Data processing: speech signal processing – linguistics – language – Speech signal processing – Recognition
Reexamination Certificate
1999-04-13
2001-11-06
Dorvil, Richemond (Department: 2741)
Data processing: speech signal processing, linguistics, language
Speech signal processing
Recognition
C704S275000
Reexamination Certificate
active
06314397
ABSTRACT:
CROSS REFERENCE TO RELATED APPLICATIONS
(Not Applicable)
STATEMENT REGARDING FEDERALLY SPONSORED RESEARCH OR DEVELOPMENT
(Not Applicable)
BACKGROUND OF THE INVENTION
1. Technical Field
This invention relates to the field of speech recognition software and more particularly to propagation of corrections in speech recognition software.
2. Description of the Related Art
Speech recognition software has a wide variety of uses. One important area where this technology has proved to be very useful is the conversion of spoken utterances into text for inclusion in a word processing document. Word processing applications may have incorporated therein a speech recognition function. Alternatively, a variety of speech recognition programs are commercially available which provide the speech recognition function to existing non-speech enabled word-processing applications. In any case, the ability of these speech enabled word processing applications to convert human speech into text has improved dramatically in recent years. Due to a variety of factors, however, errors may yet occur in the speech recognition function performed by such applications. Accordingly, it is often necessary for a user to review a document which has been dictated to a word processing application by means of such speech recognition software.
One notable aspect of error words existing in a speech enabled word processing application is that they tend not to be misspelled, but are instead words which have been misrecognized. Such error words are typically a close acoustic match to the correct word, but often involve spelling variations or in some instances may be an entirely unrelated but similar sounding word or words. These errors may be difficult to detect by automated means such as conventional spell checking features provided with word processing applications. This is primarily due to the fact that such words tend to be spelled correctly but are otherwise erroneous in the context of the text.
Moreover, when a misrecognition occurs by a speech recognition engine in the course of converting a string of spoken utterances to text, such misrecognitions can tend to be repeated as the speech recognition engine continuously misrecognizes the same spoken utterance for the same or other erroneous text words. This may occur when the spoken utterance is out of vocabulary for the speech recognition engine or may be related to some other problem such as an individual's unconventional pronunciation of such word. In any case, the ability to efficiently correct such errors throughout a text document is important to the ability to efficiently make use of speech enabled word processing applications.
SUMMARY OF THE INVENTION
The invention concerns a method to propagate corrections in dictated text stored in a word processing application document file where the word processing application is equipped for speech recognition. In response to a first user command selecting a first error word to be corrected from the dictated text, the system replaces the first error word with a correct word selected by a user. Subsequently, in response to a second user command, the system searches the remainder of the dictated text in the word processing application document to locate a second or subsequent error word that is a close acoustic match to the first error word. The second or subsequent error word may then be replaced with the correct word.
According to one aspect of the invention, the second or subsequent error word may be located by text matching a word in the dictated text to an alternate word contained in an alternate word list for the first error word. According to another aspect of the invention, the second or subsequent error word may be located by text matching the first error word to a word in the dictated text.
For the convenience of a user, an alternate word list is displayed when the first error word is selected. The first error word may then be replaced by the correct word selected from the alternate word list. If the correct word is not in the list, the user preferably adds it and the correct word is then preferably displayed as part of the alternate word list when a subsequent error word is located.
According to another aspect, the invention may be embodied in a computer system having a word processing application adapted for speech recognition. In that case, the system includes application programming responsive to a first user command for selecting a first error word to be corrected in the document file, programming for replacing the first error word in the document file with a correct word selected by a user, and further programming for searching the document file in response to a second user command to locate a second or subsequent error word that is a close acoustic match to the first error word. According to one aspect, the computer system includes application programming for replacing the second or subsequent error word in the document file with the correct word.
According to another aspect of the invention application programming is provided for locating the second or subsequent error word by text matching a word in the document file to an alternate word contained in an alternate word list for the first error word. Alternatively, or in addition thereto, application programming can be provided for locating the second or subsequent error word by text matching the first error word to a word in the document file.
Further programming is provided for displaying an alternate word list when the first error word is selected. The system also provides programming for replacing the first error word with the correct word selected from the alternate word list. The software then adds the correct word to the alternate word list and displays the alternate word list when the second or subsequent error word is located.
Finally, the invention may take the form of a machine readable storage having stored thereon a computer program having a plurality of code sections executable by a machine for causing the machine to perform a set of steps including: a) storing in a memory a word processing application document file containing dictated text; b) in response to a first user command, selecting a first error word to be corrected in the dictated text; c) replacing the first error word with a correct word selected by a user; and d) searching the dictated text in response to a second user command to locate a second or subsequent error word that is a close acoustic match to the first error word.
REFERENCES:
patent: 4914704 (1990-04-01), Cole et al.
patent: 5761687 (1998-06-01), Hon et al.
patent: 5799273 (1998-08-01), Mitchell et al.
patent: 5909667 (1999-06-01), Leontiades et al.
patent: 5970460 (1999-10-01), Bunce et al.
patent: 6064959 (2000-05-01), Young et al.
Ballard Barbara
Lewis James R.
Ortega Kerry
Dorvil Richemond
International Business Machines Corp.
Senterfitt Akerman
LandOfFree
Method and apparatus for propagating corrections in speech... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Method and apparatus for propagating corrections in speech..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Method and apparatus for propagating corrections in speech... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2583870