System and method for automating transcription services

• US 7,006,967 B1
• Filed: 02/04/2000
• Issued: 02/28/2006
• Est. Priority Date: 02/05/1999
• Status: Expired due to Fees

• Current Assignee: Custom Speech USA, Inc.
• Original Assignee: Custom Speech USA, Inc.
• Inventors: Kahn, Jonathan, Flynn, Thomas P., Tippe, Robert J., Qin, Charles
• Primary Examiner(s): McFadden, Susan
• Assistant Examiner(s): OPSASNICK, MICHAEL N
• Application Number:
  US09/889,870
• Time in Patent Office: 2,216 Days
• Field of Search: 704/235.277
• US Class Current: 704/235
• CPC Class Codes: G10L 15/075 supervised, i.e. under machine guidance
• CPC Class Codes: G10L 15/26 Speech to text systems G10L15/08 takes precedence
• CPC Class Codes: G10L 15/30 Distributed recognition, e.g. in client-server systems, for mobile phones or network applications

Abstract

A system for substantially automating transcription services for multiple users (10, 11, 12) including a manual transcription station (50), speech recognition program (40) and a routing program (200). A uniquely identified voice dictation file is generated from a user and—based on the training status—routes the voice dictation file to a manual transcription station and speech recognition program. A human transcriptionist creates transcribed files for each voice dictation file. The speech recognition program creates written text for each dictation file if the training status is training or automated. If the training status of the current user is enrollment or training, a verbatim file is manually established and the speech recognition program is trained with an acoustic model using the verbatim and voice dictation files. The transcribed file is returned to the user if the training status is enrollment or training or written text is returned if the status is automated.

Claims

1. A system for substantially automating transcription services for one or more voice users, said system comprising:

means for creating a uniquely identified voice dictation file from a current user, said current user being one of said one or more voice users;

an audio player used to audibly reproduce said uniquely identified voice dictation file;

means for manually inputting and creating a transcribed file based on humanly perceived contents of said uniquely identified voice dictation file;

means for automatically converting said uniquely identified voice dictation file into written text;

means for manually editing a copy of said written text to create a verbatim text of said uniquely identified voice dictation file;

means for training said automatic speech converting means to achieve higher accuracy with said uniquely identified voice dictation file of current user; and

means for controlling the flow of said uniquely identified voice dictation file based upon a training status of said current user, whereby said controlling means sends said uniquely identified voice dictation file to at least one of said manual input means and said automatic speech converting means.

2. The invention according to claim 1 further comprising means for transferring said written text into a written text file.

3. The invention according to claim 1 wherein said written text is at least temporarily synchronized to said uniquely identified voice dictation file, said manual editing means comprises:

means for sequentially comparing a copy of said written text with said transcribed file resulting in a sequential list of unmatched words culled from said copy of said written text, said sequential list having a beginning, an end and a current unmatched word, said current unmatched word being successively advanced from said beginning to said end;

means for incrementally searching for said current unmatched word contemporaneously within a first buffer associated with the speech recognition program containing said written text and a second buffer associated with said sequential list; and

means for correcting said current unmatched word in said second buffer, said correcting means including means for displaying said current unmatched word in a manner substantially visually isolated from other text in said copy of said written text and means for playing a portion of said synchronized voice dictation recording from said first buffer associated with said current unmatched word.

4. The invention according to claim 3 wherein said correcting means further includes means for alternatively viewing said current unmatched word in context within said copy of said written text.

5. The invention according to claim 3 further including means for determining an accuracy rate for said current user.

6. The invention according to claim 5 wherein said verbatim file is a known accurate file, invention further includes means for determining skill of a human operator based on said accuracy rate.

7. The invention according to claim 4 wherein said sequential list and said written text each have a respective number of words, said accuracy rate determining means determines the ratio of said number of words in said sequential list to said number of words in said written text.

8. The invention according to claim 1 wherein said automatic speech converting means comprises a preexisting speech recognition program intended for human interactive use, said automatic speech converting means includes means for automating responses to a series of interactive inquiries from said preexisting speech recognition program.

9. The invention according to claim 8 wherein said training means comprises a preexisting training portion of said preexisting speech recognition program intended for human interactive use, said training means includes means for automating responses to a series of interactive inquiries from said preexisting training portion of said preexisting speech recognition program.

10. The invention according to claim 1 wherein said training means comprises a preexisting training program intended for human interactive use, said training means includes means for automating responses to a series of interactive inquiries from said preexisting training program.

11. The invention according to claim 1 wherein said control means reads and modifies a user profile associated with said current user, each of said user profiles including said training status of said current user.

12. The invention according to claim 11 wherein said training status is selected from the group of pre-enrollment, enrollment, training, automation and stop automation.

13. The invention according to claim 12 when said training status is pre-enrollment said control means further includes means for creating a user identification and acoustic model within said automatic speech converting means.

14. The invention according to claim 12 when said training status is enrollment said control means routes said voice dictation file to said automatic speech converting means and said manual input means, routes said written text and said transcribed file to said manual editing means, routes said verbatim text to said training means and routes said transcribed file back to said current user as a finished text.

15. The invention according to claim 12 when said training status is training said control means routes said voice dictation file to said automatic speech converting means and said manual input means, routes said written text and said transcribed file to said manual editing means, routes said verbatim text to said training means and routes said transcribed file back to said current user as a finished text.

16. The invention according to claim 12 when said training status is automation said control means routes said voice dictation file only to said automatic speech converting means and routes said written text back to said current user as a finished text.

17. An apparatus for substantially simplifying the production of a foreign language speech model for said speech recognition program wherein said foreign language provides a sufficient set of words to teach the voice dictation recording based upon a transcribed file produced by a human transcriptionist and a written text produced by a speech recognition program, wherein said written text is at least temporarily synchronized to said voice dictation recording, said apparatus comprising:

means for sequentially comparing a copy of said written text with said transcribed file resulting in a sequential list of unmatched words culled from said copy of said written text, said sequential list having a beginning, an end and a current unmatched word, said current unmatched word being successively advanced from said beginning to said end;

means for incrementally searching for said current unmatched word contemporaneously within a first buffer associated with the speech recognition program containing said written text and a second buffer associated with said sequential list; and

means for correcting said current unmatched word in said second buffer, said correcting means including means for displaying said current unmatched word in a manner substantially visually isolated from other text in said copy of said written text and means for playing a portion of said synchronized voice dictation recording from said first buffer associated with said current unmatched word.

18. The invention according to claim 17 wherein said correcting means further includes means for alternatively viewing said current unmatched word in context within said copy of said written text.

19. The invention according to claim 18 wherein said manner substantially visually isolated from other text can be manually selected from the group containing word-by-word display, sentence-by-sentence display, and said current unmatched word display.

20. A method for automating transcription services for one or more voice users in a system including a manual transcription station and a speech recognition program, said method comprising the steps of:

establishing a profile for each of the voice users, the profile containing a training status;

creating a uniquely identified voice dictation file from a current voice user;

choosing the training status of the current voice user from the group of enrollment, training, automated and stop automation;

routing the uniquely identified voice dictation file to at least one of the manual transcription station and the speech recognition program based on the training status;

receiving the uniquely identified voice dictation file in at least one of the manual transcription station and the speech recognition program;

creating a transcribed file at the manual transcription station for each received uniquely identified voice dictation file;

automatically creating a written text with the speech recognition program for each received uniquely identified voice dictation file if the training status of the current user is training or automated;

manually establishing a verbatim file if the training status of the current user is enrollment or training;

training the speech recognition program with an acoustic model for the current user using the verbatim file and the uniquely identified voice dictation file if the training status of the current user is enrollment or training;

returning the transcribed file to the current user if the training status of the current user is enrollment or training; and

returning the written text to the current user if the training status of the current user is automated.

21. The invention according to claim 20 wherein said step of manually establishing a verbatim file includes the sub-steps of:

assisting an operator to establish the verbatim file if the training status of the current user is training by;

sequentially comparing a copy of the written text with the transcribed file resulting in a sequential list of unmatched words culled from the copy of the written text, the sequential list having a beginning, an end and a current unmatched word, the current unmatched word being successively advanced from the beginning to the end;

incrementally searching for the current unmatched word contemporaneously within a first buffer associated with the speech recognition program containing the written text and a second buffer associated with the sequential list; and

displaying the current unmatched word in a manner substantially visually isolated from other text in the copy of the written text and playing a portion of the synchronized voice dictation recording from the first buffer associated with the current unmatched word; and

correcting the current unmatched word to be a verbatim representation of the portion of the synchronized voice dictation recording.

22. A method for testing the skills of a human transcriptionist using a known accurate written text created by a speech recognition program and a transcribed file created by the human transcriptionist, the method comprising:

sequentially comparing a copy of the written text with the transcribed file resulting in a sequential list of unmatched words culled from the copy of the written text, the sequential list having a beginning, an end and a current unmatched word, the current unmatched word being successively advanced from the beginning to the end;

incrementally searching for the current unmatched word contemporaneously within a first buffer associated with the speech recognition program containing the written text and a second buffer associated with the sequential list; and

displaying the current unmatched word in a manner substantially visually isolated from other text in the copy of the written text and playing a portion of the synchronized voice dictation recording from the first buffer associated with the current unmatched word; and

calculating the accuracy rate of the human transcriptionist.