Data processing: speech signal processing – linguistics – language – Speech signal processing – Recognition
Reexamination Certificate
1998-02-03
2001-10-02
Korzuch, William (Department: 2641)
Data processing: speech signal processing, linguistics, language
Speech signal processing
Recognition
C704S273000
Reexamination Certificate
active
06298325
ABSTRACT:
BACKGROUND OF THE INVENTION
1. Field of the Invention
The present invention relates to a system for recognizing voices. More particularly, it relates to a speech recognition system for recognizing the voices of sequence data, for example, the names of stations, which ought to be inputted in regular order.
2. Description of the Related Art
A “speech recognition system for sequence data” is an apparatus which recognizes whether or not sequent words, for example, the names of the stations of the Yamanote Loop Line or the names of successive emperors, are uttered properly in regular order. The apparatus is considered to be chiefly applicable to game machines and training machines by way of example. There is not any prior-art example of the speech recognition system of the specified type. Considered as the simplest method of realizing such a system is a speech recognition method wherein the names of the stations of the Yamanote Loop Line, for example, are all employed as reference or subject words for speech recognition, and wherein voice data currently inputted is compared with all the reference words for recognition.
Such a method, however, has the problem that, since the number of the reference words for recognition is large, a long time is expended on the speech recognition. Another problem is that, when similar words, for example, “Meguro” and “Mejiro” exist within the sequence data, misrecognition is liable to occur.
SUMMARY OF THE INVENTION
An object of the present invention is to provide a speech recognition system for sequence data as shortens a time period required for recognition and as reduces the errors of the recognition.
A speech recognition system according to the present invention comprises a recognition-reference-word generation unit, a speech recognition unit, a dummied-sequence-data storage unit (namely, a storage unit for sequence data accompanied with dummy data), and a recognition-reference-word storage unit.
In the first aspect of the present invention, the recognition-reference-word generation unit adds at least one dummy data item to a sequence data item which is to be currently inputted in predetermined order, thereby to form a set of reference words for speech recognition. The speech recognition unit delivers the reference word which is the most similar to an externally inputted voice in the set of recognition reference words, as a recognized result on the basis of the feature of the input voice.
In the second aspect of the present invention, the dummied-sequence-data storage unit stores therein data items which have been obtained by adding at least one dummy data item to each sequence data item. The speech recognition unit reads out the sequence data item to be currently inputted in predetermined order, as well as the dummy data item or items corresponding to this sequence data item, from the dummied-sequence-data storage unit, thereby to form a set of reference words for speech recognition, and it delivers the reference word being the most similar to an externally inputted voice in the set of recognition reference words, as a recognized result on the basis of the feature of the input voice.
In the third aspect of the present invention, the recognition-reference-word storage unit stores all sequence data items as reference words for speech recognition therein. The speech recognition unit delivers the reference word which is the most similar to an externally inputted voice among the recognition reference words, as a recognized result on the basis of the feature of the input voice.
REFERENCES:
patent: 4156868 (1979-05-01), Levinson
patent: 4907278 (1990-03-01), Cecinnati et al.
patent: 5428707 (1995-06-01), Gould et al.
patent: 5432886 (1995-07-01), Tsukada et al.
patent: 5457768 (1995-10-01), Tsuboi et al.
patent: 5634086 (1997-05-01), Rtischev et al.
patent: 6-214590 (1994-08-01), None
Kimura Shinta
Koide Tomio
Satoh Yasuo
Fujitsu Limited
Korzuch William
Staas & Halsey , LLP
Storm Donald L.
LandOfFree
Speech recognition system for sequence data does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Speech recognition system for sequence data, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Speech recognition system for sequence data will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2551965