USING PERSONALIZED HEALTH INFORMATION TO IMPROVE SPEECH RECOGNITION

• US 20090326937A1
• Filed: 04/21/2008
• Published: 12/31/2009
• Est. Priority Date: 04/21/2008
• Status: Abandoned Application

• Current Assignee: Microsoft Technology Licensing LLC (Microsoft Corporation)
• Original Assignee: Microsoft Corporation
• Inventors: Chitsaz, Behrooz, Gounares, Alexander G., Choing, Hong
• Application Number:
  US12/106,497
• Publication Number:
  US 20090326937A1
• Time in Patent Office: Days
• US Class Current: 704/235
• CPC Class Codes: G10L 15/24 Speech recognition using non-acoustical features
• CPC Class Codes: G10L 2015/227 of the speaker; Human-factor methodology
• CPC Class Codes: G16H 10/60 for patient-specific data, e.g. for electronic patient records
• CPC Class Codes: G16H 40/63 for local operation
• CPC Class Codes: Y02A 90/10 Information and communication technologies [ICT] supporting adaptation to climate change, e.g. for weather forecasting or climate simulation

Abstract

The claimed subject matter provides systems and/or methods that improve speech recognition in the medical context. The system includes mechanisms that access personal health records associated with patients and/or analyze the personal health records for current diseases and/or past ailments. The system thereafter acquires attributes associated with the diseases or ailments and dynamically populates a speech model with these attributes. The speech model utilizes the attributes associated with the diseases or ailments to more accurately transcribe a voice pattern into text that can be projected on a visual display or persisted to a storage device.

Claims

1. A system implemented on a machine that improves speech recognition, comprising:

a component that accesses via an interface a personal health record associated with a patient, the component analyzes the personal health record for at least one of a current disease or a past ailment, acquires an attribute or a variant of the attribute associated with the current disease or the past ailment, dynamically populates a recognition component with the attribute or the variant of the attribute associated with the current disease or the past ailment, the recognition component utilizes the attribute or the variant of the attribute to transcribe a voice pattern into text projected on a visual display or persisted to a storage device.

2. The system of claim 1, the component dynamically modifies the recognition component based at least in part on a speech pattern associated with a healthcare professional.

3. The system of claim 1, the component automatically modifies the recognition component based at least in part on a level of experience associated with a healthcare professional.

4. The system of claim 1, the component modifies the recognition component based on a medical specialty associated with a healthcare professional.

5. The system of claim 1, the component facilitates a search for a medically recognized code, the medically recognized code employed by the recognition component to standardize the text projected on the visual display or persisted to the storage device.

6. The system of claim 1, the component expunges entirely from the recognition component the attribute or the variant of the attribute associated with the current disease or the past ailment obtained from the personal health record associated with the patient and replaces the attribute or the variant of the attribute associated with the current disease or the past ailment with a second attribute or a variant of the second attribute associated with another personal health record associated with a second patient.

7. The system of claim 1, the recognition component upon receipt of the voice pattern contemporaneously modifies the voice pattern to conform to a prescribed transcription formatting standard, the prescribed transcription formatting standard based at least in part on a universally recognized medical code.

8. A method implemented on a machine that improves speech recognition, comprising:

obtaining a personal health record associated with a patient;

analyzing the personal health record for at least one of a current disease or a past ailment;

extracting an attribute or a variant of the attribute associated with the current disease or the past ailment from the personal health record;

populating a recognition component with the attribute or the variant of the attribute associated with the current disease or the past ailment;

transcribing a voice pattern into text based at least in part on the attribute or the variant of the attribute; and

projecting the text onto a visual display for confirmation of accuracy by an intermediary associated with the voice pattern, and based on the confirmation of accuracy persisting the text to local or remote storage media.

9. The method of claim 8, the populating further comprising modifying the recognition component based at least in part on the voice pattern associated with the intermediary.

10. The method of claim 8, the populating further comprising modifying the recognition component based at least in part on a level of experience associated with the intermediary.

11. The method of claim 8, the populating further comprising amending the recognition component based on a medical functionality associated with the intermediary.

12. The method of claim 8, further comprising initiating a search for a medically recognized code employed to standardize text transcription.

13. The method of claim 8, the populating further comprising:

removing from the recognition component all aspects associated with the attribute or the variant of the attribute associated with the current disease or the past ailment obtained from the personal health record associated with the patient; and

substituting the attribute or the variant of the attribute associated with the current disease or the past ailment with a second attribute or a variant of the second attribute associated with another personal health record associated with a second patient.

14. The method of claim 8, the transcribing further comprising upon receipt of the voice pattern modifying the voice pattern to conform to a prescribed transcription formatting standard, the prescribed transcription formatting standard based at least in part on a universally recognized set of medical codes.

15. A system that improves speech recognition, comprising:

means for populating a recognition component with an attribute or a variant of the attribute associated with at least one of a current disease or a past ailment retrieved from a personal health record;

means for transcribing a voice pattern into text based at least in part on the attribute or the variant of the attribute associated with at least one of the current disease or the past ailment; and

means for displaying the text onto a visual display for confirmation of accuracy by an intermediary associated with the voice pattern.

16. The system of claim 15, further comprising means for removing from the recognition component all aspects associated with the attribute or the variant of the attribute associated with the current disease or the past ailment obtained from the personal health record associated with the patient.

17. The system of claim 15, further comprising means for substituting the attribute or the variant of the attribute associated with the current disease or the past ailment with a second attribute or a variant of the second attribute associated with another personal health record associated with a second patient.

18. The system of claim 15, further comprising means for modifying the voice pattern to conform to a prescribed transcription formatting standard, the prescribed transcription formatting standard based at least in part on a universally recognized set of medical codes.

19. The system of claim 15, the means for populating modifies the recognition component based at least in part on a level of experience associated with the intermediary.

20. The system of claim 15, the means for transcribing modifies the voice pattern upon receipt to conform to a prescribed transcription formatting standard, the prescribed transcription formatting standard based at least in part on a universally recognized set of medical codes.