US4833714A
Expired / Lapsed
US
⏱️ 0.0 years remaining
Speech recognition apparatus
Assignee: Mitsubishi Electric Corp
•
Filing: 1988-08-04
•
Granted: 1989-05-23
•
Expires: 2008-08-04
Abstract
Abstract A word speech recognition apparatus recognizes a speech inputted to a microphone (11). A feature extracting portion (20) extracts a feature parameter based on a aural signal waveform outputted from the microphone. The feature extracting circuit comprises a pitch cycle extraction circuit (21) for extracting pitch frequency of the speech signal waveform, a digital filter (23) for extracting, as a feature parameter, spectrum data of the speech signal waveform, and a filter coefficient setting circuit (22) for setting a filter coefficient so that a resonance frequency of the digital filter is an integral multiple of the pitch frequency. The feature parameter extracted from the feature extracting circuit is stored in an input pattern memory (3). A recognition processing portion (50) evaluates similarity between the feature parameter stored in advance in a registration pattern memory (4) and the feature parameter stored in the input pattern memory, so that speech recognition processing is made. Improved signal in noise performance results from very narrow bandwidth filters.
Inventors
- 👤 Mitsuo Shimotani
- 👤 Masahiro Hibino
- 👤 Kenji Shima
Bibliographic Details
Application Number
—
Jurisdiction
US
Data Source
google_patents
Last Synchronized
2026-11-04 18:57
CPC Classifications
G; G10; G10L; G10L25/00; G10L25/78; G10L25/87; G10L15/00; G10L25/90
Claims Analysis
Total Claims: 13
Claim 1
Independent Claim
199 words
A speech recognition apparatus operational in an environment having a large amplitude noise spectrum, comprising: audio signal input means for converting an input voice waveform to an audio signal; feature extracting means for extracting feature parameters of said audio signal; input pattern storing means for storing said feature parameters of speech to be recognized as extracted by said feature extracting means; registration pattern storing means having stored therein feature parameters of a plurality of speeches extracted by said feature extracting means; and speech recognition processing means for detecting any similarity between said feature parameters stored in said input pattern storing means and those stored in said registration pattern storing means; said feature extracting means comprising: pitch frequency detecting means for detecting a pitch frequency of said audio signal from said audio signal input means; a band-pass filter having a very narrow bandwidth and a resonance frequency dependent upon a filter coefficient thereof for extracting spectrum data of said audio signal to provide said feature parameters, and filter coefficient setting means responsive to the pitch frequency detected by said pitch frequency detecting means for setting the filter coefficient of said band-pass filter as an integral multiple of said pitch frequency.
Claim 2
Dependent Claim
199 words
A speech recognition apparatus operational in an environment having a large amplitude noise spectrum, comprising: audio signal input means for converting an input voice waveform to an audio signal; feature extracting means for extracting feature parameters of said audio signal; input pattern storing means for storing said feature parameters of speech to be recognized as extracted by said feature extracting means; registration pattern storing means having stored therein feature parameters of a plurality of speeches extracted by said feature extracting means; and speech recognition processing means for detecting any similarity between said feature parameters stored in said input pattern storing means and those stored in said registration pattern storing means; said feature extracting means comprising: pitch frequency detecting means for detecting a pitch frequency of said audio signal from said audio signal input means; a band-pass filter having a very narrow bandwidth and a resonance frequency dependent upon a filter coefficient thereof for extracting spectrum data of said audio signal to provide said feature parameters, and filter coefficient setting means responsive to the pitch frequency detected by said pitch frequency detecting means for setting the filter coefficient of said band-pass filter as an integral multiple of said pitch frequency.
Claim 3
Dependent Claim
63 words
A speech recognition apparatus in accordance with claim 1, wherein said band-pass filter is provided as a single unit, and said filter coefficient setting means includes means for setting said filter coefficient in said digital filter in time sharing manner, said digital band-pass filter constituting a group of channel filters in response to said filter coefficient set by said filter coefficient setting means.
Claim 4
Dependent Claim
63 words
A speech recognition apparatus in accordance with claim 1, wherein a plurality of said band-pass filters are provided in parallel, and said filter coefficient setting means includes means for setting different filter coefficients respectively of said parallel digital band-pass filters, said digital band-pass filters forming a group of channel filters in accordance with the filter coefficients therein set by said coefficient setting means.
Claim 5
Dependent Claim
30 words
A speech recognition apparatus in accordance with claim 1, wherein said band-pass filter further includes means for adjusting the level of said audio signal in response to a level thereof.
Claim 6
Dependent Claim
42 words
A speech recognition apparatus in accordance with claim 1, wherein said band-pass filter further includes means for adjusting an output signal level in response to a resonance frequency signal level to make an output signal level at the respective resonance frequencies constant.
Claim 7
Dependent Claim
66 words
A speech recognition apparatus in accordance with claim 1, wherein said band-pass filter includes: a primary differential circuit for receiving said audio signal from said audio signal input mans, a two-stage lattice-type filter for receiving the output from said primary differential circuit, a square-law circuit for receiving the output from said two-stage lattice-type filter, and an integrating circuit for receiving the output from said square-law circuit.
Claim 8
Dependent Claim
85 words
A speech recognition apparatus in accordance with claim 1, wherein said band-pass filter includes: a multiplier which receives said audio signal from said audio signal input means for multiplying said audio signal by a waveform multiplying coefficient, a primary differential circuit for receiving the output from said multiplier, a two-stage lattice-type filter for receiving the output from said primary differential circuit, a square-law circuit for receiving the output from said two-stage lattice-type filter, and an integrating circuit for receiving the output from said square-law circuit.
Claim 9
Dependent Claim
81 words
A speech recognition apparatus in accordance with claim 1, wherein said band-pass filter includes: a primary differential circuit for receiving said audio signal from said audio signal input means, a two-stage lattice-type filter for receiving the output from said primary differential circuit, a square-law circuit for receiving the output from said two-stage lattice-type filter, an integrating circuit for receiving the output from said square-law circuit, and a multiplier for multiplying the output from said integrating circuit by a gain correction coefficient.
Claim 10
Dependent Claim
37 words
A speed recognition apparatus in accordance with claim 1, wherein said very narrow band width of said band-pass flter is less than 80 Hertz between points attenuated 3 db from a center frequency of said band-pass filter.
Claim 11
Dependent Claim
37 words
A speed recognition apparatus in accordance with claim 1, wherein said very narrow band width of said band-pass flter is less than 50 Hertz between points attenuated 3 db from a center frequency of said band-pass filter.
Claim 12
Dependent Claim
46 words
A speech recognition apparatus in accordance with claim 1, wherein said very narrow band width of sand band-pass filter is less than half of a pitch frequency of a vowel in human speech between points attenuated 3 db from a center frequency of said band-pass filter.
Claim 13
Dependent Claim
18 words
A speech recognition apparatus in accordance with claim 1, wherein said band-pass filter comprises a digital band-pass filter.
References Cited (Backward Citations)
US3989897A
Cited
US4538295A
Cited
USPTO Annuity & Maintenance Windows
| Annuity Stage | Window Opens | Payment Deadline | Surcharge Period Until | Estimated Fee | Status |
|---|---|---|---|---|---|
| 3.5 Years (Year 4) | 1992-05-22 | 1992-11-21 | 1993-05-23 | $2000.00 | Window Closed / Lapsed |
| 7.5 Years (Year 8) | 1996-05-22 | 1996-11-21 | 1997-05-23 | $3760.00 | Window Closed / Lapsed |
| 11.5 Years (Year 12) | 2000-05-22 | 2000-11-21 | 2001-05-23 | $7700.00 | Window Closed / Lapsed |