US5369728A
Expired / Lapsed
US
⏱️ 0.0 years remaining
Method and apparatus for detecting words in input speech data
Assignee: Canon Inc
•
Filing: 1992-06-09
•
Granted: 1994-11-29
•
Expires: 2012-06-09
Abstract
Abstract An apparatus and method for recognizing speech includes a memory for storing data representing a reference pattern composed of the combination of a word reference pattern and a silence pattern, and a calculator for calculating the differences between data representing the reference pattern and data representing input speech. The use of such a silence pattern in the reference pattern permits a word such as "other" to be distinguished from the word "mother".
Inventors
- 👤 Tetsuo Kosaka
- 👤 Atsushi Sakurai
- 👤 Junichi Tamura
- 👤 Hiroshi Matsuo
Bibliographic Details
Application Number
—
Jurisdiction
US
Data Source
google_patents
Last Synchronized
2026-11-01 05:14
CPC Classifications
G; G10; G10L; G10L15/00; G10L15/08; G10L15/10
Claims Analysis
Total Claims: 24
Claim 1
Independent Claim
97 words
A method of detecting words in input speech data, comprising the steps of: storing data representing a reference pattern which comprises a combination of (i) a word reference pattern and (ii) silence reference patterns which are each sufficiently long to take into account silent sections of speech arising from a geminated consonant or a silent plosive consonant; calculating distances between data representing the reference pattern and data representing input speech; and detecting words in the data representing input speech based on the calculated distances between the data representing the reference pattern and the data representing input speech.
Claim 2
Independent Claim
97 words
A method of detecting words in input speech data, comprising the steps of: storing data representing a reference pattern which comprises a combination of (i) a word reference pattern and (ii) silence reference patterns which are each sufficiently long to take into account silent sections of speech arising from a geminated consonant or a silent plosive consonant; calculating distances between data representing the reference pattern and data representing input speech; and detecting words in the data representing input speech based on the calculated distances between the data representing the reference pattern and the data representing input speech.
Claim 3
Dependent Claim
41 words
The method as recited in claim 1, wherein said calculating step calculates distances between data representing the reference pattern and data representing input speech while shifting the data representing the reference pattern with respect to the data representing the input speech.
Claim 4
Dependent Claim
30 words
The method as recited in claim 2, wherein said storing step comprises the step of storing a non-word portion of the data representing input speech as the silence reference pattern.
Claim 5
Dependent Claim
28 words
The method as recited in claim 3, further comprising the step of sequentially updating the non-word portion of the data representing input speech as the silence reference pattern.
Claim 6
Dependent Claim
31 words
The method as recited in claim 2, further comprising the step of determining the value of speech parameters of data representing input speech and the reference pattern before said calculating step.
Claim 7
Independent Claim
106 words
An apparatus for detecting words in input speech data, comprising: input means for inputting speech data; word reference pattern storing means for storing word reference pattern data; reference pattern synthesizing means for synthesizing reference pattern data from silence reference pattern data which is sufficiently long to take into account silent sections of speech arising from a geminated consonant or a silent plosive consonant; calculating means for calculating distances between the input speech data and the synthesized reference pattern data; and detecting means for detecting words in the input speech data based on the calculated distances between the synthesized reference pattern data and the input speech data.
Claim 8
Independent Claim
106 words
An apparatus for detecting words in input speech data, comprising: input means for inputting speech data; word reference pattern storing means for storing word reference pattern data; reference pattern synthesizing means for synthesizing reference pattern data from silence reference pattern data which is sufficiently long to take into account silent sections of speech arising from a geminated consonant or a silent plosive consonant; calculating means for calculating distances between the input speech data and the synthesized reference pattern data; and detecting means for detecting words in the input speech data based on the calculated distances between the synthesized reference pattern data and the input speech data.
Claim 9
Dependent Claim
39 words
The apparatus as recited in claim 6, wherein said calculating means calculates the distances between the input speech data and the synthesized reference pattern data while shifting the synthesized reference pattern data with respect to the input speech data.
Claim 10
Dependent Claim
64 words
The apparatus as recited in claim 7, further comprising a non-word portion storing means for storing a non-word portion of the input speech data representing a portion of the input speech in which a word is absent, wherein said reference pattern synthesizing means synthesizes the reference pattern data by using the non-word portion of the input speech data as the silence reference pattern data.
Claim 11
Dependent Claim
28 words
The apparatus as recited in claim 8, further comprising means for sequentially updating the non-word portion of the input speech data stored in said non-word portion storing means.
Claim 12
Dependent Claim
27 words
The apparatus as recited in claim 6, wherein said calculating means determines values of speech parameters of the input speech data and the synthesized reference pattern data.
Claim 13
Dependent Claim
19 words
The apparatus as recited in claim 6, wherein said word reference pattern storing means is a read only memory.
Claim 14
Dependent Claim
19 words
The apparatus as recited in claim 6, wherein said word reference pattern storing means is a random access memory.
Claim 15
Independent Claim
103 words
A method of detecting words in input speech data, comprising the steps of: inputting speech data; synthesizing reference pattern data from stored word reference pattern data and stored silence reference pattern data which is at least twice the size of silence reference pattern data representing a germinated consonant; shifting the reference pattern data frame by frame with respect to the input speech data; calculating distances between the input speech data and the reference pattern data for each frame of data; and detecting words in the input speech data based on the calculated distances between the input speech data and the reference pattern data.
Claim 16
Independent Claim
103 words
A method of detecting words in input speech data, comprising the steps of: inputting speech data; synthesizing reference pattern data from stored word reference pattern data and stored silence reference pattern data which is at least twice the size of silence reference pattern data representing a germinated consonant; shifting the reference pattern data frame by frame with respect to the input speech data; calculating distances between the input speech data and the reference pattern data for each frame of data; and detecting words in the input speech data based on the calculated distances between the input speech data and the reference pattern data.
Claim 17
Dependent Claim
25 words
A method according to claim 13, wherein the silence reference pattern data is determined by calculations performed on selected portions of the input speech data.
Claim 18
Independent Claim
135 words
An apparatus for detecting words in input speech data, comprising: input means for inputting speech data; word reference storing means for storing word reference pattern data; silence reference storing means for storing silence reference pattern data which is at least twice the size of silence reference pattern data representing a geminated consonant; synthesizing means for synthesizing reference pattern data from the word reference pattern data and the silence reference pattern data; shifting means for shifting the reference pattern data frame by frame with respect to the input speech data; calculating means for calculating distances between the input speech data and the reference pattern data for each frame of data; and detecting means for detecting words in the input speech data based on the calculated distances between the input speech data and the reference pattern data.
Claim 19
Independent Claim
135 words
An apparatus for detecting words in input speech data, comprising: input means for inputting speech data; word reference storing means for storing word reference pattern data; silence reference storing means for storing silence reference pattern data which is at least twice the size of silence reference pattern data representing a geminated consonant; synthesizing means for synthesizing reference pattern data from the word reference pattern data and the silence reference pattern data; shifting means for shifting the reference pattern data frame by frame with respect to the input speech data; calculating means for calculating distances between the input speech data and the reference pattern data for each frame of data; and detecting means for detecting words in the input speech data based on the calculated distances between the input speech data and the reference pattern data.
Claim 20
Dependent Claim
31 words
An apparatus according to claim 15, further comprising silence reference pattern calculating means for performing calculations on selected portions of the input speech data to determine the silence reference pattern data.
Claim 21
Dependent Claim
18 words
The apparatus as recited in claim 13, wherein said word reference storing means is a read only memory.
Claim 22
Dependent Claim
18 words
The apparatus as recited in claim 13, wherein said word reference storing means is a random access memory.
Claim 23
Dependent Claim
18 words
The apparatus as recited in claim 13, wherein said silence reference storing means is a read only memory.
Claim 24
Dependent Claim
18 words
The apparatus as recited in claim 13, wherein said silence reference storing means is a random access memory.
References Cited (Backward Citations)
US4489435A
Cited
US4481593A
Cited
US4596032A
Cited
US4802226A
Cited
US4718095A
Cited
US4627091A
Cited
US4712243A
Cited
US4817159A
Cited
US4736429A
Cited
US4783807A
Cited
US4821325A
Cited
US4811399A
Cited
US4856067A
Cited
USPTO Annuity & Maintenance Windows
| Annuity Stage | Window Opens | Payment Deadline | Surcharge Period Until | Estimated Fee | Status |
|---|---|---|---|---|---|
| 3.5 Years (Year 4) | 1997-11-28 | 1998-05-30 | 1998-11-29 | $2000.00 | Window Closed / Lapsed |
| 7.5 Years (Year 8) | 2001-11-28 | 2002-05-30 | 2002-11-29 | $3760.00 | Window Closed / Lapsed |
| 11.5 Years (Year 12) | 2005-11-28 | 2006-05-30 | 2006-11-29 | $7700.00 | Window Closed / Lapsed |