System and method for providing user-directed constraints for handwriting recognition

• US 6,401,067 B2
• Filed: 01/28/1999
• Issued: 06/04/2002
• Est. Priority Date: 01/28/1999
• Status: Expired due to Term

• Current Assignee: Nuance Communications, Inc. (Microsoft Corporation)
• Original Assignee: International Business Machines Corporation
• Inventors: Lewis, James R., Subrahmonia, Jayashree, Ratzlaff, Eugene H., Perrone, Michael P., Pitrelli, John F.
• Primary Examiner(s): Banks-Harold, Marsha D.
• Assistant Examiner(s): McFadden, Susan
• Application Number:
  US09/238,288
• Publication Number:
  US 20010053978A1
• Time in Patent Office: 1,223 Days
• Field of Search: 704/231, 704/275, 704/270, 382/189, 382/229
• US Class Current: 704/275
• CPC Class Codes: G06F 18/40 Software arrangements specially adapted for pattern recognition, e.g. user interfaces or toolboxes therefor
• CPC Class Codes: G06V 30/10 Character recognition
• CPC Class Codes: G06V 30/262 using context analysis, e.g. lexical, syntactic or semantic context

Abstract

A data recognition system and method which allows a user to select between a “default recognition” mode and a “constrained recognition” mode via a user interface. In the default recognition mode, a recognition engine utilizes predetermined default recognition parameters to decode data (e.g., handwriting and speech). In the constrained recognition mode, the user can select one or more of a plurality of recognition constraints which temporarily modify the default recognition parameters to decode uncharacteristic and/or special data. The recognition parameters associated with the selected constraint enable the recognition engine to utilize specific information to decode the special data, thereby providing increased recognition accuracy.

Claims

1. A machine recognition system, comprising:

an input device nor receiving input data;

a user interface comprising a menu for selecting at least one of a plurality of recognition modes comprising a default recognition mode and a constrained recognition mode, wherein the constrained recognition mode enables a user to select one or more recognition constraint options, the recognition constraint options comprising a decoding algorithm constraint that can be selected to modify one of a decoding method, a decoding parameter or both;

selection means for allowing a user to select a portion the input data for recognition;

a recognition engine for recognizing the selected input/data using recognition parameters associated with the selected recognition mode; and

an output device for outputting recognition results.

2. The system of claim 1, wherein the machine recognition system is a speech recognition system and the input data is acoustic speech utterances.

3. The system of claim 1, wherein the machine recognition system is a handwritten text recognition system and the input data is handwriting data.

4. The system of claim 1, further comprising:

word list storage means for storing at least one word list; and

prototype storage means for storing at least one prototype, wherein said recognition means retrieves one of a word list, a prototype and a combination thereof, corresponding to the selected recognition mode for performing recognition.

5. The system of claim 1, wherein said selection means includes means for tagging the selected input data with a label corresponding to the selected recognition mode.

6. The system of claim 5, wherein the selection means is included within said input device.

7. A program storage device readable by a machine, tangibly embodying a program of instructions executable to perform method steps for recognizing input data in accordance with a user-selected recognition mode, the method steps comprising:

receiving the input data;

displaying a menu having a plurality of recognition mode menu options for selection by a user, the menu options comprising a default recognition mode find a constrained recognition mode, wherein the constrained recognition mode enables a user to select one or more recognition constrain options, the recognition constraint options comprising a decoding algorithm constraint that can be selected to modify one of a decoding method, a decoding parameter or both;

selecting at least a portion of the input data for recognition in accordance with the selected recognition mode; and

recognizing the selected input data using recognition parameters associated with the recognition mode selected by the user.

8. The program storage device of claim 7, wherein the machine is a speech recognition device and the input data is acoustic speech utterances.

9. The program storage device of claim 7, wherein the machine is a handwritten text recognition device and the input data is handwriting data.

10. The program storage device of claim 7, further including instructions for performing the step of retrieving one of a corresponding word list and a corresponding prototype and a combination thereof in response to the selected recognition mode, wherein recognition is performed using the retrieved word list and writer prototype.

11. The program storage device of claim 7, further including instructions for performs the steps of:

displaying a plurality of recognition constraints for selection by the user if the user selects the constrained recognition mode; and

modifying default recognition parameters in accordance with at least one selected recognition constraint.

12. The program storage device of claim 7, further including instructions for performing the step of:

resetting the selected recognition mode after recognition of the selected portion of input data.

13. The program storage device of claim 12, further including instructions for performing the steps of:

causing recognition of the selected portion of input data to be deferred at a subsequent time in response to a user request;

tagging the selected portion of input data with a label corresponding to the selected recognition mode; and

performing recognition of the selected portion of input data at the subsequent time in accordance with the recognition mode corresponding to the label.

14. The program storage device of claim 11, wherein recognition of the selected portion of input data is performed after the step of selecting the recognition mode.

15. The program storage device of claim 11, further including instructions for performing the step of:

performing recognition of subsequently selected portions of the input data in accordance with the selected recognition mode until a new recognition mode is selected, wherein recognition of additionally selected portions of the input data is performed in accordance with the newly selected recognition mode.

16. The program storage device of claim 11, further including instructions for performing the steps of:

causing recognition of the selected input data to be deferred at a subsequent time in response to a user request;

tagging the selected portion of input data with a label corresponding to the selected recognition mode; and

performing recognition of the labeled portion of input data at the subsequent time in accordance with the recognition mode corresponding to the label.

17. The program device of claim 16, further including instructions for performing the step of tagging subsequently selected portions of input data with the previously selected recognition mode until a new recognition mode is selected, wherein additionally selected portions of the input data are tagged with a label corresponding to a newly selected recognition mode after selection of the new recognition mode.

18. A method for decoding handwriting data in a handwritten text recognition system, comprising the steps of:

displaying handwriting data;

selecting between one of a default recognition mode and a constrained recognition mode the constrained recognition mode having a plurality of recognition constraints associated therewith, the recognition constraints comprising a character set constraint that constrains characters that are recognized;

recognizing a user-selected portion of the displayed handwriting data in accordance with default recognition parameters if the default recognition mode is selected for the user-selected portion; and

selecting at least one of the plurality of recognition constraints if the constrained recognition mode is selected and recognizing the user-selected portion of the displayed handwriting data in accordance with recognition parameters associated with the recognition constraint selected for the user-selected portion.

19. A method for decoding speech data in a speech recognition system, comprising the steps of:

receiving input speech data and processing the input speech data in successive frames;

decoding the processed speech data using default recognition parameters and outputting the decoding results;

selecting at least a portion of the decoding results;

selecting at least one of a plurality of recognition constraints;

recognizing the processed input speech data corresponding to the selected portion of decoding results in accordance with the selected recognition constraint.