized text. Further, a user can inserts the original natural input object for recognized text within the target application.

Claims

1. A method of generating text from a natural input object for insertion into a target application, comprising:

receiving a natural input object from which recognized text is to be recognized;

obtaining a context for insertion of recognized text recognized from the natural input object into a target application; and

generating recognized text from the natural input object based upon the obtained context.

2. The method of generating text recited in claim 1, wherein the context is obtained from the target application.

3. The method of generating text recited in claim 2, wherein

obtaining a context for insertion includes performing one or more of:

identifying a field type for a field in the target application into which the recognized text will be inserted;

identifying text that has already been inserted into the field;

identifying text adjacent to an insertion location for inserting the recognized text;

identifying a length of the insertion location for inserting the recognized text; and

identifying any text that will be replaced by the recognized text.

4. The method of generating text recited in claim 3, wherein generating text from the natural input object includes

recognizing text from the natural input object, and

formatting the recognized text.

5. The method of generating text recited in claim 4, wherein formatting the recognized text includes providing spacing for the recognized text.

6. The method of generating text recited in claim 5, further comprising:

assigning a category to the field in the target application into which the recognized text will be inserted; and

providing spacing for the recognized text according to the category for the field in the target application into which the recognized text will be inserted.

7. The method of generating text recited in claim 6, further including:

determining if context information is provided for the field in the target application into which the recognized text will be inserted; and

if context information is not provided for the field in the target application into which the recognized text will be inserted, then categorizing the field into a first category.

8. The method of generating text recited in claim 7, further including

determining if the context information allows for control of insertion of text into the field in the target application into which the recognized text will be inserted;

and

if the context information allows for control of insertion of text into the field in the target application into which the recognized text will be inserted, then categorizing the field into a second category.

9. The method of generating text recited in claim 6, further including:

providing spacing for the recognized text based upon a language of the recognized text.

10. The method of generating text recited in claim 5, further comprising:

determining if text before the insertion location forms a word when concatenated with the recognized text; and

if text before the insertion location forms a word when concatenated with the recognized text, omitting a space before the recognized text.

11. The method of generating text recited in claim 5, further comprising:

determining if text after the insertion location forms a word when concatenated with the recognized text; and

if text after the insertion location forms a word when concatenated with the recognized text, omitting a space after the recognized text.

12. The method of generating text recited in claim 4, wherein formatting the recognized text includes determining capitalization for the recognized text.

13. The method of generating text recited in claim 12, wherein formatting the recognized text includes capitalizing every character in the recognized text.

14. The method of generating text recited in claim 12, wherein formatting the recognized text includes retaining a capitalization case, recognized from the natural input object case, in the recognized text.

15. The method of generating text recited in claim 4, wherein formatting the recognized text includes determining punctuation for the recognized text.

16. The method of generating text recited in claim 15, formatting the recognized text includes adding punctuation to the recognized text.

17. The method of generating text recited in claim 3, wherein generating text from the natural input object includes

determining if the recognized text is replacing text existing in the target application;

and

if the recognized text is replacing text existing in the target application, recognizing the recognized text from the natural input object to be different from the existing text being replaced.

18. The method of generating text recited in claim 1, wherein the context is obtained from a correction process for correcting existing text.

19. The method of generating text recited in claim 18, further comprising:

determining of the recognized text is replacing existing text; and

if the recognized text is replacing text existing in the target application, recognizing the recognized text from the natural input object to be different from the existing text being replaced.

20. A method of inserting text recognized from a natural input object into a target application, comprising:

determining one or more text candidates to which the natural input object corresponds;

recognizing text from the natural input object by selecting the text candidate that most closely corresponds to the natural input object;

inserting the text object into a target application; and

providing the target application with information relating to the natural input object.

21. The method of inserting text recited in claim 20, further comprising:

providing the target application with a plurality of the one or more text candidates.

22. The method of inserting text recited in claim 20, further comprising:

providing the target application with the natural input object.

23. The method of inserting text recited in claim 20, further comprising:

providing the target application with both the natural input object and a plurality of the one or more text candidates.

24. A method of recognizing text from natural input objects, comprising:

determining one or more first text candidates to which a first natural input object corresponds;

determining one or more second text candidates to which a second natural input object corresponds;

creating a lattice of possible text combinations from the first text candidates and the second text candidates;

receiving input selecting one of the first text candidates as recognized text; and

in response to receiving input selecting one of the first text candidates as recognized text, employing the lattice to select one of the second text candidates as recognized text.

25. The method of recognizing text from natural input objects recited in claim 24, further comprising:

receiving from a recognition process;

first text corresponding to the first natural input object;

the one or more first text candidates;

second text corresponding to the second natural input object; and

the one or more second text candidates;

changing the first text corresponding to the first natural input object to the selected one of the first text candidates; and

changing the second text corresponding to the second natural input object to the selected one of the second text candidates.

26. A method of correcting text, comprising:

receiving input designating existing text;

receiving a natural input object for replacement of the existing text;

determining one or more text candidates corresponding to the natural input object, such that none of the text candidates is the same as the existing text;

selecting one of the one or more text candidates that most closely corresponds to the natural object; and

replacing the existing text with the selected text candidate that most closely corresponds to the natural object.

27. The method of correcting text recited in claim 26, further comprising:

displaying a correction interface for correcting the existing text, the correction interface including an object for inputting the natural input object to replace the existing text;

receiving input selecting the object for inputting the natural input object to replace the existing text; and

providing an interface for inputting the natural input object.

28. A method of correcting existing text, comprising:

receiving input designating existing text and selecting only a portion of the existing text;

selecting one or more text candidates corresponding to the selected portion of the existing text;

displaying the one or more text candidates;

receiving input selecting one of the one or more text candidates; and

replacing the existing text with the selected text candidate.

29. The method of correcting existing text recited in claim 28, wherein each of the one or more candidates differ from the existing text by only the selected portion of the text.

30. The method of correcting existing text recited in claim 28, wherein each of the one or more text candidates includes the selected portion of the text.

31. The method of correcting existing text recited in claim 28, wherein the one or more text candidates are selected from a recognition dictionary.

32. A recognition tool for recognizing text from natural input objects, comprising:

a recognition context module that determines a context of an insertion locating into which recognized text will be inserted;

a recognition module that

identifies one or more text candidates corresponding to a natural input object;

selects one of the one or more text candidates that most closely corresponds to the natural input object; and

generates text from the selected text candidate based upon the context determined by the recognition context module.

33. The recognition tool recited in claim 32, wherein the recognition module selects the one of the one or more text candidates based upon the context determined by the recognition context module.

34. The recognition tool recited in claim 32, wherein the recognition module formats the selected text candidate based upon the context determined by the recognition context module.

35. The recognition tool recited in claim 32, wherein the recognition module formats the selected text candidate based upon the context determined by the recognition context module.

36. The recognition tool recited in claim 35, wherein the recognition module provides spacing for the selected text candidate based upon the context determined by the recognition context module.

37. The recognition tool recited in claim 35, wherein the recognition module determines capitalization for the selected text candidate based upon the context determined by the recognition context module.

38. The recognition tool recited in claim 35, wherein the recognition module determines punctuation for the selected text candidate based upon the context determined by the recognition context module.

39. The recognition tool recited in claim 32, further comprising:

an object model module that, for each of a plurality of selected text candidates, stores

the natural input object for which the selected text candidate was selected; and

the one or more text candidates corresponding to the natural input object for which the selected text candidate was selected.

40. The recognition tool recited in claim 32, further comprising:

a correction module that provides a user interface for correcting text generated by the correction module.

41. The recognition tool recited in claim 40, wherein

if the correction module receives a natural input object to replace existing text, the correction module notifies the recognition context module of the existing text that is being replaced;

and

the recognition context module prevents the recognition module from generating text identical to the existing text that is being replaced.