US4528688A
Expired / Lapsed
US
⏱️ 0.0 years remaining
Continuous speech recognition method
Assignee: Hitachi Ltd
•
Filing: 1984-04-19
•
Granted: 1985-07-09
•
Expires: 2004-04-19
Abstract
Abstract This speech signal recognition system compares the two-dimensionals pattern (time sequence of feature vectors) of an unknown signal to prestored standard references patterns for recognition, thus forming a corresponding two-dimensional comparison pattern of points of elemental Hamming distance differences. The sum of the pattern point distances is the similarity measure. To improve accuracy, partial patterns are selected (or "masked") and tested sequentially, and the point values weighted relative to their location within the mask. The mask may be rectangular or oblique.
Inventors
- 👤 Akira Ichikawa
- 👤 Akio Komatsu
- 👤 Yoshiaki Asakawa
Bibliographic Details
Application Number
—
Jurisdiction
US
Data Source
google_patents
Last Synchronized
2026-10-25 05:26
CPC Classifications
G; G10; G10L; G10L15/00
Claims Analysis
Total Claims: 8
Claim 1
Independent Claim
164 words
A continuous speech recognition method comprising the steps: analyzing a continuous input speech pattern in time series, setting a predetermined two-dimensional mask for successively appointing a one unit region for matching which includes a plurality of analytical points of each of a standard pattern analyzed in time series in advance and the input speech pattern, calculating a parameter corresponding to a degree of similarity between said standard pattern and said input speech pattern within said one unit region on the basis of analytical results at said plurality of analytical points of said standard pattern and aid input speech pattern within said mask including the operations of allotting predetermined weights to respective values corresponding to differences of the analytical results at said plurality of analytical points of said standard pattern and said input speech pattern within said two-dimensional mask, and comparing said parameter corresponding to said degree of similarity with a predetermined threshold value, whereby said input speech pattern is successively recognized in time series.
Claim 2
Dependent Claim
164 words
A continuous speech recognition method comprising the steps: analyzing a continuous input speech pattern in time series, setting a predetermined two-dimensional mask for successively appointing a one unit region for matching which includes a plurality of analytical points of each of a standard pattern analyzed in time series in advance and the input speech pattern, calculating a parameter corresponding to a degree of similarity between said standard pattern and said input speech pattern within said one unit region on the basis of analytical results at said plurality of analytical points of said standard pattern and aid input speech pattern within said mask including the operations of allotting predetermined weights to respective values corresponding to differences of the analytical results at said plurality of analytical points of said standard pattern and said input speech pattern within said two-dimensional mask, and comparing said parameter corresponding to said degree of similarity with a predetermined threshold value, whereby said input speech pattern is successively recognized in time series.
Claim 3
Dependent Claim
38 words
A continuous speech recognition method as defined in claim 1, wherein said predetermined weights are assigned within said one unit region such that at least selected interior points are assigned greater weights than exterior parts of the region.
Claim 4
Dependent Claim
17 words
A continuous speech recognition method as defined in claim 2, wherein said one unit region is inclined.
Claim 5
Dependent Claim
159 words
A continuous speech recognition method comprising the steps: analyzing a continuous input speech pattern in time series at time separated analytical points, setting a predetermined two-dimensional mask for successively appointing a one unit region for matching which includes a plurality of the time separated analytical points from an input speech pattern analyzed in time series, calculating a distance parameter corresponding to a degree of similarity between said standard pattern and said input speech pattern within said one unit region from a mathematical analysis of a plurality of the time separated analytical points of said standard pattern and said input speech pattern within said region including operations of alloting predetermined weights to distances calculated at said plurality of analytical points of said standard pattern and said input speech pattern within said two-dimensional mask, and comparing said distance parameter corresponding to said degree of similarity with a predetermined threshold value, whereby said input speech pattern is successively recognized in time series.
Claim 6
Dependent Claim
159 words
A continuous speech recognition method comprising the steps: analyzing a continuous input speech pattern in time series at time separated analytical points, setting a predetermined two-dimensional mask for successively appointing a one unit region for matching which includes a plurality of the time separated analytical points from an input speech pattern analyzed in time series, calculating a distance parameter corresponding to a degree of similarity between said standard pattern and said input speech pattern within said one unit region from a mathematical analysis of a plurality of the time separated analytical points of said standard pattern and said input speech pattern within said region including operations of alloting predetermined weights to distances calculated at said plurality of analytical points of said standard pattern and said input speech pattern within said two-dimensional mask, and comparing said distance parameter corresponding to said degree of similarity with a predetermined threshold value, whereby said input speech pattern is successively recognized in time series.
Claim 7
Dependent Claim
42 words
A continuous speech recognition method as defined in claim 4, wherein said predetermined weights are assigned within said one unit region such that a greater weight is assigned to an interior part of the region than an exterior part of the region.
Claim 8
Dependent Claim
15 words
A continuous speech recognition method as defined in claim 5, wherein said region is inclined.
References Cited (Backward Citations)
USPTO Annuity & Maintenance Windows
| Annuity Stage | Window Opens | Payment Deadline | Surcharge Period Until | Estimated Fee | Status |
|---|---|---|---|---|---|
| 3.5 Years (Year 4) | 1988-07-08 | 1989-01-07 | 1989-07-09 | $2000.00 | Window Closed / Lapsed |
| 7.5 Years (Year 8) | 1992-07-08 | 1993-01-07 | 1993-07-09 | $3760.00 | Window Closed / Lapsed |
| 11.5 Years (Year 12) | 1996-07-08 | 1997-01-07 | 1997-07-09 | $7700.00 | Window Closed / Lapsed |