Data processing: speech signal processing – linguistics – language – Speech signal processing – Recognition
Reexamination Certificate
1998-03-25
2001-02-27
Hudspeth, David R. (Department: 2741)
Data processing: speech signal processing, linguistics, language
Speech signal processing
Recognition
C704S250000, C704S256000
Reexamination Certificate
active
06195637
ABSTRACT:
BACKGROUND OF THE INVENTION
1. Field of the Invention
This invention relates generally to the field of speech recognition applications, and in particular, to a method for correcting dictation recognition errors.
2. Description of Related Art
Speech dictation systems improve over time if recognition of dictated words improves. The recognition of dictated words improves over time only if users take the time to correct misrecognized words. Thus, if a user sees an error, the user should correct the error. However, most users want to just finish their work and do not take the time to correct misrecognized words. Currently, it is often suggested that users dictate a paragraph and then correct the errors. Waiting any longer is likely to result in errors being overlooked or forgotten. According to presently accepted procedures to correct the errors, the user must first find the error or errors, then open a dialog box or window for correcting the error, and then make the correction, thus frequently interrupting the flow of the user's thoughts.
Error correction tools in speech applications now available, for example Vocabulary Expander, provides for automatically generating a list of words in the out of recognition category, that is, those words which appear to the speech application as having been misrecognized. This list is usually generated at the conclusion of dictation, although users can invoke the tool more frequently, at the cost of interrupting their dictation.
In order to improve speech recognition in speech systems, the user needs to correct the errors using the software application's correction method rather than just typing or dictating over the text. Only through such error correction does speech recognition improve. Accordingly, there is a long-felt and unsatisfied need to provide a method for error correction which does not interfere with or unduly interrupt the user's dictation.
SUMMARY OF THE INVENTION
In accordance with the inventive arrangements taught herein, this need is satisfied by a method which enables users to mark the misrecognition errors using a marking tool, and then at a later time, run a utility within the speech recognition application to correct the errors. This allows users to finish their work and then correct all the errors at once, instead of stopping every time they see an error or even at the end of every paragraph.
In a presently preferred embodiment, the misrecognized word or words can be marked as easily as highlighting or underlining the words as they appear on the screen, using a mouse or other pointing tool. The misrecognized words can also be marked by a speech command. Irrespective of the marking method, all misrecognized words can be marked for deferred correction, so that the user marked words and, if desired, additional words marked automatically by the application, can be corrected at the conclusion of the dictation. Marking and deferring correction of misrecognized words can significantly reduce the likelihood that misrecognized words will be missed, overlooked or forgotten.
A method for correcting misrecognition errors in accordance with an inventive arrangement comprises the steps of: dictating to a speech application; observing recognized words during the dictating step; marking misrecognized words during the dictating step; concluding dictation; and, displaying and correcting the marked misrecognized words, whereby the correcting of the misrecognized words is deferred until after the dictating step is concluded and the dictating step is not significantly interrupted.
The displaying and correcting step is advantageously implemented by invoking a correction tool of the speech application, whereby the correcting of the misrecognized words trains the speech application.
The marking step is implemented by activating a pointing tool or uttering voice commands or both.
The displaying and correcting step can comprise the steps of: sequentially displaying each sentence with a marked word; and, displaying a list of user selectable words to substitute for the marked word. The displaying and correcting step can further comprise the step of enabling users to enter a correction of the marked word by typing a word not on the list or by uttering a word not on the list.
The method displaying and correcting step can further comprise the step of enabling the user to select format corrections for the marked word in the sentence and for user selected text in the sentence.
The method can further comprise the step of generating a graphical user interface to implement the displaying and correcting step or generating respective graphical user interfaces for correcting text errors and format errors.
A method for correcting misrecognition errors in accordance with another inventive arrangement comprises the steps of: dictating to a speech application; marking misrecognized words during the dictating step; and, after the dictating and marking steps, displaying and correcting the marked misrecognized words, whereby the correcting of the misrecognized words is deferred until after the dictating step is concluded and the dictating step is not significantly interrupted. The displaying and correcting step is advantageously implemented by invoking a correction tool of the speech application, whereby the correcting of the misrecognized words trains the speech application.
REFERENCES:
patent: 5799273 (1998-08-01), Mitchell et al.
patent: 5799279 (1998-08-01), Gould et al.
patent: 5855000 (1998-12-01), Waibel et al.
patent: 5857099 (1999-01-01), Mitchell et al.
patent: 5864805 (1999-01-01), Chen et al.
patent: 5884258 (1999-03-01), Rozak et al.
patent: 5899976 (1999-05-01), Rozak
patent: 6006183 (1999-12-01), Lai et al.
Ballard Barbara E.
Ortega Kerry A.
Abebe Daniel
Hudspeth David R.
International Business Machines Corp.
Quarles & Brady LLP
LandOfFree
Marking and deferring correction of misrecognition errors does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Marking and deferring correction of misrecognition errors, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Marking and deferring correction of misrecognition errors will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2586669